{"id":"05e1e2a5-5503-471f-8a52-01fb3affc389","entityType":"agent","slug":"clawhub-deciqai-social-proof","name":"Social Proof","canonicalUrl":"https://www.xpersona.co/agent/clawhub-deciqai-social-proof","canonicalPath":"/agent/clawhub-deciqai-social-proof","generatedAt":"2026-10-11T20:58:45.156Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-11T16:39:46.460Z","emptyReason":null},"description":"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or con... Skill: Social Proof Owner: deciqai Summary: Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or con... Tags: latest:1.0.5 Version history: v1.0.5 | 2026-07-16T18:16:28.146Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/social-proof.json) v1.0.4 | 2026-07-10T10:28:09.728Z | user A","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1K downloads reported by the source. Last updated 10/11/2026.","installCommand":"clawhub skill install s17a4mqcnk515kvaca5ze55d0x88pfpx:social-proof","sourceUrl":"https://clawhub.ai/deciqai/social-proof","homepage":"https://clawhub.ai/deciqai/skills/social-proof","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/deciqai/social-proof","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/deciqai/skills/social-proof","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":60,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or con..."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-11T16:39:46.460Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T16:39:46.460Z","emptyReason":null},"stars":null,"forks":null,"downloads":1026,"packageName":null,"latestVersion":"1.0.5","tractionLabel":"1K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T16:39:46.391Z","emptyReason":null},"lastUpdatedAt":"2026-10-11T16:39:46.460Z","lastCrawledAt":"2026-10-11T16:39:46.391Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-12T16:39:46.391Z","lastVerifiedAt":null,"highlights":[{"version":"1.0.5","createdAt":"2026-07-16T18:16:28.146Z","changelog":"Description tail link + agents machine-readable metadata line (deciqai.com/s/social-proof.json)","fileCount":6,"zipByteSize":17943},{"version":"1.0.4","createdAt":"2026-07-10T10:28:09.728Z","changelog":"Add 2024-2026 AI-era worked example + updated sources","fileCount":6,"zipByteSize":17940},{"version":"1.0.3","createdAt":"2026-07-08T11:19:54.704Z","changelog":"Footer now uses /c/<slug> short link (fixes UTM truncation when SKILL.md is read in a terminal)","fileCount":5,"zipByteSize":12855},{"version":"1.0.2","createdAt":"2026-07-08T01:04:45.309Z","changelog":"Refreshed content + GitHub star link in footer","fileCount":5,"zipByteSize":12956},{"version":"1.0.1","createdAt":"2026-07-07T22:33:29.825Z","changelog":"Add catalog categories and topics","fileCount":5,"zipByteSize":12758},{"version":"1.0.0","createdAt":"2026-07-02T14:17:55.577Z","changelog":"Initial publish","fileCount":5,"zipByteSize":12758}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17a4mqcnk515kvaca5ze55d0x88pfpx:social-proof","setupComplexity":"low","setupSteps":["Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-11T20:58:45.153Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-social-proof/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"high","updatedAt":"2026-10-11T16:39:46.460Z","emptyReason":null},"readme":"Skill: Social Proof\n\nOwner: deciqai\n\nSummary: Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or con...\n\nTags: latest:1.0.5\n\nVersion history:\n\nv1.0.5 | 2026-07-16T18:16:28.146Z | user\n\nDescription tail link + agents machine-readable metadata line (deciqai.com/s/social-proof.json)\n\nv1.0.4 | 2026-07-10T10:28:09.728Z | user\n\nAdd 2024-2026 AI-era worked example + updated sources\n\nv1.0.3 | 2026-07-08T11:19:54.704Z | user\n\nFooter now uses /c/<slug> short link (fixes UTM truncation when SKILL.md is read in a terminal)\n\nv1.0.2 | 2026-07-08T01:04:45.309Z | user\n\nRefreshed content + GitHub star link in footer\n\nv1.0.1 | 2026-07-07T22:33:29.825Z | user\n\nAdd catalog categories and topics\n\nv1.0.0 | 2026-07-02T14:17:55.577Z | user\n\nInitial publish\n\nArchive index:\n\nArchive v1.0.5: 6 files, 17943 bytes\n\nFiles: examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md (9328b), examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md (11318b), references/sources.md (3748b), skill-card.md (2746b), SKILL.md (10170b), _meta.json (131b)\n\nFile v1.0.5:SKILL.md\n\n---\nname: social-proof\ndescription: \"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or consensus is real, suspects fake reviews or manufactured engagement, or is designing testimonials/UX to convert users. Do NOT activate when: the decision is low-stakes and reversible (choosing a lunch spot) or the user already has direct measured evidence stronger than any consensus signal. More: deciqai.com/c/social-proof\"\n---\n\n# Social Proof\n\n## Overview\n\n**Social proof**: we judge what is correct, normal, or worth doing by observing what others — especially similar others — are doing. Usually efficient; failure mode is severe: under unanimous consensus, people publicly endorse answers they privately know are wrong (Asch 1951–56: error rate <1% alone, ~37% under group pressure). Two amplifiers: **uncertainty** (social proof fills the vacuum) and **similarity** (same-type peers drive far stronger conformity than generic crowds).\n\nComposes with `reciprocity` (Cialdini's two primary levers), `anchoring` (price tiers often function as quasi-social-proof), and `critical-thinking` (structured fallback when consensus has been engineered).\n\n## When to Use\n\n**Use when:** purchase/hiring/investment decision leaning on what others chose; proposal cites \"everyone is doing this\"; designing growth/marketing/UX with social-proof patterns; decision feels unsafe alone without a clear reason; suspecting manufactured consensus (bots, paid reviews, astroturf); a trend is accelerating and private doubt is being suppressed by the fact everyone is on board; a \"we must adopt AI because every competitor is deploying it\" mandate is driving procurement or a pilot ahead of any validated ROI (AI hype / FOMO buying).\n\n**Do NOT use when:** decision is low-stakes and reversible; you have direct measured evidence stronger than any consensus; the \"consensus\" is from verified domain experts with better epistemic position; you want to rationalize a contrarian position that lacks independent evidence.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has a concrete case → run The Process directly.\n- **Coach mode:** user is unfamiliar or has no concrete case → guide, don't lecture.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** We judge what's correct by looking at what others do — useful most of the time, but under enough unanimous consensus, people will publicly agree with answers they privately know are wrong, even on obvious questions.\n2. **Check fit** against When to Use / When NOT to use. If direct evidence is stronger, point there.\n3. **Elicit the real situation.** A concrete decision shaped by what others are doing, or a design problem deploying social proof. Never run on hypotheticals.\n> **[WAIT — do not advance until user responds]**\n4. **One element at a time.** Walk through: what's the consensus, who are the consensus-makers, are they similar to you / informed, would you decide the same way if alone — wait for input.\n> **[WAIT — do not advance until user responds]**\n5. **Close by naming the payoff.** The one move — accept the consensus, reject it, or seek independent evidence — that fits their situation.\n> **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Social-Proof Analysis**. Diagnose the source of consensus, then decide whether to use it as evidence.\n\n1. **Name the consensus precisely.** Not \"everyone uses Salesforce\" but \"three cohort companies I respect use Salesforce.\" Vague consensus cannot be analyzed.\n2. **Identify consensus-makers.** Who exactly, how many, how similar to you in ways relevant to the decision?\n3. **Classify consensus type.** Informational (converged on evidence) | Social (converged because others did — cascade risk) | Manufactured (engineered appearance via bots, paid reviews, cherry-picked cases).\n4. **Test signal strength.** Did consensus form independently or in chain? What are dissenters saying? What is the base rate for consensus being wrong in this domain?\n5. **Run the Asch counterfactual.** Alone, with only the underlying evidence, would you reach the same conclusion? If no — you've been pulled in by the rule itself.\n6. **Check manufactured-consensus signs:** astroturf, survivorship bias, selected testimonials, engagement-metric inflation.\n7. **As a sender:** real named testimonials, third-party reviews, transparent distributions, limitations in case studies.\n8. **Stop-rule:** if you cannot defend the decision independently of \"many others are doing it,\" treat it as provisional. Plan a fallback.\n\n### Output template\n\n```\nConsensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]\n```\n\n*→ Method in Action: [Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md)*\n\n*→ 2026 lens: [Enterprise AI Copilot FOMO Procurement (2024–2026)](examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md)*\n\n## Pack: Social Proof in Practice\n\n- **Sender (marketing/sales):** named logos, real attributed testimonials, third-party reviews, transparent star distributions, case studies with limitations stated. Reference customers similar to the prospect.\n- **Receiver (evaluation):** check distribution shapes not averages; distinguish trial users from paying customers; named accounts are top-decile success cases — ask about failures.\n- **Product UX:** \"popular choice\" defaults are powerful — notice when a default is doing your decision-making for you.\n- **Astroturf defense:** anomalous account creation timing, repeated language across \"independent\" voices, engagement metrics inconsistent with audience size → drop consensus signal to near-zero.\n\n## Applying It Well\n\n- Similarity is load-bearing: \"other founders chose this\" works on a founder; \"many people chose this\" does not. Always identify whether consensus-makers are similar to you in relevant ways.\n- Three well-placed testimonials approach the power of fifteen — the rule saturates at small numbers.\n- A single visible dissenter destroys most conformity pressure. Find the dissenter or be the dissenter.\n- The rule operates below introspection. The Asch counterfactual is the test, not the self-report.\n- Manufactured social proof has a reputational cliff when discovered. Real social proof compounds.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] \"Everyone is doing it, so it must be right\" | Asch showed this fails on tasks where the right answer was visually obvious. \"Everyone\" is one signal to weigh — not the conclusion. |\n| [D] \"Many smart people are doing X, so X is right\" | Smart people are more invested in being seen as smart, raising the cost of public dissent. The \"smart people\" filter does not defend against engineered consensus. |\n| [D] \"I have my own opinion regardless of what others do\" | Asch's 75% applies even to people who predicted this of themselves. The rule operates below introspection — the counterfactual is the test. |\n| [D] \"Bestseller / most-popular must be the best\" | Popularity reflects discoverability and marketing, not necessarily quality. Treat it as a prior, then update on evidence. |\n| [D] \"If it were wrong, more people would have noticed\" | Public dissent is rare even when private dissent is widespread (Theranos, FTX, 2008 housing market). |\n| [D] \"I noticed the manufactured social proof, so I'm immune\" | Recognition reduces but does not eliminate the pull. Treat recognition as the start of defense, not the conclusion. |\n| [D] Confusing aggregated wisdom with social proof | Markets aggregate information. Social proof aggregates behavior — which may not reflect information. Distinguish \"independent estimators converged\" from \"people followed early movers.\" |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Decision driven by \"many others are doing it\" with no underlying analysis\n- Consensus-makers cannot be verified as similar to you in relevant ways\n- No dissenters visible in a domain where you would expect dissent\n- Consensus measured in low-cost actions (likes, sign-ups) not high-cost ones (purchases, repeat usage)\n- Star distributions are suspiciously skewed (all 5-stars or U-shaped)\n- Social-proof claim growing faster than the underlying user/evidence base could plausibly support\n\n## Verification\n\n- [ ] Consensus claim named precisely (specific group, not generic mass)\n- [ ] Consensus-makers identified — number, similarity, base of information\n- [ ] Consensus classified: informational / social / manufactured\n- [ ] Asch counterfactual performed: same conclusion alone?\n- [ ] Dissenting voices sought; their reasoning evaluated\n- [ ] Manufactured-consensus signs checked\n- [ ] If following: early-warning indicator specified\n- [ ] If sending: proof is real, named, verifiable, representative\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 227 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/c/social-proof** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\n*Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/social-proof.json*\n\nFile v1.0.5:_meta.json\n\n{\n  \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n  \"slug\": \"social-proof\",\n  \"version\": \"1.0.5\",\n  \"publishedAt\": 1784225788146\n}\n\nFile v1.0.5:references/sources.md\n\n# Sources — social-proof\n\n> *Primary sources for the [social-proof](../SKILL.md) skill.*\n\n- Asch, S. E. (1951). \"Effects of Group Pressure upon the Modification and Distortion of Judgments.\" In H. Guetzkow (Ed.), *Groups, Leadership, and Men: Research in Human Relations* (pp. 177–190). Pittsburgh: Carnegie Press. The original publication of the line-matching conformity experiments at Swarthmore.\n- Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), 31–35. The most accessible account, written by Asch for a general audience; primary source for the verbatim quotations on the 36.8% conformity rate, the categories of conformity (distortion of perception/judgment/action), the group-size threshold, and the dissenter effect. https://doi.org/10.1038/scientificamerican1155-31\n- Asch, S. E. (1956). \"Studies of Independence and Conformity: I. A Minority of One Against a Unanimous Majority.\" *Psychological Monographs: General and Applied*, 70(9), 1–70. The full monograph documentation. https://doi.org/10.1037/h0093718\n- Sherif, M. (1935). \"A Study of Some Social Factors in Perception.\" *Archives of Psychology*, 27(187), 1–60. The autokinetic-effect experiments; the canonical demonstration that under ambiguity, group norms substitute for objective information. The historical precursor to the Asch line.\n- Milgram, S., Bickman, L., & Berkowitz, L. (1969). \"Note on the Drawing Power of Crowds of Different Size.\" *Journal of Personality and Social Psychology*, 13(2), 79–82. The New York City sidewalk field experiment: one confederate looking up at a building draws ~4% of passers-by, fifteen confederates draw ~86%. https://doi.org/10.1037/h0028070\n- Cialdini, R. B. (1984; 2007 rev.). *Influence: The Psychology of Persuasion*. Harper Business. Chapter 4 is the canonical applied treatment of social proof, including the uncertainty and similarity amplifiers. ISBN 978-0061241895.\n- Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). \"A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades.\" *Journal of Political Economy*, 100(5), 992–1026. The formal model of information cascades — how rational agents looking at the choices of earlier agents can produce socially-proof-style mass behavior even when each individual's private information would have led to a different conclusion. https://doi.org/10.1086/261849\n- Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). \"Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market.\" *Science*, 311(5762), 854–856. The Music Lab experiment: 14,000+ subjects shown otherwise-identical songs converged on radically different \"winners\" depending on the social-proof signals they were shown — empirical demonstration that the consensus a market produces is often determined by social-proof dynamics rather than underlying quality. https://doi.org/10.1126/science.1121066\n- Gartner (2024). \"Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025.\" Press release, July 29, 2024. Contemporary evidence that enterprise generative-AI adoption in 2024–2025 outran validated ROI — cited factors include poor data quality, inadequate risk controls, escalating costs, and unclear business value. Illustrates social-proof / FOMO-driven procurement in a live hype cycle. https://www.gartner.com/en/newsroom\n- McKinsey & Company (2024–2025). \"The State of AI\" (annual global survey). Documents the widening gap between rapid generative-AI adoption and the share of organizations able to demonstrate material financial return — the empirical backdrop for consensus-driven (\"everyone is deploying copilots\") AI buying. https://www.mckinsey.com\n\nFile v1.0.5:examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md\n\n# Method in Action: Enterprise AI Copilot FOMO Procurement (2024–2026)\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nBetween the launch of ChatGPT in late 2022 and 2026, generative AI moved from novelty to boardroom mandate. By 2024–2025 a recognizable pattern had set in across large organizations: an executive would say some version of *\"every competitor is deploying AI copilots — we can't be the ones left behind,\"* and a procurement cycle would begin. The dominant justification for buying was not a measured business case; it was the observation that **everyone else was buying**. This is a textbook social-proof failure mode — consensus behavior substituting for independent evidence — and it is worth walking through this skill's own Process, because the same pattern recurs in every hype cycle.\n\nThe pattern was widely documented. Industry surveys and analyst commentary through 2024–2025 repeatedly noted a gap between adoption rates and realized returns: adoption of generative AI rose sharply, while a large share of organizations reported difficulty demonstrating financial return from their AI initiatives. A frequently discussed data point was a 2024 Gartner prediction that a substantial share of generative-AI projects would be abandoned after proof-of-concept by the end of 2025, citing poor data quality, unclear business value, escalating costs, and inadequate risk controls. The *phenomenon* — FOMO-driven procurement outrunning validated ROI — was well established before this analysis. Let us run it through the eight Process steps.\n\n## 1. Name the consensus precisely\n\nThe default framing — *\"everyone is deploying AI copilots\"* — is exactly the vague consensus the skill warns against. It cannot be analyzed. Sharpen it:\n\n- Not \"everyone is deploying copilots\" but: *\"three named direct competitors announced Microsoft 365 Copilot or a comparable assistant rollout in press releases and earnings calls, and our board saw those announcements.\"*\n- Distinguish an **announced pilot** (\"we are exploring generative AI\") from a **validated, ROI-positive production deployment**. Public announcements are overwhelmingly the former. The consensus you can actually verify is a consensus of *announcements*, not a consensus of *proven results*.