What makes Jobs to Be Done (JTBD) notable?

1.3K downloads. Trust evidence available

How should Jobs to Be Done (JTBD) be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

agentCLAWHUBUnverified

Jobs to Be Done (JTBD)

Activate when: user says 'customers aren't switching to us even though we're better,' 'our churn surveys aren't predicting who actually leaves,' 'we're debat... Skill: Jobs to Be Done (JTBD) Owner: deciqai Summary: Activate when: user says 'customers aren't switching to us even though we're better,' 'our churn surveys aren't predicting who actually leaves,' 'we're debat... Tags: latest:1.0.5 Version history: v1.0.5 | 2026-07-16T18:03:44.748Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/jobs-to-be-done.json) v1.0.4 | 2026-07-09T11:18:26.

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

1.3k

Updated

Oct 10, 2026

Version

1.0.5

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1.3K downloads reported by the source. Last updated 10/10/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 10, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 10, 2026
Adoption signal: 1.3K downloadsadoption · observed Oct 10, 2026
Latest release: 1.0.5release · observed Jul 16, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s17a4mqcnk515kvaca5ze55d0x88pfpx:jobs-to-be-done

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: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-deciqai-jobs-to-be-done/snapshot"

Documentation

CLAWHUB

128,164 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: jobs-to-be-done
description: "Activate when: user says 'customers aren't switching to us even though we're better,' 'our churn surveys aren't predicting who actually leaves,' 'we're debating features instead of what the customer actually needs,' 'who is our real customer,' or mentions 'JTBD / jobs to be done / what are they hiring this for.'
  Do NOT activate when: product is a commodity with no job-level differentiation (electricity, raw materials); purchase is driven entirely by regulatory/legal compliance with no real customer choice. More: deciqai.com/c/jobs-to-be-done"
---

# Jobs to Be Done (JTBD)

## Overview

People don't buy products — they **hire** products to do a job (make progress in a specific circumstance, across functional, emotional, and social dimensions). Customers switch when a new hire does the job better; they churn when your product stops serving the job. Developed by Christensen, Moesta, and Taddy Hall; codified in *Competing Against Luck* (2016). Rooted in Levitt's 1960 insight: "People don't want a quarter-inch drill, they want a quarter-inch hole."

Composes with `pmf-crossing-the-chasm`, `mvp`, `switching-costs`, `first-principles`.

## When to Use

- Product is technically excellent but customers don't switch from incumbents
- Demographic segmentation produces segments that don't behave alike
- Churn is high but exit surveys don't predict it; roadmap debate is feature-vs-feature
- New market entry: "who is our customer" instead of "what job"
- Building an AI-native product or "AI wrapper": are users hiring us for output/features, or for progress (get unblocked, ship faster) — and are we losing to AI adoption, the base model's own app, or non-consumption?

**Not when:** commodity; regulatory-compliance purchase; org buyer with different motivations than end-user.

## Coaching Novices (Adaptive Front Door)

- **Engine mode:** concrete product/customer case → run The Process directly.
- **Coach mode:** unfamiliar or no concrete case → guide step by step.

In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.

1. One-line: people hire products to do a job — the competitor set is *everything* the buyer considered, not just your category.
2. Check fit: commodity / regulatory-buy / no-choice → not this lens.
3. Elicit the real product and customer behavior they're trying to understand.
> **[WAIT — do not advance until user responds]**
4. One question at a time: what job is the customer hiring this for? what circumstance? what did they hire before?
> **[WAIT — do not advance until user responds]**
5. Close: job statement + the non-obvious competitor they're actually choosing between.
> **[WAIT — do not advance until user responds]**

## The Process

**Step 1 — State product + assumed customer** (starting point; will be dismantled).

**Step 2 — Switch Interviews.** Interview recent switchers to/from your product. Reconstruct the switching moment: (1) Firs