\n\nOnce named precisely, most of the perceived consensus evaporates: what \"everyone\" has done is issue a press release and stand up a pilot — not demonstrate return.\n\n## 2. Identify consensus-makers\n\nWho exactly is the consensus, how many, and how similar to you on dimensions relevant to *whether the tool will pay off*?\n\n- **Vendors and their platform partners** have direct commercial incentive to project a \"everyone is adopting\" narrative.\n- **Competitors' investor-relations announcements** are similar to you in industry but are optimizing for a stock-price and talent-signaling audience, not for disclosing pilot failure rates.\n- **Analysts and the trade press** amplify the loud adopters; the companies that quietly ran a pilot and shelved it rarely issue a press release.\n\nRelevant similarity is thin. The consensus-makers share your *industry* but not your *data readiness, workflow, or the honest internal ROI numbers* — the variables that actually determine whether a copilot pays off.\n\n## 3. Classify consensus type\n\n- **Informational** (converged on evidence): weak. Very few adopters had published rigorous, independent ROI data by 2024–2025; many were still in pilot.\n- **Social** (converged because others did — cascade risk): dominant. Each competitor's visible move raised the perceived cost of *not* moving, independent of any new evidence — the exact structure of an informational cascade (Bikhchandani, Hirshleifer & Welch, 1992).\n- **Manufactured** (engineered appearance): partial. Vendor-sponsored case studies, curated success stories, and \"N% of enterprises are adopting AI\" marketing statistics inflate the apparent consensus.\n\nVerdict: **primarily social with a manufactured overlay, thin on informational content.**\n\n## 4. Test signal strength\n\nDid the consensus form independently or in a chain? Almost entirely a chain: Company A's announcement fed Company B's board deck fed Company C's mandate. This is cascade behavior, not independent estimators converging.\n\nWhat are dissenters saying? The dissenting evidence was public and loud by 2024–2025 — analyst warnings of high proof-of-concept abandonment, surveys showing most firms could not yet quantify financial return, and widely reported early-mover disappointments. The base rate for consensus being wrong in enterprise technology hype cycles is high (compare the earlier big-data and blockchain waves, where adoption announcements far outran realized ROI).\n\n## 5. Run the Asch counterfactual\n\nThe core test: *alone, with only the underlying evidence and no awareness of what competitors were doing, would we buy this now, at this price, at this scale?*\n\nFor most FOMO buyers the honest answer is **no** — absent the competitive signal, they would run a small, instrumented pilot with a defined success metric before a broad rollout. If the decision flips the moment you remove the \"everyone is doing it\" signal, you have been pulled in by the rule itself, not by evidence. This is the Asch \"distortion of action\" pattern at organizational scale: decision-makers who privately doubt the ROI still publicly back the purchase rather than be the executive who \"fell behind on AI.\"\n\n## 6. Check manufactured-consensus signs\n\n- **Astroturf / curated cases:** vendor case studies are top-decile successes; ask for the failure distribution and the denominator, not the showcase logo.\n- **Survivorship bias:** the shelved pilots don't issue press releases, so the visible population is biased toward apparent winners.\n- **Metric inflation:** \"adoption\" is often measured by seats provisioned or logins (low-cost actions), not by sustained usage, workflow integration, or booked savings (high-cost actions). Seats purchased is not value realized.\n\n## 7. As a sender\n\nIf *you* are marketing an AI tool and want durable rather than fragile social proof: use real named reference customers similar to the prospect, disclose the *distribution* of outcomes (including where it underperformed), and quantify ROI with stated limitations rather than a single hero anecdote. Manufactured AI hype has a reputational cliff when the abandonment numbers surface; honest, representative proof compounds as the market matures past the peak of the cycle.\n\n## 8. Stop-rule\n\nIf you cannot defend the AI purchase independently of \"our competitors are all doing it,\" treat it as **provisional**. The disciplined move is not \"don't adopt AI\" — the technology is real and the upside can be large. The disciplined move is to **replace the cascade with your own evidence**: a bounded pilot, a pre-committed success metric (a specific productivity or cost KPI on a specific workflow), and a fallback plan if the pilot misses.\n\n## Output template, filled\n\n```\nConsensus claim: \"3 named competitors announced copilot rollouts\" (announcements, not proven ROI); consensus-makers similar in industry, not in data-readiness or disclosed results\nConsensus type: primarily social (cascade) + manufactured overlay; thin informational content\nIndependent vs. chain: chain — each competitor's move fed the next; high cascade risk\nDissenters: analysts warning of high proof-of-concept abandonment; surveys showing most firms cannot yet quantify AI financial return\nAsch counterfactual: alone with only evidence, most would run a metered pilot first, not a broad rollout → no\nManufactured-consensus check: curated vendor case studies, survivorship (shelved pilots stay silent), adoption measured by seats not sustained value\nDecision: seek independent evidence — bounded pilot with a pre-committed ROI metric before scaling\nEarly-warning trigger: pilot misses its pre-set KPI; usage decays after novelty; per-seat cost rises without measured workflow savings\n```\n\n## What this case teaches\n\nThe 2024–2026 enterprise AI wave is not an argument against adopting AI — it is a clean, recent instance of consensus behavior masquerading as evidence. The tell was structural and visible in real time: the consensus was a chain of *announcements* rather than a convergence of independent, ROI-validated results; it was amplified by vendors and investor-relations incentives; and the dissenting signal (analyst abandonment warnings, unmeasured returns) was public for anyone who looked. Running the Asch counterfactual — *would we buy this at this scale if no competitor had moved?* — is what separates disciplined AI adoption from FOMO procurement. The rule operates below introspection, so the counterfactual, not the self-report, is the test.\n\n*Sources: Gartner press release, \"Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025\" (July 29, 2024), gartner.com; McKinsey, \"The State of AI\" global survey reports (2024–2025), mckinsey.com; Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992), \"A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades,\" Journal of Political Economy 100(5), 992–1026; Asch, S. E. (1955), \"Opinions and Social Pressure,\" Scientific American 193(5), 31–35.*\n\nFile v1.0.5:examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md\n\n# Method in Action: Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nThe empirical foundation of modern social-proof theory rests on a series of experiments conducted by **Solomon Asch** at Swarthmore College between 1951 and 1956. The setup was almost embarrassingly simple. Yet the result it produced has appeared in every introductory social-psychology textbook for the past seventy years, and was selected by the American Psychological Association as one of the twenty most important studies of the 20th century.\n\nAsch recruited male college students under the cover story that he was running a \"visual perception experiment.\" Each session brought seven or nine people into a room and seated them around a table. The participants were shown a pair of cards, one displaying a \"standard\" vertical line, the other displaying three \"comparison\" vertical lines labeled A, B, and C. One of the comparison lines was the same length as the standard; the other two differed visibly. The participants' task was to say aloud, in the order they were seated, which comparison line matched the standard.\n\nIn the control condition — subjects making the judgment alone — the task was trivially easy. Across baseline trials, error rates were under 1%. Anyone with normal vision could see which line matched.\n\nWhat the subjects did not know was that, in the experimental condition, every person in the room except one was a **confederate**. The confederates had been instructed in advance to give the same wrong answer on critical trials. The single real subject was seated near the end of the line, so that he heard the others' answers before giving his own. There were 18 trials total; on 12 of them (\"critical trials\"), the confederates gave a unanimous wrong answer. On the other 6 trials, they answered correctly, so the situation would not feel obviously rigged.\n\nThe question Asch wanted to answer was: given a perceptual judgment so easy that error rates are under 1% in isolation, **how often will a subject publicly endorse a clearly wrong answer simply because everyone else in the room has done so?**\n\nThe empirical result, summarized in Asch's 1955 *Scientific American* article and his 1956 monograph in *Psychological Monographs*, was striking:\n\n> \"Whereas in ordinary circumstances individuals matching the lines will make mistakes less than 1 per cent of the time, under group pressure the minority subjects swung to acceptance of the misleading majority's wrong judgments in 36.8 per cent of the selections.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 33.\n\nAcross all critical trials, naïve subjects gave the wrong answer matching the group's wrong answer approximately one third of the time. Looking at individuals rather than trials, **about 75% of subjects conformed at least once during the experiment, and about 25% never conformed**. Subjects varied: some conformed on nearly every trial, some on a few, some never. But the modal subject was not the independent thinker; the modal subject conformed at least sometimes.\n\nAsch was particularly careful to follow up. In post-experiment interviews, he asked subjects to explain what had happened. The interviews were the part of the work that elevated the experiment from \"interesting curiosity\" to \"foundational finding,\" because they distinguished what people *said* publicly from what they *believed* privately. Three distinct patterns emerged:\n\n> \"Among the independent subjects there were several who showed great steadfastness of purpose and complete confidence in their own judgments. They were able to dismiss the disagreement and remain firmly anchored to their own perceptions... Among the yielding subjects, three categories emerged. (1) *Distortion of perception*. A very few subjects came to perceive the majority estimates as correct. (2) *Distortion of judgment*. Most of the subjects who went along with the majority concluded that their own perceptions were inaccurate and that the majority was right. (3) *Distortion of action*. The third and largest group of yielding subjects did not suffer a modification of perception nor did they conclude that they were wrong. They yielded because of an overmastering need not to appear different from or inferior to others, because they did not want to expose themselves to the scorn of the majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), pp. 34–35.\n\nThis third category — **\"distortion of action\"** — is the central finding for modern social proof analysis. The largest group of conforming subjects **knew the right answer**, **privately maintained that the majority was wrong**, **and still publicly said the majority's answer aloud** rather than be the lone dissenter. They were not deceived. They were not persuaded. They were not even confused. They simply could not bear the social cost of audibly disagreeing.\n\nAsch ran several variants to identify which features of the setup drove the effect. Three findings from the variants are operationally critical.\n\n**First**, group size matters, but not in a linear way. With one or two confederates, the conformity effect was small. With three confederates, it reached most of its strength. Adding more confederates beyond three produced only marginal increases. Asch wrote:\n\n> \"The opposition of two persons produces 12.8 per cent of errors; with three opponents this becomes 31.8 per cent, but with larger opposition the percentage of errors remains substantially the same... The size of the opposition is important only up to a point.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThe practical implication: it does not take a large crowd to produce the rule. Three same-direction signals are nearly as powerful as fifteen. This is why three well-placed customer testimonials, or three \"everyone is doing it\" references in a sales conversation, can be as effective as much larger consensus signals.\n\n**Second, and most important for defense**, the introduction of a single dissenter — even one — dramatically broke the conformity effect. Asch ran a condition where one of the confederates was instructed to give the correct answer, breaking the unanimity. The naïve subjects' conformity rate dropped from roughly 36% to about 5%:\n\n> \"The presence of a supporting partner depleted the majority of much of its power. Its pressure on the dissenting individual was reduced to one fourth: that is, subjects answered incorrectly only one fourth as often as under the pressure of a unanimous majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThis is the most operationally useful finding in the entire literature. **A single visible dissenter — one person willing to say \"no, the right answer is the one you can plainly see\" — destroys most of the conformity pressure**. Not because the dissenter is persuasive on the merits, but because they make the public act of disagreement *survivable*. In any decision setting where you fear social proof is producing manufactured consensus, the single most effective intervention is to **find and amplify the dissenter**, or to *be* the dissenter when needed. The social cost of being one-of-two is qualitatively different from the social cost of being one-of-one.\n\n**Third**, when subjects were allowed to write their answers privately rather than say them aloud, conformity dropped substantially. The effect is largely a *public* phenomenon: the rule operates most strongly when others can witness your answer. This explains why anonymous surveys frequently produce different results from public polls or focus groups, and why decisions that are publicly debated produce different outcomes than decisions made privately and then aggregated.\n\nThe Asch experiments matter for this skill in five specific ways.\n\n**First**, **the rule does not require deception, ambiguity, or weakness**. Asch's task was visually obvious. The subjects were intelligent young men at an elite college. The confederates had no power over them, no formal authority, no claim to expertise. The room had no high stakes. Yet the rule fired in a third of trials. **If social proof can produce conformity in conditions that benign, it can produce it in your business and personal decision-making — particularly when stakes, ambiguity, and reputational costs are all higher.**\n\n**Second**, **the rule fires in the public domain, not the private one**. Most subjects conformed *publicly* while *privately* maintaining the correct answer. This is the core mechanism behind \"the emperor has no clothes\" dynamics: situations where almost everyone privately doubts the conventional wisdom but no one says so publicly because everyone else seems to be on board. The cost of public dissent is real and is structurally what the rule exploits. *Bubbles, fads, doomed business strategies that everyone privately doubts, and political consensuses that quietly invert overnight all operate on this mechanism.*\n\n**Third**, **a single dissenter is enough to break the conformity field**. This is the operational core of the defense. When you suspect a decision is being made under social-proof pressure rather than evidence, the act of being or finding the first dissenter is disproportionately effective. You do not need to overturn the consensus alone; you need to make it possible for others — who already privately agree with you — to say so.\n\n**Fourth**, **the rule is amplified by similarity, uncertainty, and visibility**. Cialdini's framework derives directly from Asch and Sherif: ambiguity (Sherif) and similarity (Cialdini) amplify the effect, and visibility (Asch's public-vs-private finding) is what makes it operate at all. When you deploy social proof in marketing or product design, these are the levers; when you defend against it in your decision-making, these are the warning signs that the rule is likely firing.\n\n**Fifth**, and most critical: **the rule operates below introspection**. Asch's subjects could not, in real time, tell that the group's wrong answer was making them say a wrong answer. The interviews revealed the mechanism only afterward. **In real time, the rule presents as your own judgment, not as external pressure.** This is why social-proof analysis cannot be done by introspection alone — it requires structured exercises (the Asch counterfactual: \"what would I have concluded with only the evidence and no awareness of others?\") to expose what introspection cannot.\n\nThe Asch experiments did not invent the observation that humans conform; folk psychology has known that for millennia. What they did was provide the precise empirical claim that the rule fires *even when consensus is visibly wrong*, that it operates *predominantly in public*, and that *one dissenter is sufficient to break it*. Every modern application of social proof — in marketing, in growth, in product design, in political organizing, in defense against manipulation — ultimately rests on those three findings. They were established in a small experimental room in Pennsylvania in the early 1950s, and they have not been overturned.\n\nFile v1.0.5:skill-card.md\n\n## Description:\n\nHelps an agent analyze when social proof, popularity claims, testimonials, trend pressure, fake reviews, manufactured engagement, or FOMO are influencing a decision or design.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[deciqai](https://clawhub.ai/user/deciqai)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nEmployees, external users, and developers use this skill to evaluate whether consensus signals are reliable evidence or manufactured pressure. It supports purchasing, hiring, investment, marketing, sales, and product UX decisions involving testimonials, popularity claims, trends, or FOMO.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill can influence how an agent discusses social proof, testimonials, fake consensus, and FOMO-driven decisions.\n\nMitigation: Use it as a reasoning aid and require independent evidence before relying on consensus signals for consequential business decisions.\n\nRisk: Examples and cited source claims may become outdated when applied to current business decisions.\n\nMitigation: Independently verify factual claims in the examples and sources before using them as current decision evidence.\n\n## Reference(s):\n\n- [ClawHub Skill Page](https://clawhub.ai/deciqai/skills/social-proof)\n- [Social Proof Sources](references/sources.md)\n- [Solomon Asch Conformity Example](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md)\n- [Enterprise AI Copilot FOMO Procurement Example](examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md)\n- [Social Proof Machine-Readable Metadata](https://www.deciqai.com/s/social-proof.json)\n- [Social Proof Skill Page](https://www.deciqai.com/c/social-proof)\n- [Asch 1955 Scientific American DOI](https://doi.org/10.1038/scientificamerican1155-31)\n- [Asch 1956 Psychological Monographs DOI](https://doi.org/10.1037/h0093718)\n- [Bikhchandani, Hirshleifer, and Welch 1992 DOI](https://doi.org/10.1086/261849)\n- [Salganik, Dodds, and Watts 2006 DOI](https://doi.org/10.1126/science.1121066)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, guidance]\n\n**Output Format:** [Markdown analysis template and coaching prompts]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Produces structured consensus analysis, decision guidance, and early-warning triggers; no files, code, shell commands, or configuration are produced.]\n\n## Skill Version(s):\n\n1.0.5 (source: release evidence)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.\n\nArchive v1.0.4: 6 files, 17940 bytes\n\nFiles: examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md (9328b), examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md (11318b), references/sources.md (3748b), skill-card.md (2855b), SKILL.md (10035b), _meta.json (131b)\n\nFile v1.0.4:SKILL.md\n\n---\nname: social-proof\ndescription: \"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or consensus is real, suspects fake reviews or manufactured engagement, or is designing testimonials/UX to convert users. Do NOT activate when: the decision is low-stakes and reversible (choosing a lunch spot) or the user already has direct measured evidence stronger than any consensus signal.\"\n---\n\n# Social Proof\n\n## Overview\n\n**Social proof**: we judge what is correct, normal, or worth doing by observing what others — especially similar others — are doing. Usually efficient; failure mode is severe: under unanimous consensus, people publicly endorse answers they privately know are wrong (Asch 1951–56: error rate <1% alone, ~37% under group pressure). Two amplifiers: **uncertainty** (social proof fills the vacuum) and **similarity** (same-type peers drive far stronger conformity than generic crowds).\n\nComposes with `reciprocity` (Cialdini's two primary levers), `anchoring` (price tiers often function as quasi-social-proof), and `critical-thinking` (structured fallback when consensus has been engineered).\n\n## When to Use\n\n**Use when:** purchase/hiring/investment decision leaning on what others chose; proposal cites \"everyone is doing this\"; designing growth/marketing/UX with social-proof patterns; decision feels unsafe alone without a clear reason; suspecting manufactured consensus (bots, paid reviews, astroturf); a trend is accelerating and private doubt is being suppressed by the fact everyone is on board; a \"we must adopt AI because every competitor is deploying it\" mandate is driving procurement or a pilot ahead of any validated ROI (AI hype / FOMO buying).\n\n**Do NOT use when:** decision is low-stakes and reversible; you have direct measured evidence stronger than any consensus; the \"consensus\" is from verified domain experts with better epistemic position; you want to rationalize a contrarian position that lacks independent evidence.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has a concrete case → run The Process directly.\n- **Coach mode:** user is unfamiliar or has no concrete case → guide, don't lecture.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** We judge what's correct by looking at what others do — useful most of the time, but under enough unanimous consensus, people will publicly agree with answers they privately know are wrong, even on obvious questions.\n2. **Check fit** against When to Use / When NOT to use. If direct evidence is stronger, point there.\n3. **Elicit the real situation.** A concrete decision shaped by what others are doing, or a design problem deploying social proof. Never run on hypotheticals.\n> **[WAIT — do not advance until user responds]**\n4. **One element at a time.** Walk through: what's the consensus, who are the consensus-makers, are they similar to you / informed, would you decide the same way if alone — wait for input.\n> **[WAIT — do not advance until user responds]**\n5. **Close by naming the payoff.** The one move — accept the consensus, reject it, or seek independent evidence — that fits their situation.\n> **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Social-Proof Analysis**. Diagnose the source of consensus, then decide whether to use it as evidence.\n\n1. **Name the consensus precisely.** Not \"everyone uses Salesforce\" but \"three cohort companies I respect use Salesforce.\" Vague consensus cannot be analyzed.\n2. **Identify consensus-makers.** Who exactly, how many, how similar to you in ways relevant to the decision?\n3. **Classify consensus type.** Informational (converged on evidence) | Social (converged because others did — cascade risk) | Manufactured (engineered appearance via bots, paid reviews, cherry-picked cases).\n4. **Test signal strength.** Did consensus form independently or in chain? What are dissenters saying? What is the base rate for consensus being wrong in this domain?\n5. **Run the Asch counterfactual.** Alone, with only the underlying evidence, would you reach the same conclusion? If no — you've been pulled in by the rule itself.\n6. **Check manufactured-consensus signs:** astroturf, survivorship bias, selected testimonials, engagement-metric inflation.\n7. **As a sender:** real named testimonials, third-party reviews, transparent distributions, limitations in case studies.\n8. **Stop-rule:** if you cannot defend the decision independently of \"many others are doing it,\" treat it as provisional. Plan a fallback.\n\n### Output template\n\n```\nConsensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]\n```\n\n*→ Method in Action: [Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md)*\n\n*→ 2026 lens: [Enterprise AI Copilot FOMO Procurement (2024–2026)](examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md)*\n\n## Pack: Social Proof in Practice\n\n- **Sender (marketing/sales):** named logos, real attributed testimonials, third-party reviews, transparent star distributions, case studies with limitations stated. Reference customers similar to the prospect.\n- **Receiver (evaluation):** check distribution shapes not averages; distinguish trial users from paying customers; named accounts are top-decile success cases — ask about failures.\n- **Product UX:** \"popular choice\" defaults are powerful — notice when a default is doing your decision-making for you.\n- **Astroturf defense:** anomalous account creation timing, repeated language across \"independent\" voices, engagement metrics inconsistent with audience size → drop consensus signal to near-zero.\n\n## Applying It Well\n\n- Similarity is load-bearing: \"other founders chose this\" works on a founder; \"many people chose this\" does not. Always identify whether consensus-makers are similar to you in relevant ways.\n- Three well-placed testimonials approach the power of fifteen — the rule saturates at small numbers.\n- A single visible dissenter destroys most conformity pressure. Find the dissenter or be the dissenter.\n- The rule operates below introspection. The Asch counterfactual is the test, not the self-report.\n- Manufactured social proof has a reputational cliff when discovered. Real social proof compounds.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] \"Everyone is doing it, so it must be right\" | Asch showed this fails on tasks where the right answer was visually obvious. \"Everyone\" is one signal to weigh — not the conclusion. |\n| [D] \"Many smart people are doing X, so X is right\" | Smart people are more invested in being seen as smart, raising the cost of public dissent. The \"smart people\" filter does not defend against engineered consensus. |\n| [D] \"I have my own opinion regardless of what others do\" | Asch's 75% applies even to people who predicted this of themselves. The rule operates below introspection — the counterfactual is the test. |\n| [D] \"Bestseller / most-popular must be the best\" | Popularity reflects discoverability and marketing, not necessarily quality. Treat it as a prior, then update on evidence. |\n| [D] \"If it were wrong, more people would have noticed\" | Public dissent is rare even when private dissent is widespread (Theranos, FTX, 2008 housing market). |\n| [D] \"I noticed the manufactured social proof, so I'm immune\" | Recognition reduces but does not eliminate the pull. Treat recognition as the start of defense, not the conclusion. |\n| [D] Confusing aggregated wisdom with social proof | Markets aggregate information. Social proof aggregates behavior — which may not reflect information. Distinguish \"independent estimators converged\" from \"people followed early movers.\" |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Decision driven by \"many others are doing it\" with no underlying analysis\n- Consensus-makers cannot be verified as similar to you in relevant ways\n- No dissenters visible in a domain where you would expect dissent\n- Consensus measured in low-cost actions (likes, sign-ups) not high-cost ones (purchases, repeat usage)\n- Star distributions are suspiciously skewed (all 5-stars or U-shaped)\n- Social-proof claim growing faster than the underlying user/evidence base could plausibly support\n\n## Verification\n\n- [ ] Consensus claim named precisely (specific group, not generic mass)\n- [ ] Consensus-makers identified — number, similarity, base of information\n- [ ] Consensus classified: informational / social / manufactured\n- [ ] Asch counterfactual performed: same conclusion alone?\n- [ ] Dissenting voices sought; their reasoning evaluated\n- [ ] Manufactured-consensus signs checked\n- [ ] If following: early-warning indicator specified\n- [ ] If sending: proof is real, named, verifiable, representative\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 223 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/c/social-proof** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\nFile v1.0.4:_meta.json\n\n{\n  \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n  \"slug\": \"social-proof\",\n  \"version\": \"1.0.4\",\n  \"publishedAt\": 1783679289728\n}\n\nFile v1.0.4:references/sources.md\n\n# Sources — social-proof\n\n> *Primary sources for the [social-proof](../SKILL.md) skill.*\n\n- Asch, S. E. (1951). \"Effects of Group Pressure upon the Modification and Distortion of Judgments.\" In H. Guetzkow (Ed.), *Groups, Leadership, and Men: Research in Human Relations* (pp. 177–190). Pittsburgh: Carnegie Press. The original publication of the line-matching conformity experiments at Swarthmore.\n- Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), 31–35. The most accessible account, written by Asch for a general audience; primary source for the verbatim quotations on the 36.8% conformity rate, the categories of conformity (distortion of perception/judgment/action), the group-size threshold, and the dissenter effect. https://doi.org/10.1038/scientificamerican1155-31\n- Asch, S. E. (1956). \"Studies of Independence and Conformity: I. A Minority of One Against a Unanimous Majority.\" *Psychological Monographs: General and Applied*, 70(9), 1–70. The full monograph documentation. https://doi.org/10.1037/h0093718\n- Sherif, M. (1935). \"A Study of Some Social Factors in Perception.\" *Archives of Psychology*, 27(187), 1–60. The autokinetic-effect experiments; the canonical demonstration that under ambiguity, group norms substitute for objective information. The historical precursor to the Asch line.\n- Milgram, S., Bickman, L., & Berkowitz, L. (1969). \"Note on the Drawing Power of Crowds of Different Size.\" *Journal of Personality and Social Psychology*, 13(2), 79–82. The New York City sidewalk field experiment: one confederate looking up at a building draws ~4% of passers-by, fifteen confederates draw ~86%. https://doi.org/10.1037/h0028070\n- Cialdini, R. B. (1984; 2007 rev.). *Influence: The Psychology of Persuasion*. Harper Business. Chapter 4 is the canonical applied treatment of social proof, including the uncertainty and similarity amplifiers. ISBN 978-0061241895.\n- Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). \"A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades.\" *Journal of Political Economy*, 100(5), 992–1026. The formal model of information cascades — how rational agents looking at the choices of earlier agents can produce socially-proof-style mass behavior even when each individual's private information would have led to a different conclusion. https://doi.org/10.1086/261849\n- Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). \"Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market.\" *Science*, 311(5762), 854–856. The Music Lab experiment: 14,000+ subjects shown otherwise-identical songs converged on radically different \"winners\" depending on the social-proof signals they were shown — empirical demonstration that the consensus a market produces is often determined by social-proof dynamics rather than underlying quality. https://doi.org/10.1126/science.1121066\n- Gartner (2024). \"Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025.\" Press release, July 29, 2024. Contemporary evidence that enterprise generative-AI adoption in 2024–2025 outran validated ROI — cited factors include poor data quality, inadequate risk controls, escalating costs, and unclear business value. Illustrates social-proof / FOMO-driven procurement in a live hype cycle. https://www.gartner.com/en/newsroom\n- McKinsey & Company (2024–2025). \"The State of AI\" (annual global survey). Documents the widening gap between rapid generative-AI adoption and the share of organizations able to demonstrate material financial return — the empirical backdrop for consensus-driven (\"everyone is deploying copilots\") AI buying. https://www.mckinsey.com\n\nFile v1.0.4:examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md\n\n# Method in Action: Enterprise AI Copilot FOMO Procurement (2024–2026)\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nBetween the launch of ChatGPT in late 2022 and 2026, generative AI moved from novelty to boardroom mandate. By 2024–2025 a recognizable pattern had set in across large organizations: an executive would say some version of *\"every competitor is deploying AI copilots — we can't be the ones left behind,\"* and a procurement cycle would begin. The dominant justification for buying was not a measured business case; it was the observation that **everyone else was buying**. This is a textbook social-proof failure mode — consensus behavior substituting for independent evidence — and it is worth walking through this skill's own Process, because the same pattern recurs in every hype cycle.\n\nThe pattern was widely documented. Industry surveys and analyst commentary through 2024–2025 repeatedly noted a gap between adoption rates and realized returns: adoption of generative AI rose sharply, while a large share of organizations reported difficulty demonstrating financial return from their AI initiatives. A frequently discussed data point was a 2024 Gartner prediction that a substantial share of generative-AI projects would be abandoned after proof-of-concept by the end of 2025, citing poor data quality, unclear business value, escalating costs, and inadequate risk controls. The *phenomenon* — FOMO-driven procurement outrunning validated ROI — was well established before this analysis. Let us run it through the eight Process steps.\n\n## 1. Name the consensus precisely\n\nThe default framing — *\"everyone is deploying AI copilots\"* — is exactly the vague consensus the skill warns against. It cannot be analyzed. Sharpen it:\n\n- Not \"everyone is deploying copilots\" but: *\"three named direct competitors announced Microsoft 365 Copilot or a comparable assistant rollout in press releases and earnings calls, and our board saw those announcements.\"*\n- Distinguish an **announced pilot** (\"we are exploring generative AI\") from a **validated, ROI-positive production deployment**. Public announcements are overwhelmingly the former. The consensus you can actually verify is a consensus of *announcements*, not a consensus of *proven results*.\n\nOnce named precisely, most of the perceived consensus evaporates: what \"everyone\" has done is issue a press release and stand up a pilot — not demonstrate return.\n\n## 2. Identify consensus-makers\n\nWho exactly is the consensus, how many, and how similar to you on dimensions relevant to *whether the tool will pay off*?\n\n- **Vendors and their platform partners** have direct commercial incentive to project a \"everyone is adopting\" narrative.\n- **Competitors' investor-relations announcements** are similar to you in industry but are optimizing for a stock-price and talent-signaling audience, not for disclosing pilot failure rates.\n- **Analysts and the trade press** amplify the loud adopters; the companies that quietly ran a pilot and shelved it rarely issue a press release.\n\nRelevant similarity is thin. The consensus-makers share your *industry* but not your *data readiness, workflow, or the honest internal ROI numbers* — the variables that actually determine whether a copilot pays off.\n\n## 3. Classify consensus type\n\n- **Informational** (converged on evidence): weak. Very few adopters had published rigorous, independent ROI data by 2024–2025; many were still in pilot.\n- **Social** (converged because others did — cascade risk): dominant. Each competitor's visible move raised the perceived cost of *not* moving, independent of any new evidence — the exact structure of an informational cascade (Bikhchandani, Hirshleifer & Welch, 1992).\n- **Manufactured** (engineered appearance): partial. Vendor-sponsored case studies, curated success stories, and \"N% of enterprises are adopting AI\" marketing statistics inflate the apparent consensus.\n\nVerdict: **primarily social with a manufactured overlay, thin on informational content.**\n\n## 4. Test signal strength\n\nDid the consensus form independently or in a chain? Almost entirely a chain: Company A's announcement fed Company B's board deck fed Company C's mandate. This is cascade behavior, not independent estimators converging.\n\nWhat are dissenters saying? The dissenting evidence was public and loud by 2024–2025 — analyst warnings of high proof-of-concept abandonment, surveys showing most firms could not yet quantify financial return, and widely reported early-mover disappointments. The base rate for consensus being wrong in enterprise technology hype cycles is high (compare the earlier big-data and blockchain waves, where adoption announcements far outran realized ROI).\n\n## 5. Run the Asch counterfactual\n\nThe core test: *alone, with only the underlying evidence and no awareness of what competitors were doing, would we buy this now, at this price, at this scale?*\n\nFor most FOMO buyers the honest answer is **no** — absent the competitive signal, they would run a small, instrumented pilot with a defined success metric before a broad rollout. If the decision flips the moment you remove the \"everyone is doing it\" signal, you have been pulled in by the rule itself, not by evidence. This is the Asch \"distortion of action\" pattern at organizational scale: decision-makers who privately doubt the ROI still publicly back the purchase rather than be the executive who \"fell behind on AI.\"\n\n## 6. Check manufactured-consensus signs\n\n- **Astroturf / curated cases:** vendor case studies are top-decile successes; ask for the failure distribution and the denominator, not the showcase logo.\n- **Survivorship bias:** the shelved pilots don't issue press releases, so the visible population is biased toward apparent winners.\n- **Metric inflation:** \"adoption\" is often measured by seats provisioned or logins (low-cost actions), not by sustained usage, workflow integration, or booked savings (high-cost actions). Seats purchased is not value realized.\n\n## 7. As a sender\n\nIf *you* are marketing an AI tool and want durable rather than fragile social proof: use real named reference customers similar to the prospect, disclose the *distribution* of outcomes (including where it underperformed), and quantify ROI with stated limitations rather than a single hero anecdote. Manufactured AI hype has a reputational cliff when the abandonment numbers surface; honest, representative proof compounds as the market matures past the peak of the cycle.\n\n## 8. Stop-rule\n\nIf you cannot defend the AI purchase independently of \"our competitors are all doing it,\" treat it as **provisional**. The disciplined move is not \"don't adopt AI\" — the technology is real and the upside can be large. The disciplined move is to **replace the cascade with your own evidence**: a bounded pilot, a pre-committed success metric (a specific productivity or cost KPI on a specific workflow), and a fallback plan if the pilot misses.\n\n## Output template, filled\n\n```\nConsensus claim: \"3 named competitors announced copilot rollouts\" (announcements, not proven ROI); consensus-makers similar in industry, not in data-readiness or disclosed results\nConsensus type: primarily social (cascade) + manufactured overlay; thin informational content\nIndependent vs. chain: chain — each competitor's move fed the next; high cascade risk\nDissenters: analysts warning of high proof-of-concept abandonment; surveys showing most firms cannot yet quantify AI financial return\nAsch counterfactual: alone with only evidence, most would run a metered pilot first, not a broad rollout → no\nManufactured-consensus check: curated vendor case studies, survivorship (shelved pilots stay silent), adoption measured by seats not sustained value\nDecision: seek independent evidence — bounded pilot with a pre-committed ROI metric before scaling\nEarly-warning trigger: pilot misses its pre-set KPI; usage decays after novelty; per-seat cost rises without measured workflow savings\n```\n\n## What this case teaches\n\nThe 2024–2026 enterprise AI wave is not an argument against adopting AI — it is a clean, recent instance of consensus behavior masquerading as evidence. The tell was structural and visible in real time: the consensus was a chain of *announcements* rather than a convergence of independent, ROI-validated results; it was amplified by vendors and investor-relations incentives; and the dissenting signal (analyst abandonment warnings, unmeasured returns) was public for anyone who looked. Running the Asch counterfactual — *would we buy this at this scale if no competitor had moved?* — is what separates disciplined AI adoption from FOMO procurement. The rule operates below introspection, so the counterfactual, not the self-report, is the test.\n\n*Sources: Gartner press release, \"Gartner Predicts 30% of Generative AI Projects Will Be Abandoned After Proof of Concept by End of 2025\" (July 29, 2024), gartner.com; McKinsey, \"The State of AI\" global survey reports (2024–2025), mckinsey.com; Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992), \"A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades,\" Journal of Political Economy 100(5), 992–1026; Asch, S. E. (1955), \"Opinions and Social Pressure,\" Scientific American 193(5), 31–35.*\n\nFile v1.0.4:examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md\n\n# Method in Action: Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nThe empirical foundation of modern social-proof theory rests on a series of experiments conducted by **Solomon Asch** at Swarthmore College between 1951 and 1956. The setup was almost embarrassingly simple. Yet the result it produced has appeared in every introductory social-psychology textbook for the past seventy years, and was selected by the American Psychological Association as one of the twenty most important studies of the 20th century.\n\nAsch recruited male college students under the cover story that he was running a \"visual perception experiment.\" Each session brought seven or nine people into a room and seated them around a table. The participants were shown a pair of cards, one displaying a \"standard\" vertical line, the other displaying three \"comparison\" vertical lines labeled A, B, and C. One of the comparison lines was the same length as the standard; the other two differed visibly. The participants' task was to say aloud, in the order they were seated, which comparison line matched the standard.\n\nIn the control condition — subjects making the judgment alone — the task was trivially easy. Across baseline trials, error rates were under 1%. Anyone with normal vision could see which line matched.\n\nWhat the subjects did not know was that, in the experimental condition, every person in the room except one was a **confederate**. The confederates had been instructed in advance to give the same wrong answer on critical trials. The single real subject was seated near the end of the line, so that he heard the others' answers before giving his own. There were 18 trials total; on 12 of them (\"critical trials\"), the confederates gave a unanimous wrong answer. On the other 6 trials, they answered correctly, so the situation would not feel obviously rigged.\n\nThe question Asch wanted to answer was: given a perceptual judgment so easy that error rates are under 1% in isolation, **how often will a subject publicly endorse a clearly wrong answer simply because everyone else in the room has done so?**\n\nThe empirical result, summarized in Asch's 1955 *Scientific American* article and his 1956 monograph in *Psychological Monographs*, was striking:\n\n> \"Whereas in ordinary circumstances individuals matching the lines will make mistakes less than 1 per cent of the time, under group pressure the minority subjects swung to acceptance of the misleading majority's wrong judgments in 36.8 per cent of the selections.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 33.\n\nAcross all critical trials, naïve subjects gave the wrong answer matching the group's wrong answer approximately one third of the time. Looking at individuals rather than trials, **about 75% of subjects conformed at least once during the experiment, and about 25% never conformed**. Subjects varied: some conformed on nearly every trial, some on a few, some never. But the modal subject was not the independent thinker; the modal subject conformed at least sometimes.\n\nAsch was particularly careful to follow up. In post-experiment interviews, he asked subjects to explain what had happened. The interviews were the part of the work that elevated the experiment from \"interesting curiosity\" to \"foundational finding,\" because they distinguished what people *said* publicly from what they *believed* privately. Three distinct patterns emerged:\n\n> \"Among the independent subjects there were several who showed great steadfastness of purpose and complete confidence in their own judgments. They were able to dismiss the disagreement and remain firmly anchored to their own perceptions... Among the yielding subjects, three categories emerged. (1) *Distortion of perception*. A very few subjects came to perceive the majority estimates as correct. (2) *Distortion of judgment*. Most of the subjects who went along with the majority concluded that their own perceptions were inaccurate and that the majority was right. (3) *Distortion of action*. The third and largest group of yielding subjects did not suffer a modification of perception nor did they conclude that they were wrong. They yielded because of an overmastering need not to appear different from or inferior to others, because they did not want to expose themselves to the scorn of the majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), pp. 34–35.\n\nThis third category — **\"distortion of action\"** — is the central finding for modern social proof analysis. The largest group of conforming subjects **knew the right answer**, **privately maintained that the majority was wrong**, **and still publicly said the majority's answer aloud** rather than be the lone dissenter. They were not deceived. They were not persuaded. They were not even confused. They simply could not bear the social cost of audibly disagreeing.\n\nAsch ran several variants to identify which features of the setup drove the effect. Three findings from the variants are operationally critical.\n\n**First**, group size matters, but not in a linear way. With one or two confederates, the conformity effect was small. With three confederates, it reached most of its strength. Adding more confederates beyond three produced only marginal increases. Asch wrote:\n\n> \"The opposition of two persons produces 12.8 per cent of errors; with three opponents this becomes 31.8 per cent, but with larger opposition the percentage of errors remains substantially the same... The size of the opposition is important only up to a point.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThe practical implication: it does not take a large crowd to produce the rule. Three same-direction signals are nearly as powerful as fifteen. This is why three well-placed customer testimonials, or three \"everyone is doing it\" references in a sales conversation, can be as effective as much larger consensus signals.\n\n**Second, and most important for defense**, the introduction of a single dissenter — even one — dramatically broke the conformity effect. Asch ran a condition where one of the confederates was instructed to give the correct answer, breaking the unanimity. The naïve subjects' conformity rate dropped from roughly 36% to about 5%:\n\n> \"The presence of a supporting partner depleted the majority of much of its power. Its pressure on the dissenting individual was reduced to one fourth: that is, subjects answered incorrectly only one fourth as often as under the pressure of a unanimous majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThis is the most operationally useful finding in the entire literature. **A single visible dissenter — one person willing to say \"no, the right answer is the one you can plainly see\" — destroys most of the conformity pressure**. Not because the dissenter is persuasive on the merits, but because they make the public act of disagreement *survivable*. In any decision setting where you fear social proof is producing manufactured consensus, the single most effective intervention is to **find and amplify the dissenter**, or to *be* the dissenter when needed. The social cost of being one-of-two is qualitatively different from the social cost of being one-of-one.\n\n**Third**, when subjects were allowed to write their answers privately rather than say them aloud, conformity dropped substantially. The effect is largely a *public* phenomenon: the rule operates most strongly when others can witness your answer. This explains why anonymous surveys frequently produce different results from public polls or focus groups, and why decisions that are publicly debated produce different outcomes than decisions made privately and then aggregated.\n\nThe Asch experiments matter for this skill in five specific ways.\n\n**First**, **the rule does not require deception, ambiguity, or weakness**. Asch's task was visually obvious. The subjects were intelligent young men at an elite college. The confederates had no power over them, no formal authority, no claim to expertise. The room had no high stakes. Yet the rule fired in a third of trials. **If social proof can produce conformity in conditions that benign, it can produce it in your business and personal decision-making — particularly when stakes, ambiguity, and reputational costs are all higher.**\n\n**Second**, **the rule fires in the public domain, not the private one**. Most subjects conformed *publicly* while *privately* maintaining the correct answer. This is the core mechanism behind \"the emperor has no clothes\" dynamics: situations where almost everyone privately doubts the conventional wisdom but no one says so publicly because everyone else seems to be on board. The cost of public dissent is real and is structurally what the rule exploits. *Bubbles, fads, doomed business strategies that everyone privately doubts, and political consensuses that quietly invert overnight all operate on this mechanism.*\n\n**Third**, **a single dissenter is enough to break the conformity field**. This is the operational core of the defense. When you suspect a decision is being made under social-proof pressure rather than evidence, the act of being or finding the first dissenter is disproportionately effective. You do not need to overturn the consensus alone; you need to make it possible for others — who already privately agree with you — to say so.\n\n**Fourth**, **the rule is amplified by similarity, uncertainty, and visibility**. Cialdini's framework derives directly from Asch and Sherif: ambiguity (Sherif) and similarity (Cialdini) amplify the effect, and visibility (Asch's public-vs-private finding) is what makes it operate at all. When you deploy social proof in marketing or product design, these are the levers; when you defend against it in your decision-making, these are the warning signs that the rule is likely firing.\n\n**Fifth**, and most critical: **the rule operates below introspection**. Asch's subjects could not, in real time, tell that the group's wrong answer was making them say a wrong answer. The interviews revealed the mechanism only afterward. **In real time, the rule presents as your own judgment, not as external pressure.** This is why social-proof analysis cannot be done by introspection alone — it requires structured exercises (the Asch counterfactual: \"what would I have concluded with only the evidence and no awareness of others?\") to expose what introspection cannot.\n\nThe Asch experiments did not invent the observation that humans conform; folk psychology has known that for millennia. What they did was provide the precise empirical claim that the rule fires *even when consensus is visibly wrong*, that it operates *predominantly in public*, and that *one dissenter is sufficient to break it*. Every modern application of social proof — in marketing, in growth, in product design, in political organizing, in defense against manipulation — ultimately rests on those three findings. They were established in a small experimental room in Pennsylvania in the early 1950s, and they have not been overturned.\n\nFile v1.0.4:skill-card.md\n\n## Description: <br>\nHelps agents analyze social-proof claims, distinguish informational, social, and manufactured consensus, and decide whether to follow, reject, or seek independent evidence. <br>\n\nThis skill is ready for commercial/non-commercial use. <br>\n\n## Publisher: <br>\n[deciqai](https://clawhub.ai/user/deciqai) <br>\n\n### License/Terms of Use: <br>\nMIT-0 <br>\n\n\n## Use Case: <br>\nExternal users, developers, and business operators use this skill when a decision, procurement, marketing message, UX pattern, or trend evaluation depends on what other people appear to be doing. It guides the agent to identify the consensus source, test whether the signal is independent or manufactured, and produce a structured social-proof analysis. <br>\n\n### Deployment Geography for Use: <br>\nGlobal <br>\n\n## Known Risks and Mitigations: <br>\nRisk: Analytical outputs can be inaccurate or misleading if the agent overweights a weak consensus signal. <br>\nMitigation: Review the analysis for accuracy and require independent evidence before relying on conclusions. <br>\nRisk: Users may paste confidential decision context into the active agent session. <br>\nMitigation: Avoid sharing confidential material unless it is appropriate for the active agent environment. <br>\nRisk: The skill can be used on consequential purchasing, hiring, investment, or UX decisions. <br>\nMitigation: Use the skill's stop rule: treat decisions as provisional when they cannot be defended independently of what others are doing. <br>\n\n\n## Reference(s): <br>\n- [ClawHub listing](https://clawhub.ai/deciqai/skills/social-proof) <br>\n- [Primary sources](references/sources.md) <br>\n- [Solomon Asch conformity example](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md) <br>\n- [Enterprise AI copilot FOMO example](examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md) <br>\n- [Asch 1955, Opinions and Social Pressure](https://doi.org/10.1038/scientificamerican1155-31) <br>\n- [Bikhchandani, Hirshleifer, and Welch 1992, Information Cascades](https://doi.org/10.1086/261849) <br>\n- [Salganik, Dodds, and Watts 2006, Music Lab](https://doi.org/10.1126/science.1121066) <br>\n\n\n## Skill Output: <br>\n**Output Type(s):** [text, markdown, guidance] <br>\n**Output Format:** [Markdown analysis with a structured template and short coaching questions when needed] <br>\n**Output Parameters:** [1D] <br>\n**Other Properties Related to Output:** [May include WAIT stops in coach mode; does not execute tools or code.] <br>\n\n## Skill Version(s): <br>\n1.0.4 (source: server release evidence) <br>\n\n## Ethical Considerations: <br>\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment. <br>\n\nArchive v1.0.3: 5 files, 12855 bytes\n\nFiles: examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md (11318b), references/sources.md (2933b), skill-card.md (2775b), SKILL.md (9744b), _meta.json (131b)\n\nFile v1.0.3:SKILL.md\n\n---\nname: social-proof\ndescription: \"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or consensus is real, suspects fake reviews or manufactured engagement, or is designing testimonials/UX to convert users. Do NOT activate when: the decision is low-stakes and reversible (choosing a lunch spot) or the user already has direct measured evidence stronger than any consensus signal.\"\n---\n\n# Social Proof\n\n## Overview\n\n**Social proof**: we judge what is correct, normal, or worth doing by observing what others — especially similar others — are doing. Usually efficient; failure mode is severe: under unanimous consensus, people publicly endorse answers they privately know are wrong (Asch 1951–56: error rate <1% alone, ~37% under group pressure). Two amplifiers: **uncertainty** (social proof fills the vacuum) and **similarity** (same-type peers drive far stronger conformity than generic crowds).\n\nComposes with `reciprocity` (Cialdini's two primary levers), `anchoring` (price tiers often function as quasi-social-proof), and `critical-thinking` (structured fallback when consensus has been engineered).\n\n## When to Use\n\n**Use when:** purchase/hiring/investment decision leaning on what others chose; proposal cites \"everyone is doing this\"; designing growth/marketing/UX with social-proof patterns; decision feels unsafe alone without a clear reason; suspecting manufactured consensus (bots, paid reviews, astroturf); a trend is accelerating and private doubt is being suppressed by the fact everyone is on board.\n\n**Do NOT use when:** decision is low-stakes and reversible; you have direct measured evidence stronger than any consensus; the \"consensus\" is from verified domain experts with better epistemic position; you want to rationalize a contrarian position that lacks independent evidence.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has a concrete case → run The Process directly.\n- **Coach mode:** user is unfamiliar or has no concrete case → guide, don't lecture.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** We judge what's correct by looking at what others do — useful most of the time, but under enough unanimous consensus, people will publicly agree with answers they privately know are wrong, even on obvious questions.\n2. **Check fit** against When to Use / When NOT to use. If direct evidence is stronger, point there.\n3. **Elicit the real situation.** A concrete decision shaped by what others are doing, or a design problem deploying social proof. Never run on hypotheticals.\n> **[WAIT — do not advance until user responds]**\n4. **One element at a time.** Walk through: what's the consensus, who are the consensus-makers, are they similar to you / informed, would you decide the same way if alone — wait for input.\n> **[WAIT — do not advance until user responds]**\n5. **Close by naming the payoff.** The one move — accept the consensus, reject it, or seek independent evidence — that fits their situation.\n> **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Social-Proof Analysis**. Diagnose the source of consensus, then decide whether to use it as evidence.\n\n1. **Name the consensus precisely.** Not \"everyone uses Salesforce\" but \"three cohort companies I respect use Salesforce.\" Vague consensus cannot be analyzed.\n2. **Identify consensus-makers.** Who exactly, how many, how similar to you in ways relevant to the decision?\n3. **Classify consensus type.** Informational (converged on evidence) | Social (converged because others did — cascade risk) | Manufactured (engineered appearance via bots, paid reviews, cherry-picked cases).\n4. **Test signal strength.** Did consensus form independently or in chain? What are dissenters saying? What is the base rate for consensus being wrong in this domain?\n5. **Run the Asch counterfactual.** Alone, with only the underlying evidence, would you reach the same conclusion? If no — you've been pulled in by the rule itself.\n6. **Check manufactured-consensus signs:** astroturf, survivorship bias, selected testimonials, engagement-metric inflation.\n7. **As a sender:** real named testimonials, third-party reviews, transparent distributions, limitations in case studies.\n8. **Stop-rule:** if you cannot defend the decision independently of \"many others are doing it,\" treat it as provisional. Plan a fallback.\n\n### Output template\n\n```\nConsensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]\n```\n\n*→ Method in Action: [Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md)*\n\n## Pack: Social Proof in Practice\n\n- **Sender (marketing/sales):** named logos, real attributed testimonials, third-party reviews, transparent star distributions, case studies with limitations stated. Reference customers similar to the prospect.\n- **Receiver (evaluation):** check distribution shapes not averages; distinguish trial users from paying customers; named accounts are top-decile success cases — ask about failures.\n- **Product UX:** \"popular choice\" defaults are powerful — notice when a default is doing your decision-making for you.\n- **Astroturf defense:** anomalous account creation timing, repeated language across \"independent\" voices, engagement metrics inconsistent with audience size → drop consensus signal to near-zero.\n\n## Applying It Well\n\n- Similarity is load-bearing: \"other founders chose this\" works on a founder; \"many people chose this\" does not. Always identify whether consensus-makers are similar to you in relevant ways.\n- Three well-placed testimonials approach the power of fifteen — the rule saturates at small numbers.\n- A single visible dissenter destroys most conformity pressure. Find the dissenter or be the dissenter.\n- The rule operates below introspection. The Asch counterfactual is the test, not the self-report.\n- Manufactured social proof has a reputational cliff when discovered. Real social proof compounds.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] \"Everyone is doing it, so it must be right\" | Asch showed this fails on tasks where the right answer was visually obvious. \"Everyone\" is one signal to weigh — not the conclusion. |\n| [D] \"Many smart people are doing X, so X is right\" | Smart people are more invested in being seen as smart, raising the cost of public dissent. The \"smart people\" filter does not defend against engineered consensus. |\n| [D] \"I have my own opinion regardless of what others do\" | Asch's 75% applies even to people who predicted this of themselves. The rule operates below introspection — the counterfactual is the test. |\n| [D] \"Bestseller / most-popular must be the best\" | Popularity reflects discoverability and marketing, not necessarily quality. Treat it as a prior, then update on evidence. |\n| [D] \"If it were wrong, more people would have noticed\" | Public dissent is rare even when private dissent is widespread (Theranos, FTX, 2008 housing market). |\n| [D] \"I noticed the manufactured social proof, so I'm immune\" | Recognition reduces but does not eliminate the pull. Treat recognition as the start of defense, not the conclusion. |\n| [D] Confusing aggregated wisdom with social proof | Markets aggregate information. Social proof aggregates behavior — which may not reflect information. Distinguish \"independent estimators converged\" from \"people followed early movers.\" |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Decision driven by \"many others are doing it\" with no underlying analysis\n- Consensus-makers cannot be verified as similar to you in relevant ways\n- No dissenters visible in a domain where you would expect dissent\n- Consensus measured in low-cost actions (likes, sign-ups) not high-cost ones (purchases, repeat usage)\n- Star distributions are suspiciously skewed (all 5-stars or U-shaped)\n- Social-proof claim growing faster than the underlying user/evidence base could plausibly support\n\n## Verification\n\n- [ ] Consensus claim named precisely (specific group, not generic mass)\n- [ ] Consensus-makers identified — number, similarity, base of information\n- [ ] Consensus classified: informational / social / manufactured\n- [ ] Asch counterfactual performed: same conclusion alone?\n- [ ] Dissenting voices sought; their reasoning evaluated\n- [ ] Manufactured-consensus signs checked\n- [ ] If following: early-warning indicator specified\n- [ ] If sending: proof is real, named, verifiable, representative\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 164 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/c/social-proof** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\nFile v1.0.3:_meta.json\n\n{\n  \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n  \"slug\": \"social-proof\",\n  \"version\": \"1.0.3\",\n  \"publishedAt\": 1783509594704\n}\n\nFile v1.0.3:references/sources.md\n\n# Sources — social-proof\n\n> *Primary sources for the [social-proof](../SKILL.md) skill.*\n\n- Asch, S. E. (1951). \"Effects of Group Pressure upon the Modification and Distortion of Judgments.\" In H. Guetzkow (Ed.), *Groups, Leadership, and Men: Research in Human Relations* (pp. 177–190). Pittsburgh: Carnegie Press. The original publication of the line-matching conformity experiments at Swarthmore.\n- Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), 31–35. The most accessible account, written by Asch for a general audience; primary source for the verbatim quotations on the 36.8% conformity rate, the categories of conformity (distortion of perception/judgment/action), the group-size threshold, and the dissenter effect. https://doi.org/10.1038/scientificamerican1155-31\n- Asch, S. E. (1956). \"Studies of Independence and Conformity: I. A Minority of One Against a Unanimous Majority.\" *Psychological Monographs: General and Applied*, 70(9), 1–70. The full monograph documentation. https://doi.org/10.1037/h0093718\n- Sherif, M. (1935). \"A Study of Some Social Factors in Perception.\" *Archives of Psychology*, 27(187), 1–60. The autokinetic-effect experiments; the canonical demonstration that under ambiguity, group norms substitute for objective information. The historical precursor to the Asch line.\n- Milgram, S., Bickman, L., & Berkowitz, L. (1969). \"Note on the Drawing Power of Crowds of Different Size.\" *Journal of Personality and Social Psychology*, 13(2), 79–82. The New York City sidewalk field experiment: one confederate looking up at a building draws ~4% of passers-by, fifteen confederates draw ~86%. https://doi.org/10.1037/h0028070\n- Cialdini, R. B. (1984; 2007 rev.). *Influence: The Psychology of Persuasion*. Harper Business. Chapter 4 is the canonical applied treatment of social proof, including the uncertainty and similarity amplifiers. ISBN 978-0061241895.\n- Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). \"A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades.\" *Journal of Political Economy*, 100(5), 992–1026. The formal model of information cascades — how rational agents looking at the choices of earlier agents can produce socially-proof-style mass behavior even when each individual's private information would have led to a different conclusion. https://doi.org/10.1086/261849\n- Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). \"Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market.\" *Science*, 311(5762), 854–856. The Music Lab experiment: 14,000+ subjects shown otherwise-identical songs converged on radically different \"winners\" depending on the social-proof signals they were shown — empirical demonstration that the consensus a market produces is often determined by social-proof dynamics rather than underlying quality. https://doi.org/10.1126/science.1121066\n\nFile v1.0.3:examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md\n\n# Method in Action: Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nThe empirical foundation of modern social-proof theory rests on a series of experiments conducted by **Solomon Asch** at Swarthmore College between 1951 and 1956. The setup was almost embarrassingly simple. Yet the result it produced has appeared in every introductory social-psychology textbook for the past seventy years, and was selected by the American Psychological Association as one of the twenty most important studies of the 20th century.\n\nAsch recruited male college students under the cover story that he was running a \"visual perception experiment.\" Each session brought seven or nine people into a room and seated them around a table. The participants were shown a pair of cards, one displaying a \"standard\" vertical line, the other displaying three \"comparison\" vertical lines labeled A, B, and C. One of the comparison lines was the same length as the standard; the other two differed visibly. The participants' task was to say aloud, in the order they were seated, which comparison line matched the standard.\n\nIn the control condition — subjects making the judgment alone — the task was trivially easy. Across baseline trials, error rates were under 1%. Anyone with normal vision could see which line matched.\n\nWhat the subjects did not know was that, in the experimental condition, every person in the room except one was a **confederate**. The confederates had been instructed in advance to give the same wrong answer on critical trials. The single real subject was seated near the end of the line, so that he heard the others' answers before giving his own. There were 18 trials total; on 12 of them (\"critical trials\"), the confederates gave a unanimous wrong answer. On the other 6 trials, they answered correctly, so the situation would not feel obviously rigged.\n\nThe question Asch wanted to answer was: given a perceptual judgment so easy that error rates are under 1% in isolation, **how often will a subject publicly endorse a clearly wrong answer simply because everyone else in the room has done so?**\n\nThe empirical result, summarized in Asch's 1955 *Scientific American* article and his 1956 monograph in *Psychological Monographs*, was striking:\n\n> \"Whereas in ordinary circumstances individuals matching the lines will make mistakes less than 1 per cent of the time, under group pressure the minority subjects swung to acceptance of the misleading majority's wrong judgments in 36.8 per cent of the selections.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 33.\n\nAcross all critical trials, naïve subjects gave the wrong answer matching the group's wrong answer approximately one third of the time. Looking at individuals rather than trials, **about 75% of subjects conformed at least once during the experiment, and about 25% never conformed**. Subjects varied: some conformed on nearly every trial, some on a few, some never. But the modal subject was not the independent thinker; the modal subject conformed at least sometimes.\n\nAsch was particularly careful to follow up. In post-experiment interviews, he asked subjects to explain what had happened. The interviews were the part of the work that elevated the experiment from \"interesting curiosity\" to \"foundational finding,\" because they distinguished what people *said* publicly from what they *believed* privately. Three distinct patterns emerged:\n\n> \"Among the independent subjects there were several who showed great steadfastness of purpose and complete confidence in their own judgments. They were able to dismiss the disagreement and remain firmly anchored to their own perceptions... Among the yielding subjects, three categories emerged. (1) *Distortion of perception*. A very few subjects came to perceive the majority estimates as correct. (2) *Distortion of judgment*. Most of the subjects who went along with the majority concluded that their own perceptions were inaccurate and that the majority was right. (3) *Distortion of action*. The third and largest group of yielding subjects did not suffer a modification of perception nor did they conclude that they were wrong. They yielded because of an overmastering need not to appear different from or inferior to others, because they did not want to expose themselves to the scorn of the majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), pp. 34–35.\n\nThis third category — **\"distortion of action\"** — is the central finding for modern social proof analysis. The largest group of conforming subjects **knew the right answer**, **privately maintained that the majority was wrong**, **and still publicly said the majority's answer aloud** rather than be the lone dissenter. They were not deceived. They were not persuaded. They were not even confused. They simply could not bear the social cost of audibly disagreeing.\n\nAsch ran several variants to identify which features of the setup drove the effect. Three findings from the variants are operationally critical.\n\n**First**, group size matters, but not in a linear way. With one or two confederates, the conformity effect was small. With three confederates, it reached most of its strength. Adding more confederates beyond three produced only marginal increases. Asch wrote:\n\n> \"The opposition of two persons produces 12.8 per cent of errors; with three opponents this becomes 31.8 per cent, but with larger opposition the percentage of errors remains substantially the same... The size of the opposition is important only up to a point.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThe practical implication: it does not take a large crowd to produce the rule. Three same-direction signals are nearly as powerful as fifteen. This is why three well-placed customer testimonials, or three \"everyone is doing it\" references in a sales conversation, can be as effective as much larger consensus signals.\n\n**Second, and most important for defense**, the introduction of a single dissenter — even one — dramatically broke the conformity effect. Asch ran a condition where one of the confederates was instructed to give the correct answer, breaking the unanimity. The naïve subjects' conformity rate dropped from roughly 36% to about 5%:\n\n> \"The presence of a supporting partner depleted the majority of much of its power. Its pressure on the dissenting individual was reduced to one fourth: that is, subjects answered incorrectly only one fourth as often as under the pressure of a unanimous majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThis is the most operationally useful finding in the entire literature. **A single visible dissenter — one person willing to say \"no, the right answer is the one you can plainly see\" — destroys most of the conformity pressure**. Not because the dissenter is persuasive on the merits, but because they make the public act of disagreement *survivable*. In any decision setting where you fear social proof is producing manufactured consensus, the single most effective intervention is to **find and amplify the dissenter**, or to *be* the dissenter when needed. The social cost of being one-of-two is qualitatively different from the social cost of being one-of-one.\n\n**Third**, when subjects were allowed to write their answers privately rather than say them aloud, conformity dropped substantially. The effect is largely a *public* phenomenon: the rule operates most strongly when others can witness your answer. This explains why anonymous surveys frequently produce different results from public polls or focus groups, and why decisions that are publicly debated produce different outcomes than decisions made privately and then aggregated.\n\nThe Asch experiments matter for this skill in five specific ways.\n\n**First**, **the rule does not require deception, ambiguity, or weakness**. Asch's task was visually obvious. The subjects were intelligent young men at an elite college. The confederates had no power over them, no formal authority, no claim to expertise. The room had no high stakes. Yet the rule fired in a third of trials. **If social proof can produce conformity in conditions that benign, it can produce it in your business and personal decision-making — particularly when stakes, ambiguity, and reputational costs are all higher.**\n\n**Second**, **the rule fires in the public domain, not the private one**. Most subjects conformed *publicly* while *privately* maintaining the correct answer. This is the core mechanism behind \"the emperor has no clothes\" dynamics: situations where almost everyone privately doubts the conventional wisdom but no one says so publicly because everyone else seems to be on board. The cost of public dissent is real and is structurally what the rule exploits. *Bubbles, fads, doomed business strategies that everyone privately doubts, and political consensuses that quietly invert overnight all operate on this mechanism.*\n\n**Third**, **a single dissenter is enough to break the conformity field**. This is the operational core of the defense. When you suspect a decision is being made under social-proof pressure rather than evidence, the act of being or finding the first dissenter is disproportionately effective. You do not need to overturn the consensus alone; you need to make it possible for others — who already privately agree with you — to say so.\n\n**Fourth**, **the rule is amplified by similarity, uncertainty, and visibility**. Cialdini's framework derives directly from Asch and Sherif: ambiguity (Sherif) and similarity (Cialdini) amplify the effect, and visibility (Asch's public-vs-private finding) is what makes it operate at all. When you deploy social proof in marketing or product design, these are the levers; when you defend against it in your decision-making, these are the warning signs that the rule is likely firing.\n\n**Fifth**, and most critical: **the rule operates below introspection**. Asch's subjects could not, in real time, tell that the group's wrong answer was making them say a wrong answer. The interviews revealed the mechanism only afterward. **In real time, the rule presents as your own judgment, not as external pressure.** This is why social-proof analysis cannot be done by introspection alone — it requires structured exercises (the Asch counterfactual: \"what would I have concluded with only the evidence and no awareness of others?\") to expose what introspection cannot.\n\nThe Asch experiments did not invent the observation that humans conform; folk psychology has known that for millennia. What they did was provide the precise empirical claim that the rule fires *even when consensus is visibly wrong*, that it operates *predominantly in public*, and that *one dissenter is sufficient to break it*. Every modern application of social proof — in marketing, in growth, in product design, in political organizing, in defense against manipulation — ultimately rests on those three findings. They were established in a small experimental room in Pennsylvania in the early 1950s, and they have not been overturned.\n\nFile v1.0.3:skill-card.md\n\n## Description: <br>\nGuides an agent to evaluate whether consensus, popularity, testimonials, reviews, or trend signals are reliable evidence or manufactured social proof. <br>\n\nThis skill is ready for commercial/non-commercial use. <br>\n\n## Publisher: <br>\n[deciqai](https://clawhub.ai/user/deciqai) <br>\n\n### License/Terms of Use: <br>\nMIT-0 <br>\n\n\n## Use Case: <br>\nExternal users, employees, and agents use this skill to analyze decisions or messaging that rely on popularity, peer behavior, testimonials, reviews, trends, or apparent consensus. It helps distinguish informational consensus from social cascades or manufactured proof and produces a practical next step. <br>\n\n### Deployment Geography for Use: <br>\nGlobal <br>\n\n## Known Risks and Mitigations: <br>\nRisk: The skill may be used to critique or design social-proof messaging in business, hiring, investment, or public communication contexts. <br>\nMitigation: Treat outputs as advisory, apply the skill ethically, and verify source claims before relying on the analysis for high-stakes decisions or public messaging. <br>\nRisk: Consensus signals, reviews, testimonials, or engagement metrics may be misleading or manufactured. <br>\nMitigation: Check for named and verifiable sources, representative evidence, dissenting views, suspicious distribution patterns, and independent support before using consensus as evidence. <br>\n\n\n## Reference(s): <br>\n- [Sources - social-proof](references/sources.md) <br>\n- [Solomon Asch's Conformity Experiments, Swarthmore, 1951-1956](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md) <br>\n- [Asch, Opinions and Social Pressure](https://doi.org/10.1038/scientificamerican1155-31) <br>\n- [Asch, Studies of Independence and Conformity](https://doi.org/10.1037/h0093718) <br>\n- [Milgram, Bickman, and Berkowitz, Drawing Power of Crowds](https://doi.org/10.1037/h0028070) <br>\n- [Bikhchandani, Hirshleifer, and Welch, Informational Cascades](https://doi.org/10.1086/261849) <br>\n- [Salganik, Dodds, and Watts, Artificial Cultural Market](https://doi.org/10.1126/science.1121066) <br>\n\n\n## Skill Output: <br>\n**Output Type(s):** [Text, Markdown, Guidance] <br>\n**Output Format:** [Markdown] <br>\n**Output Parameters:** [1D] <br>\n**Other Properties Related to Output:** [May include a structured social-proof analysis template with consensus type, dissenters, counterfactual, decision, and early-warning trigger.] <br>\n\n## Skill Version(s): <br>\n1.0.3 (source: release evidence) <br>\n\n## Ethical Considerations: <br>\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment. <br>\n\nArchive v1.0.2: 5 files, 12956 bytes\n\nFiles: examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md (11318b), references/sources.md (2933b), skill-card.md (2995b), SKILL.md (9846b), _meta.json (131b)\n\nFile v1.0.2:SKILL.md\n\n---\nname: social-proof\ndescription: \"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or consensus is real, suspects fake reviews or manufactured engagement, or is designing testimonials/UX to convert users. Do NOT activate when: the decision is low-stakes and reversible (choosing a lunch spot) or the user already has direct measured evidence stronger than any consensus signal.\"\n---\n\n# Social Proof\n\n## Overview\n\n**Social proof**: we judge what is correct, normal, or worth doing by observing what others — especially similar others — are doing. Usually efficient; failure mode is severe: under unanimous consensus, people publicly endorse answers they privately know are wrong (Asch 1951–56: error rate <1% alone, ~37% under group pressure). Two amplifiers: **uncertainty** (social proof fills the vacuum) and **similarity** (same-type peers drive far stronger conformity than generic crowds).\n\nComposes with `reciprocity` (Cialdini's two primary levers), `anchoring` (price tiers often function as quasi-social-proof), and `critical-thinking` (structured fallback when consensus has been engineered).\n\n## When to Use\n\n**Use when:** purchase/hiring/investment decision leaning on what others chose; proposal cites \"everyone is doing this\"; designing growth/marketing/UX with social-proof patterns; decision feels unsafe alone without a clear reason; suspecting manufactured consensus (bots, paid reviews, astroturf); a trend is accelerating and private doubt is being suppressed by the fact everyone is on board.\n\n**Do NOT use when:** decision is low-stakes and reversible; you have direct measured evidence stronger than any consensus; the \"consensus\" is from verified domain experts with better epistemic position; you want to rationalize a contrarian position that lacks independent evidence.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has a concrete case → run The Process directly.\n- **Coach mode:** user is unfamiliar or has no concrete case → guide, don't lecture.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** We judge what's correct by looking at what others do — useful most of the time, but under enough unanimous consensus, people will publicly agree with answers they privately know are wrong, even on obvious questions.\n2. **Check fit** against When to Use / When NOT to use. If direct evidence is stronger, point there.\n3. **Elicit the real situation.** A concrete decision shaped by what others are doing, or a design problem deploying social proof. Never run on hypotheticals.\n> **[WAIT — do not advance until user responds]**\n4. **One element at a time.** Walk through: what's the consensus, who are the consensus-makers, are they similar to you / informed, would you decide the same way if alone — wait for input.\n> **[WAIT — do not advance until user responds]**\n5. **Close by naming the payoff.** The one move — accept the consensus, reject it, or seek independent evidence — that fits their situation.\n> **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Social-Proof Analysis**. Diagnose the source of consensus, then decide whether to use it as evidence.\n\n1. **Name the consensus precisely.** Not \"everyone uses Salesforce\" but \"three cohort companies I respect use Salesforce.\" Vague consensus cannot be analyzed.\n2. **Identify consensus-makers.** Who exactly, how many, how similar to you in ways relevant to the decision?\n3. **Classify consensus type.** Informational (converged on evidence) | Social (converged because others did — cascade risk) | Manufactured (engineered appearance via bots, paid reviews, cherry-picked cases).\n4. **Test signal strength.** Did consensus form independently or in chain? What are dissenters saying? What is the base rate for consensus being wrong in this domain?\n5. **Run the Asch counterfactual.** Alone, with only the underlying evidence, would you reach the same conclusion? If no — you've been pulled in by the rule itself.\n6. **Check manufactured-consensus signs:** astroturf, survivorship bias, selected testimonials, engagement-metric inflation.\n7. **As a sender:** real named testimonials, third-party reviews, transparent distributions, limitations in case studies.\n8. **Stop-rule:** if you cannot defend the decision independently of \"many others are doing it,\" treat it as provisional. Plan a fallback.\n\n### Output template\n\n```\nConsensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]\n```\n\n*→ Method in Action: [Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md)*\n\n## Pack: Social Proof in Practice\n\n- **Sender (marketing/sales):** named logos, real attributed testimonials, third-party reviews, transparent star distributions, case studies with limitations stated. Reference customers similar to the prospect.\n- **Receiver (evaluation):** check distribution shapes not averages; distinguish trial users from paying customers; named accounts are top-decile success cases — ask about failures.\n- **Product UX:** \"popular choice\" defaults are powerful — notice when a default is doing your decision-making for you.\n- **Astroturf defense:** anomalous account creation timing, repeated language across \"independent\" voices, engagement metrics inconsistent with audience size → drop consensus signal to near-zero.\n\n## Applying It Well\n\n- Similarity is load-bearing: \"other founders chose this\" works on a founder; \"many people chose this\" does not. Always identify whether consensus-makers are similar to you in relevant ways.\n- Three well-placed testimonials approach the power of fifteen — the rule saturates at small numbers.\n- A single visible dissenter destroys most conformity pressure. Find the dissenter or be the dissenter.\n- The rule operates below introspection. The Asch counterfactual is the test, not the self-report.\n- Manufactured social proof has a reputational cliff when discovered. Real social proof compounds.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] \"Everyone is doing it, so it must be right\" | Asch showed this fails on tasks where the right answer was visually obvious. \"Everyone\" is one signal to weigh — not the conclusion. |\n| [D] \"Many smart people are doing X, so X is right\" | Smart people are more invested in being seen as smart, raising the cost of public dissent. The \"smart people\" filter does not defend against engineered consensus. |\n| [D] \"I have my own opinion regardless of what others do\" | Asch's 75% applies even to people who predicted this of themselves. The rule operates below introspection — the counterfactual is the test. |\n| [D] \"Bestseller / most-popular must be the best\" | Popularity reflects discoverability and marketing, not necessarily quality. Treat it as a prior, then update on evidence. |\n| [D] \"If it were wrong, more people would have noticed\" | Public dissent is rare even when private dissent is widespread (Theranos, FTX, 2008 housing market). |\n| [D] \"I noticed the manufactured social proof, so I'm immune\" | Recognition reduces but does not eliminate the pull. Treat recognition as the start of defense, not the conclusion. |\n| [D] Confusing aggregated wisdom with social proof | Markets aggregate information. Social proof aggregates behavior — which may not reflect information. Distinguish \"independent estimators converged\" from \"people followed early movers.\" |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Decision driven by \"many others are doing it\" with no underlying analysis\n- Consensus-makers cannot be verified as similar to you in relevant ways\n- No dissenters visible in a domain where you would expect dissent\n- Consensus measured in low-cost actions (likes, sign-ups) not high-cost ones (purchases, repeat usage)\n- Star distributions are suspiciously skewed (all 5-stars or U-shaped)\n- Social-proof claim growing faster than the underlying user/evidence base could plausibly support\n\n## Verification\n\n- [ ] Consensus claim named precisely (specific group, not generic mass)\n- [ ] Consensus-makers identified — number, similarity, base of information\n- [ ] Consensus classified: informational / social / manufactured\n- [ ] Asch counterfactual performed: same conclusion alone?\n- [ ] Dissenting voices sought; their reasoning evaluated\n- [ ] Manufactured-consensus signs checked\n- [ ] If following: early-warning indicator specified\n- [ ] If sending: proof is real, named, verifiable, representative\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 163 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/skills/social-proof?utm_source=clawhub&utm_medium=marketplace&utm_campaign=knowledge-skills&utm_content=social-proof** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\nFile v1.0.2:_meta.json\n\n{\n  \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n  \"slug\": \"social-proof\",\n  \"version\": \"1.0.2\",\n  \"publishedAt\": 1783472685309\n}\n\nFile v1.0.2:references/sources.md\n\n# Sources — social-proof\n\n> *Primary sources for the [social-proof](../SKILL.md) skill.*\n\n- Asch, S. E. (1951). \"Effects of Group Pressure upon the Modification and Distortion of Judgments.\" In H. Guetzkow (Ed.), *Groups, Leadership, and Men: Research in Human Relations* (pp. 177–190). Pittsburgh: Carnegie Press. The original publication of the line-matching conformity experiments at Swarthmore.\n- Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), 31–35. The most accessible account, written by Asch for a general audience; primary source for the verbatim quotations on the 36.8% conformity rate, the categories of conformity (distortion of perception/judgment/action), the group-size threshold, and the dissenter effect. https://doi.org/10.1038/scientificamerican1155-31\n- Asch, S. E. (1956). \"Studies of Independence and Conformity: I. A Minority of One Against a Unanimous Majority.\" *Psychological Monographs: General and Applied*, 70(9), 1–70. The full monograph documentation. https://doi.org/10.1037/h0093718\n- Sherif, M. (1935). \"A Study of Some Social Factors in Perception.\" *Archives of Psychology*, 27(187), 1–60. The autokinetic-effect experiments; the canonical demonstration that under ambiguity, group norms substitute for objective information. The historical precursor to the Asch line.\n- Milgram, S., Bickman, L., & Berkowitz, L. (1969). \"Note on the Drawing Power of Crowds of Different Size.\" *Journal of Personality and Social Psychology*, 13(2), 79–82. The New York City sidewalk field experiment: one confederate looking up at a building draws ~4% of passers-by, fifteen confederates draw ~86%. https://doi.org/10.1037/h0028070\n- Cialdini, R. B. (1984; 2007 rev.). *Influence: The Psychology of Persuasion*. Harper Business. Chapter 4 is the canonical applied treatment of social proof, including the uncertainty and similarity amplifiers. ISBN 978-0061241895.\n- Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). \"A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades.\" *Journal of Political Economy*, 100(5), 992–1026. The formal model of information cascades — how rational agents looking at the choices of earlier agents can produce socially-proof-style mass behavior even when each individual's private information would have led to a different conclusion. https://doi.org/10.1086/261849\n- Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). \"Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market.\" *Science*, 311(5762), 854–856. The Music Lab experiment: 14,000+ subjects shown otherwise-identical songs converged on radically different \"winners\" depending on the social-proof signals they were shown — empirical demonstration that the consensus a market produces is often determined by social-proof dynamics rather than underlying quality. https://doi.org/10.1126/science.1121066\n\nFile v1.0.2:examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md\n\n# Method in Action: Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nThe empirical foundation of modern social-proof theory rests on a series of experiments conducted by **Solomon Asch** at Swarthmore College between 1951 and 1956. The setup was almost embarrassingly simple. Yet the result it produced has appeared in every introductory social-psychology textbook for the past seventy years, and was selected by the American Psychological Association as one of the twenty most important studies of the 20th century.\n\nAsch recruited male college students under the cover story that he was running a \"visual perception experiment.\" Each session brought seven or nine people into a room and seated them around a table. The participants were shown a pair of cards, one displaying a \"standard\" vertical line, the other displaying three \"comparison\" vertical lines labeled A, B, and C. One of the comparison lines was the same length as the standard; the other two differed visibly. The participants' task was to say aloud, in the order they were seated, which comparison line matched the standard.\n\nIn the control condition — subjects making the judgment alone — the task was trivially easy. Across baseline trials, error rates were under 1%. Anyone with normal vision could see which line matched.\n\nWhat the subjects did not know was that, in the experimental condition, every person in the room except one was a **confederate**. The confederates had been instructed in advance to give the same wrong answer on critical trials. The single real subject was seated near the end of the line, so that he heard the others' answers before giving his own. There were 18 trials total; on 12 of them (\"critical trials\"), the confederates gave a unanimous wrong answer. On the other 6 trials, they answered correctly, so the situation would not feel obviously rigged.\n\nThe question Asch wanted to answer was: given a perceptual judgment so easy that error rates are under 1% in isolation, **how often will a subject publicly endorse a clearly wrong answer simply because everyone else in the room has done so?**\n\nThe empirical result, summarized in Asch's 1955 *Scientific American* article and his 1956 monograph in *Psychological Monographs*, was striking:\n\n> \"Whereas in ordinary circumstances individuals matching the lines will make mistakes less than 1 per cent of the time, under group pressure the minority subjects swung to acceptance of the misleading majority's wrong judgments in 36.8 per cent of the selections.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 33.\n\nAcross all critical trials, naïve subjects gave the wrong answer matching the group's wrong answer approximately one third of the time. Looking at individuals rather than trials, **about 75% of subjects conformed at least once during the experiment, and about 25% never conformed**. Subjects varied: some conformed on nearly every trial, some on a few, some never. But the modal subject was not the independent thinker; the modal subject conformed at least sometimes.\n\nAsch was particularly careful to follow up. In post-experiment interviews, he asked subjects to explain what had happened. The interviews were the part of the work that elevated the experiment from \"interesting curiosity\" to \"foundational finding,\" because they distinguished what people *said* publicly from what they *believed* privately. Three distinct patterns emerged:\n\n> \"Among the independent subjects there were several who showed great steadfastness of purpose and complete confidence in their own judgments. They were able to dismiss the disagreement and remain firmly anchored to their own perceptions... Among the yielding subjects, three categories emerged. (1) *Distortion of perception*. A very few subjects came to perceive the majority estimates as correct. (2) *Distortion of judgment*. Most of the subjects who went along with the majority concluded that their own perceptions were inaccurate and that the majority was right. (3) *Distortion of action*. The third and largest group of yielding subjects did not suffer a modification of perception nor did they conclude that they were wrong. They yielded because of an overmastering need not to appear different from or inferior to others, because they did not want to expose themselves to the scorn of the majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), pp. 34–35.\n\nThis third category — **\"distortion of action\"** — is the central finding for modern social proof analysis. The largest group of conforming subjects **knew the right answer**, **privately maintained that the majority was wrong**, **and still publicly said the majority's answer aloud** rather than be the lone dissenter. They were not deceived. They were not persuaded. They were not even confused. They simply could not bear the social cost of audibly disagreeing.\n\nAsch ran several variants to identify which features of the setup drove the effect. Three findings from the variants are operationally critical.\n\n**First**, group size matters, but not in a linear way. With one or two confederates, the conformity effect was small. With three confederates, it reached most of its strength. Adding more confederates beyond three produced only marginal increases. Asch wrote:\n\n> \"The opposition of two persons produces 12.8 per cent of errors; with three opponents this becomes 31.8 per cent, but with larger opposition the percentage of errors remains substantially the same... The size of the opposition is important only up to a point.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThe practical implication: it does not take a large crowd to produce the rule. Three same-direction signals are nearly as powerful as fifteen. This is why three well-placed customer testimonials, or three \"everyone is doing it\" references in a sales conversation, can be as effective as much larger consensus signals.\n\n**Second, and most important for defense**, the introduction of a single dissenter — even one — dramatically broke the conformity effect. Asch ran a condition where one of the confederates was instructed to give the correct answer, breaking the unanimity. The naïve subjects' conformity rate dropped from roughly 36% to about 5%:\n\n> \"The presence of a supporting partner depleted the majority of much of its power. Its pressure on the dissenting individual was reduced to one fourth: that is, subjects answered incorrectly only one fourth as often as under the pressure of a unanimous majority.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 34.\n\nThis is the most operationally useful finding in the entire literature. **A single visible dissenter — one person willing to say \"no, the right answer is the one you can plainly see\" — destroys most of the conformity pressure**. Not because the dissenter is persuasive on the merits, but because they make the public act of disagreement *survivable*. In any decision setting where you fear social proof is producing manufactured consensus, the single most effective intervention is to **find and amplify the dissenter**, or to *be* the dissenter when needed. The social cost of being one-of-two is qualitatively different from the social cost of being one-of-one.\n\n**Third**, when subjects were allowed to write their answers privately rather than say them aloud, conformity dropped substantially. The effect is largely a *public* phenomenon: the rule operates most strongly when others can witness your answer. This explains why anonymous surveys frequently produce different results from public polls or focus groups, and why decisions that are publicly debated produce different outcomes than decisions made privately and then aggregated.\n\nThe Asch experiments matter for this skill in five specific ways.\n\n**First**, **the rule does not require deception, ambiguity, or weakness**. Asch's task was visually obvious. The subjects were intelligent young men at an elite college. The confederates had no power over them, no formal authority, no claim to expertise. The room had no high stakes. Yet the rule fired in a third of trials. **If social proof can produce conformity in conditions that benign, it can produce it in your business and personal decision-making — particularly when stakes, ambiguity, and reputational costs are all higher.**\n\n**Second**, **the rule fires in the public domain, not the private one**. Most subjects conformed *publicly* while *privately* maintaining the correct answer. This is the core mechanism behind \"the emperor has no clothes\" dynamics: situations where almost everyone privately doubts the conventional wisdom but no one says so publicly because everyone else seems to be on board. The cost of public dissent is real and is structurally what the rule exploits. *Bubbles, fads, doomed business strategies that everyone privately doubts, and political consensuses that quietly invert overnight all operate on this mechanism.*\n\n**Third**, **a single dissenter is enough to break the conformity field**. This is the operational core of the defense. When you suspect a decision is being made under social-proof pressure rather than evidence, the act of being or finding the first dissenter is disproportionately effective. You do not need to overturn the consensus alone; you need to make it possible for others — who already privately agree with you — to say so.\n\n**Fourth**, **the rule is amplified by similarity, uncertainty, and visibility**. Cialdini's framework derives directly from Asch and Sherif: ambiguity (Sherif) and similarity (Cialdini) amplify the effect, and visibility (Asch's public-vs-private finding) is what makes it operate at all. When you deploy social proof in marketing or product design, these are the levers; when you defend against it in your decision-making, these are the warning signs that the rule is likely firing.\n\n**Fifth**, and most critical: **the rule operates below introspection**. Asch's subjects could not, in real time, tell that the group's wrong answer was making them say a wrong answer. The interviews revealed the mechanism only afterward. **In real time, the rule presents as your own judgment, not as external pressure.** This is why social-proof analysis cannot be done by introspection alone — it requires structured exercises (the Asch counterfactual: \"what would I have concluded with only the evidence and no awareness of others?\") to expose what introspection cannot.\n\nThe Asch experiments did not invent the observation that humans conform; folk psychology has known that for millennia. What they did was provide the precise empirical claim that the rule fires *even when consensus is visibly wrong*, that it operates *predominantly in public*, and that *one dissenter is sufficient to break it*. Every modern application of social proof — in marketing, in growth, in product design, in political organizing, in defense against manipulation — ultimately rests on those three findings. They were established in a small experimental room in Pennsylvania in the early 1950s, and they have not been overturned.\n\nFile v1.0.2:skill-card.md\n\n## Description: <br>\nHelps agents analyze consensus claims, testimonials, reviews, popularity signals, and manufactured engagement before using social proof as evidence or designing with it. <br>\n\nThis skill is ready for commercial/non-commercial use. <br>\n\n## Publisher: <br>\n[deciqai](https://clawhub.ai/user/deciqai) <br>\n\n### License/Terms of Use: <br>\nMIT-0 <br>\n\n\n## Use Case: <br>\nEmployees, external users, and agents use this skill to evaluate decisions influenced by what other people appear to be doing. It is also useful when designing or reviewing marketing, sales, and product experiences that use testimonials, popularity claims, reviews, or other social-proof patterns. <br>\n\n### Deployment Geography for Use: <br>\nGlobal <br>\n\n## Known Risks and Mitigations: <br>\nRisk: The skill can influence decisions about consensus, testimonials, reviews, and popularity signals, so poor application could produce misleading guidance. <br>\nMitigation: Review outputs against the skill's verification checklist, especially the consensus claim, dissenting evidence, manufactured-consensus signs, and whether the same decision would be made without social pressure. <br>\nRisk: The security evidence is clean, but the supplied scanner guidance still recommends review before granting access to sensitive repositories, credentials, or production services. <br>\nMitigation: Inspect the skill files and install source before deployment in sensitive environments, and avoid granting unnecessary access. <br>\n\n\n## Reference(s): <br>\n- [ClawHub Skill Page](https://clawhub.ai/deciqai/skills/social-proof) <br>\n- [deciqAI Publisher Profile](https://clawhub.ai/user/deciqai) <br>\n- [Sources](references/sources.md) <br>\n- [Solomon Asch Conformity Experiments Example](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md) <br>\n- [Asch 1955: Opinions and Social Pressure](https://doi.org/10.1038/scientificamerican1155-31) <br>\n- [Asch 1956: Studies of Independence and Conformity](https://doi.org/10.1037/h0093718) <br>\n- [Milgram, Bickman, and Berkowitz 1969](https://doi.org/10.1037/h0028070) <br>\n- [Bikhchandani, Hirshleifer, and Welch 1992](https://doi.org/10.1086/261849) <br>\n- [Salganik, Dodds, and Watts 2006](https://doi.org/10.1126/science.1121066) <br>\n\n\n## Skill Output: <br>\n**Output Type(s):** [analysis, markdown, guidance] <br>\n**Output Format:** [Markdown or plain text structured as a social-proof analysis] <br>\n**Output Parameters:** [1D] <br>\n**Other Properties Related to Output:** [May include a consensus claim, consensus type, dissenters, manufactured-consensus check, decision, and early-warning trigger.] <br>\n\n## Skill Version(s): <br>\n1.0.2 (source: server release evidence) <br>\n\n## Ethical Considerations: <br>\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment. <br>\n\nArchive v1.0.1: 5 files, 12758 bytes\n\nFiles: examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md (11318b), references/sources.md (2933b), skill-card.md (2623b), SKILL.md (9679b), _meta.json (131b)\n\nFile v1.0.1:SKILL.md\n\n---\nname: social-proof\ndescription: \"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or consensus is real, suspects fake reviews or manufactured engagement, or is designing testimonials/UX to convert users. Do NOT activate when: the decision is low-stakes and reversible (choosing a lunch spot) or the user already has direct measured evidence stronger than any consensus signal.\"\n---\n\n# Social Proof\n\n## Overview\n\n**Social proof**: we judge what is correct, normal, or worth doing by observing what others — especially similar others — are doing. Usually efficient; failure mode is severe: under unanimous consensus, people publicly endorse answers they privately know are wrong (Asch 1951–56: error rate <1% alone, ~37% under group pressure). Two amplifiers: **uncertainty** (social proof fills the vacuum) and **similarity** (same-type peers drive far stronger conformity than generic crowds).\n\nComposes with [`reciprocity`](../reciprocity/SKILL.md) (Cialdini's two primary levers), [`anchoring`](../anchoring/SKILL.md) (price tiers often function as quasi-social-proof), and [`critical-thinking`](../critical-thinking/SKILL.md) (structured fallback when consensus has been engineered).\n\n## When to Use\n\n**Use when:** purchase/hiring/investment decision leaning on what others chose; proposal cites \"everyone is doing this\"; designing growth/marketing/UX with social-proof patterns; decision feels unsafe alone without a clear reason; suspecting manufactured consensus (bots, paid reviews, astroturf); a trend is accelerating and private doubt is being suppressed by the fact everyone is on board.\n\n**Do NOT use when:** decision is low-stakes and reversible; you have direct measured evidence stronger than any consensus; the \"consensus\" is from verified domain experts with better epistemic position; you want to rationalize a contrarian position that lacks independent evidence.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has a concrete case → run The Process directly.\n- **Coach mode:** user is unfamiliar or has no concrete case → guide, don't lecture.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** We judge what's correct by looking at what others do — useful most of the time, but under enough unanimous consensus, people will publicly agree with answers they privately know are wrong, even on obvious questions.\n2. **Check fit** against When to Use / When NOT to use. If direct evidence is stronger, point there.\n3. **Elicit the real situation.** A concrete decision shaped by what others are doing, or a design problem deploying social proof. Never run on hypotheticals.\n> **[WAIT — do not advance until user responds]**\n4. **One element at a time.** Walk through: what's the consensus, who are the consensus-makers, are they similar to you / informed, would you decide the same way if alone — wait for input.\n> **[WAIT — do not advance until user responds]**\n5. **Close by naming the payoff.** The one move — accept the consensus, reject it, or seek independent evidence — that fits their situation.\n> **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Social-Proof Analysis**. Diagnose the source of consensus, then decide whether to use it as evidence.\n\n1. **Name the consensus precisely.** Not \"everyone uses Salesforce\" but \"three cohort companies I respect use Salesforce.\" Vague consensus cannot be analyzed.\n2. **Identify consensus-makers.** Who exactly, how many, how similar to you in ways relevant to the decision?\n3. **Classify consensus type.** Informational (converged on evidence) | Social (converged because others did — cascade risk) | Manufactured (engineered appearance via bots, paid reviews, cherry-picked cases).\n4. **Test signal strength.** Did consensus form independently or in chain? What are dissenters saying? What is the base rate for consensus being wrong in this domain?\n5. **Run the Asch counterfactual.** Alone, with only the underlying evidence, would you reach the same conclusion? If no — you've been pulled in by the rule itself.\n6. **Check manufactured-consensus signs:** astroturf, survivorship bias, selected testimonials, engagement-metric inflation.\n7. **As a sender:** real named testimonials, third-party reviews, transparent distributions, limitations in case studies.\n8. **Stop-rule:** if you cannot defend the decision independently of \"many others are doing it,\" treat it as provisional. Plan a fallback.\n\n### Output template\n\n```\nConsensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]\n```\n\n*→ Method in Action: [Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956](examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md)*\n\n## Pack: Social Proof in Practice\n\n- **Sender (marketing/sales):** named logos, real attributed testimonials, third-party reviews, transparent star distributions, case studies with limitations stated. Reference customers similar to the prospect.\n- **Receiver (evaluation):** check distribution shapes not averages; distinguish trial users from paying customers; named accounts are top-decile success cases — ask about failures.\n- **Product UX:** \"popular choice\" defaults are powerful — notice when a default is doing your decision-making for you.\n- **Astroturf defense:** anomalous account creation timing, repeated language across \"independent\" voices, engagement metrics inconsistent with audience size → drop consensus signal to near-zero.\n\n## Applying It Well\n\n- Similarity is load-bearing: \"other founders chose this\" works on a founder; \"many people chose this\" does not. Always identify whether consensus-makers are similar to you in relevant ways.\n- Three well-placed testimonials approach the power of fifteen — the rule saturates at small numbers.\n- A single visible dissenter destroys most conformity pressure. Find the dissenter or be the dissenter.\n- The rule operates below introspection. The Asch counterfactual is the test, not the self-report.\n- Manufactured social proof has a reputational cliff when discovered. Real social proof compounds.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] \"Everyone is doing it, so it must be right\" | Asch showed this fails on tasks where the right answer was visually obvious. \"Everyone\" is one signal to weigh — not the conclusion. |\n| [D] \"Many smart people are doing X, so X is right\" | Smart people are more invested in being seen as smart, raising the cost of public dissent. The \"smart people\" filter does not defend against engineered consensus. |\n| [D] \"I have my own opinion regardless of what others do\" | Asch's 75% applies even to people who predicted this of themselves. The rule operates below introspection — the counterfactual is the test. |\n| [D] \"Bestseller / most-popular must be the best\" | Popularity reflects discoverability and marketing, not necessarily quality. Treat it as a prior, then update on evidence. |\n| [D] \"If it were wrong, more people would have noticed\" | Public dissent is rare even when private dissent is widespread (Theranos, FTX, 2008 housing market). |\n| [D] \"I noticed the manufactured social proof, so I'm immune\" | Recognition reduces but does not eliminate the pull. Treat recognition as the start of defense, not the conclusion. |\n| [D] Confusing aggregated wisdom with social proof | Markets aggregate information. Social proof aggregates behavior — which may not reflect information. Distinguish \"independent estimators converged\" from \"people followed early movers.\" |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Decision driven by \"many others are doing it\" with no underlying analysis\n- Consensus-makers cannot be verified as similar to you in relevant ways\n- No dissenters visible in a domain where you would expect dissent\n- Consensus measured in low-cost actions (likes, sign-ups) not high-cost ones (purchases, repeat usage)\n- Star distributions are suspiciously skewed (all 5-stars or U-shaped)\n- Social-proof claim growing faster than the underlying user/evidence base could plausibly support\n\n## Verification\n\n- [ ] Consensus claim named precisely (specific group, not generic mass)\n- [ ] Consensus-makers identified — number, similarity, base of information\n- [ ] Consensus classified: informational / social / manufactured\n- [ ] Asch counterfactual performed: same conclusion alone?\n- [ ] Dissenting voices sought; their reasoning evaluated\n- [ ] Manufactured-consensus signs checked\n- [ ] If following: early-warning indicator specified\n- [ ] If sending: proof is real, named, verifiable, representative\n\n---\n\n*Part of **deciqAI Knowledge Skills** — open-source thinking skills that make rigor executable for AI agents. Built by deciqAI · https://deciqai.com · Contributions welcome — see the template at the repo root.*\n\nFile v1.0.1:_meta.json\n\n{\n  \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n  \"slug\": \"social-proof\",\n  \"version\": \"1.0.1\",\n  \"publishedAt\": 1783463609825\n}\n\nFile v1.0.1:references/sources.md\n\n# Sources — social-proof\n\n> *Primary sources for the [social-proof](../SKILL.md) skill.*\n\n- Asch, S. E. (1951). \"Effects of Group Pressure upon the Modification and Distortion of Judgments.\" In H. Guetzkow (Ed.), *Groups, Leadership, and Men: Research in Human Relations* (pp. 177–190). Pittsburgh: Carnegie Press. The original publication of the line-matching conformity experiments at Swarthmore.\n- Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), 31–35. The most accessible account, written by Asch for a general audience; primary source for the verbatim quotations on the 36.8% conformity rate, the categories of conformity (distortion of perception/judgment/action), the group-size threshold, and the dissenter effect. https://doi.org/10.1038/scientificamerican1155-31\n- Asch, S. E. (1956). \"Studies of Independence and Conformity: I. A Minority of One Against a Unanimous Majority.\" *Psychological Monographs: General and Applied*, 70(9), 1–70. The full monograph documentation. https://doi.org/10.1037/h0093718\n- Sherif, M. (1935). \"A Study of Some Social Factors in Perception.\" *Archives of Psychology*, 27(187), 1–60. The autokinetic-effect experiments; the canonical demonstration that under ambiguity, group norms substitute for objective information. The historical precursor to the Asch line.\n- Milgram, S., Bickman, L., & Berkowitz, L. (1969). \"Note on the Drawing Power of Crowds of Different Size.\" *Journal of Personality and Social Psychology*, 13(2), 79–82. The New York City sidewalk field experiment: one confederate looking up at a building draws ~4% of passers-by, fifteen confederates draw ~86%. https://doi.org/10.1037/h0028070\n- Cialdini, R. B. (1984; 2007 rev.). *Influence: The Psychology of Persuasion*. Harper Business. Chapter 4 is the canonical applied treatment of social proof, including the uncertainty and similarity amplifiers. ISBN 978-0061241895.\n- Bikhchandani, S., Hirshleifer, D., & Welch\n\nArchive v1.0.0: 5 files, 12758 bytes\n\nFiles: examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md (11318b), references/sources.md (2933b), skill-card.md (2667b), SKILL.md (9679b), _meta.json (131b)","readmeExcerpt":"Skill: Social Proof Owner: deciqai Summary: Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or con... Tags: latest:1.0.5 Version history: v1.0.5 | 2026-07-16T18:16:28.146Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/social-proof.json) v1.0.4 | 2026-07-10T10:28:09.728Z | user A","codeSnippets":[],"executableExamples":[{"language":"text","snippet":"Consensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]"},{"language":"text","snippet":"Consensus claim: \"3 named competitors announced copilot rollouts\" (announcements, not proven ROI); consensus-makers similar in industry, not in data-readiness or disclosed results\nConsensus type: primarily social (cascade) + manufactured overlay; thin informational content\nIndependent vs. chain: chain — each competitor's move fed the next; high cascade risk\nDissenters: analysts warning of high proof-of-concept abandonment; surveys showing most firms cannot yet quantify AI financial return\nAsch counterfactual: alone with only evidence, most would run a metered pilot first, not a broad rollout → no\nManufactured-consensus check: curated vendor case studies, survivorship (shelved pilots stay silent), adoption measured by seats not sustained value\nDecision: seek independent evidence — bounded pilot with a pre-committed ROI metric before scaling\nEarly-warning trigger: pilot misses its pre-set KPI; usage decays after novelty; per-seat cost rises without measured workflow savings"},{"language":"text","snippet":"Consensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]"},{"language":"text","snippet":"Consensus claim: \"3 named competitors announced copilot rollouts\" (announcements, not proven ROI); consensus-makers similar in industry, not in data-readiness or disclosed results\nConsensus type: primarily social (cascade) + manufactured overlay; thin informational content\nIndependent vs. chain: chain — each competitor's move fed the next; high cascade risk\nDissenters: analysts warning of high proof-of-concept abandonment; surveys showing most firms cannot yet quantify AI financial return\nAsch counterfactual: alone with only evidence, most would run a metered pilot first, not a broad rollout → no\nManufactured-consensus check: curated vendor case studies, survivorship (shelved pilots stay silent), adoption measured by seats not sustained value\nDecision: seek independent evidence — bounded pilot with a pre-committed ROI metric before scaling\nEarly-warning trigger: pilot misses its pre-set KPI; usage decays after novelty; per-seat cost rises without measured workflow savings"},{"language":"text","snippet":"Consensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]"},{"language":"text","snippet":"Consensus claim: [specific group, not generic mass; how many; similar to me how?]\nConsensus type: [informational / social / manufactured / mix]\nIndependent vs. chain: [yes/no — cascade risk]\nDissenters: [who, what they say]\nAsch counterfactual: [same conclusion alone? yes/no]\nManufactured-consensus check: [astroturf / survivorship / testimonials / metric inflation]\nDecision: [follow / depart / seek independent evidence] — because [reason]\nEarly-warning trigger: [what would signal consensus is wrong]"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: social-proof\ndescription: \"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or consensus is real, suspects fake reviews or manufactured engagement, or is designing testimonials/UX to convert users. Do NOT activate when: the decision is low-stakes and reversible (choosing a lunch spot) or the user already has direct measured evidence stronger than any consensus signal. More: deciqai.com/c/social-proof\"\n---\n\n# Social Proof\n\n## Overview\n\n**Social proof**: we judge what is correct, normal, or worth doing by observing what others — especially similar others — are doing. Usually efficient; failure mode is severe: under unanimous consensus, people publicly endorse answers they privately know are wrong (Asch 1951–56: error rate <1% alone, ~37% under group pressure). Two amplifiers: **uncertainty** (social proof fills the vacuum) and **similarity** (same-type peers drive far stronger conformity than generic crowds).\n\nComposes with `reciprocity` (Cialdini's two primary levers), `anchoring` (price tiers often function as quasi-social-proof), and `critical-thinking` (structured fallback when consensus has been engineered).\n\n## When to Use\n\n**Use when:** purchase/hiring/investment decision leaning on what others chose; proposal cites \"everyone is doing this\"; designing growth/marketing/UX with social-proof patterns; decision feels unsafe alone without a clear reason; suspecting manufactured consensus (bots, paid reviews, astroturf); a trend is accelerating and private doubt is being suppressed by the fact everyone is on board; a \"we must adopt AI because every competitor is deploying it\" mandate is driving procurement or a pilot ahead of any validated ROI (AI hype / FOMO buying).\n\n**Do NOT use when:** decision is low-stakes and reversible; you have direct measured evidence stronger than any consensus; the \"consensus\" is from verified domain experts with better epistemic position; you want to rationalize a contrarian position that lacks independent evidence.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has a concrete case → run The Process directly.\n- **Coach mode:** user is unfamiliar or has no concrete case → guide, don't lecture.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** We judge what's correct by looking at what others do — useful most of the time, but under enough unanimous consensus, people will publicly agree with answers they privately know are wrong, even on obvious questions.\n2. **Check fit** against When to Use / When NOT to use. If direct evidence is stronger, point there.\n3. **Elicit the real situation.** A concrete decision shaped by what others are doing, or a design problem deploying social proof. Never run on hypotheticals.\n> **[WAIT — do not advance until user responds]**\n4. **One element at a time.** Walk "},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n  \"slug\": \"social-proof\",\n  \"version\": \"1.0.5\",\n  \"publishedAt\": 1784225788146\n}"},{"path":"references/sources.md","content":"# Sources — social-proof\n\n> *Primary sources for the [social-proof](../SKILL.md) skill.*\n\n- Asch, S. E. (1951). \"Effects of Group Pressure upon the Modification and Distortion of Judgments.\" In H. Guetzkow (Ed.), *Groups, Leadership, and Men: Research in Human Relations* (pp. 177–190). Pittsburgh: Carnegie Press. The original publication of the line-matching conformity experiments at Swarthmore.\n- Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), 31–35. The most accessible account, written by Asch for a general audience; primary source for the verbatim quotations on the 36.8% conformity rate, the categories of conformity (distortion of perception/judgment/action), the group-size threshold, and the dissenter effect. https://doi.org/10.1038/scientificamerican1155-31\n- Asch, S. E. (1956). \"Studies of Independence and Conformity: I. A Minority of One Against a Unanimous Majority.\" *Psychological Monographs: General and Applied*, 70(9), 1–70. The full monograph documentation. https://doi.org/10.1037/h0093718\n- Sherif, M. (1935). \"A Study of Some Social Factors in Perception.\" *Archives of Psychology*, 27(187), 1–60. The autokinetic-effect experiments; the canonical demonstration that under ambiguity, group norms substitute for objective information. The historical precursor to the Asch line.\n- Milgram, S., Bickman, L., & Berkowitz, L. (1969). \"Note on the Drawing Power of Crowds of Different Size.\" *Journal of Personality and Social Psychology*, 13(2), 79–82. The New York City sidewalk field experiment: one confederate looking up at a building draws ~4% of passers-by, fifteen confederates draw ~86%. https://doi.org/10.1037/h0028070\n- Cialdini, R. B. (1984; 2007 rev.). *Influence: The Psychology of Persuasion*. Harper Business. Chapter 4 is the canonical applied treatment of social proof, including the uncertainty and similarity amplifiers. ISBN 978-0061241895.\n- Bikhchandani, S., Hirshleifer, D., & Welch, I. (1992). \"A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades.\" *Journal of Political Economy*, 100(5), 992–1026. The formal model of information cascades — how rational agents looking at the choices of earlier agents can produce socially-proof-style mass behavior even when each individual's private information would have led to a different conclusion. https://doi.org/10.1086/261849\n- Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). \"Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market.\" *Science*, 311(5762), 854–856. The Music Lab experiment: 14,000+ subjects shown otherwise-identical songs converged on radically different \"winners\" depending on the social-proof signals they were shown — empirical demonstration that the consensus a market produces is often determined by social-proof dynamics rather than underlying quality. https://doi.org/10.1126/science.1121066\n- Gartner (2024). \"Gartner Predicts 30% of Generative AI Projects Will Be Abandoned Aft"},{"path":"examples/enterprise-ai-copilot-fomo-procurement-2024-2026.md","content":"# Method in Action: Enterprise AI Copilot FOMO Procurement (2024–2026)\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nBetween the launch of ChatGPT in late 2022 and 2026, generative AI moved from novelty to boardroom mandate. By 2024–2025 a recognizable pattern had set in across large organizations: an executive would say some version of *\"every competitor is deploying AI copilots — we can't be the ones left behind,\"* and a procurement cycle would begin. The dominant justification for buying was not a measured business case; it was the observation that **everyone else was buying**. This is a textbook social-proof failure mode — consensus behavior substituting for independent evidence — and it is worth walking through this skill's own Process, because the same pattern recurs in every hype cycle.\n\nThe pattern was widely documented. Industry surveys and analyst commentary through 2024–2025 repeatedly noted a gap between adoption rates and realized returns: adoption of generative AI rose sharply, while a large share of organizations reported difficulty demonstrating financial return from their AI initiatives. A frequently discussed data point was a 2024 Gartner prediction that a substantial share of generative-AI projects would be abandoned after proof-of-concept by the end of 2025, citing poor data quality, unclear business value, escalating costs, and inadequate risk controls. The *phenomenon* — FOMO-driven procurement outrunning validated ROI — was well established before this analysis. Let us run it through the eight Process steps.\n\n## 1. Name the consensus precisely\n\nThe default framing — *\"everyone is deploying AI copilots\"* — is exactly the vague consensus the skill warns against. It cannot be analyzed. Sharpen it:\n\n- Not \"everyone is deploying copilots\" but: *\"three named direct competitors announced Microsoft 365 Copilot or a comparable assistant rollout in press releases and earnings calls, and our board saw those announcements.\"*\n- Distinguish an **announced pilot** (\"we are exploring generative AI\") from a **validated, ROI-positive production deployment**. Public announcements are overwhelmingly the former. The consensus you can actually verify is a consensus of *announcements*, not a consensus of *proven results*.\n\nOnce named precisely, most of the perceived consensus evaporates: what \"everyone\" has done is issue a press release and stand up a pilot — not demonstrate return.\n\n## 2. Identify consensus-makers\n\nWho exactly is the consensus, how many, and how similar to you on dimensions relevant to *whether the tool will pay off*?\n\n- **Vendors and their platform partners** have direct commercial incentive to project a \"everyone is adopting\" narrative.\n- **Competitors' investor-relations announcements** are similar to you in industry but are optimizing for a stock-price and talent-signaling audience, not for disclosing pilot failure rates.\n- **Analysts and the trade press** amplify the loud adopters; the companies that quietly ran "},{"path":"examples/solomon-asch-conformity-experiments-swarthmore-1951-1956.md","content":"# Method in Action: Solomon Asch's Conformity Experiments, Swarthmore, 1951–1956\n\n> *Example for the [social-proof](../SKILL.md) skill.*\n\nThe empirical foundation of modern social-proof theory rests on a series of experiments conducted by **Solomon Asch** at Swarthmore College between 1951 and 1956. The setup was almost embarrassingly simple. Yet the result it produced has appeared in every introductory social-psychology textbook for the past seventy years, and was selected by the American Psychological Association as one of the twenty most important studies of the 20th century.\n\nAsch recruited male college students under the cover story that he was running a \"visual perception experiment.\" Each session brought seven or nine people into a room and seated them around a table. The participants were shown a pair of cards, one displaying a \"standard\" vertical line, the other displaying three \"comparison\" vertical lines labeled A, B, and C. One of the comparison lines was the same length as the standard; the other two differed visibly. The participants' task was to say aloud, in the order they were seated, which comparison line matched the standard.\n\nIn the control condition — subjects making the judgment alone — the task was trivially easy. Across baseline trials, error rates were under 1%. Anyone with normal vision could see which line matched.\n\nWhat the subjects did not know was that, in the experimental condition, every person in the room except one was a **confederate**. The confederates had been instructed in advance to give the same wrong answer on critical trials. The single real subject was seated near the end of the line, so that he heard the others' answers before giving his own. There were 18 trials total; on 12 of them (\"critical trials\"), the confederates gave a unanimous wrong answer. On the other 6 trials, they answered correctly, so the situation would not feel obviously rigged.\n\nThe question Asch wanted to answer was: given a perceptual judgment so easy that error rates are under 1% in isolation, **how often will a subject publicly endorse a clearly wrong answer simply because everyone else in the room has done so?**\n\nThe empirical result, summarized in Asch's 1955 *Scientific American* article and his 1956 monograph in *Psychological Monographs*, was striking:\n\n> \"Whereas in ordinary circumstances individuals matching the lines will make mistakes less than 1 per cent of the time, under group pressure the minority subjects swung to acceptance of the misleading majority's wrong judgments in 36.8 per cent of the selections.\"\n\n— Asch, S. E. (1955). \"Opinions and Social Pressure.\" *Scientific American*, 193(5), p. 33.\n\nAcross all critical trials, naïve subjects gave the wrong answer matching the group's wrong answer approximately one third of the time. Looking at individuals rather than trials, **about 75% of subjects conformed at least once during the experiment, and about 25% never conformed**. Subjects varied: some conformed on nearly"}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or con... Skill: Social Proof Owner: deciqai Summary: Activate when: user says 'everyone is doing this,' 'join thousands of others,' 'it's the popular choice,' 'trusted by X customers,' wonders if a trend or con... Tags: latest:1.0.5 Version history: v1.0.5 | 2026-07-16T18:16:28.146Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/social-proof.json) v1.0.4 | 2026-07-10T10:28:09.728Z | user A","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":2167,"uniquenessScore":52,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-11T16:39:46.460Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-11T16:39:46.460Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-11T20:58:45.156Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}