_meta.json

{
  "ownerId": "kn754b8sk22s8c6gjxt02bftbn88q7ye",
  "slug": "jobs-to-be-done",
  "version": "1.0.5",
  "publishedAt": 1784225024748
}

references/sources.md

# Sources — jobs-to-be-done

> *Primary sources for the [jobs-to-be-done](../SKILL.md) skill.*

- Christensen, C. M., Hall, T., Dillon, K., & Duncan, D. S. (2016). *Competing Against Luck: The Story of Innovation and Customer Choice.* HarperBusiness. ISBN 978-0062435613. The canonical book-length treatment.
- Christensen, C. M., Anthony, S. D., Berstell, G., & Nitterhouse, D. (2007). "Finding the right job for your product." *MIT Sloan Management Review*, 48(3), 38-47. Early academic articulation of the framework.
- Levitt, T. (1960). "Marketing Myopia." *Harvard Business Review*, 38(4), 45-56. The intellectual ancestor: "People don't want a quarter-inch drill, they want a quarter-inch hole."
- Moesta, B. & Spiek, C. (2020). *Demand-Side Sales 101: Stop Selling and Help Your Customers Make Progress.* Lioncrest. Practical Switch Interview methodology.
- Ulwick, A. W. (2016). *Jobs to be Done: Theory to Practice.* Idea Bite Press. A parallel "outcome-driven innovation" school of JTBD with stronger quantitative emphasis.
- "Clay Christensen's Milkshake Marketing." *HBS Working Knowledge*, Feb 2007. https://hbswk.hbs.edu/item/clay-christensens-milkshake-marketing
- Stack Overflow. *Developer Survey* (2023 and 2024 editions). https://survey.stackoverflow.co — documents rapid adoption of AI coding tools among professional developers (majority using or intending to use AI in their workflow), evidence for the 2023–2026 AI-assistant JTBD example.
- GitHub. Public reporting and research on AI-assisted development / Copilot adoption and productivity (2023–2024). https://github.blog — supports the "get unblocked, ship faster" job framing for AI coding assistants.

examples/christensen-and-the-milkshake-study-2003.md

# Method in Action: Christensen and the Milkshake Study, 2003

> *Example for the [jobs-to-be-done](../SKILL.md) skill.*

The most famous illustration of JTBD comes from a study Clayton Christensen and Bob Moesta conducted for a major fast-food chain in 2002-2003. The chain wanted to sell more milkshakes. They had spent considerable resources on the conventional approach: customer demographic surveys, segmentation into target customer profiles, focus groups asking "would you buy more milkshakes if they were thicker / sweeter / cheaper / had more flavors?"

The result of conventional research was a thicker, sweeter, more-flavored milkshake. Sales did not move.

Christensen's team tried a different approach: they stood in the restaurant and observed who bought milkshakes and when. Two patterns emerged immediately:

1. A surprisingly large share of milkshakes were sold *before 9 AM*.
2. The before-9-AM buyers were almost all solo adults in business attire, buying just a milkshake and nothing else, drinking it in their car.

The team began interviewing the morning milkshake buyers as they left the restaurant. The question was not "what could we add to make this better?" but "what *job* did you hire that milkshake for this morning?"

The answers converged on a specific job statement:

> *I have a long, boring drive to work. I need something to do with my hand and mouth that will make the commute less boring. It has to last the whole drive — about 20 minutes — so something I finish in 3 bites won't do. It has to not make a mess so I can eat it one-handed while driving. It has to not leave me hungry by 10 AM.*

The job framework reframed the entire competitive picture:

- **Bagels** were a competitor — but bagels crumb, need cream cheese, and require two hands.
- **Bananas** were a competitor — but bananas are eaten in 90 seconds.
- **Donuts** were a competitor — but donuts leave the customer hungry within an hour.
- **Coffee** was a competitor — but coffee doesn't fill you up.
- **Doing nothing** was the most common competitor — many morning commuters were defaulting to a boring drive with nothing.

The milkshake won this competition because it was thick enough to last 20 minutes, dense enough to not leave the buyer hungry, and consumable one-handed without mess. The conventional research had been measuring whether the *milkshake* could be better — but the milkshake was already winning the morning job. **The optimization opportunity was to make the milkshake even more efficient at the existing job** (thicker, faster to serve so commuters didn't wait), not to make it taste like more flavors.

Critically, the team also found a second morning shake-buyer pattern: a parent buying a milkshake for a child in the late afternoon, as a treat. **Same product, completely different job.** The afternoon shake needed to be small (the parent didn't want the child to get full), and the parent's emotional need was "I want to be a good parent who treats my child sometim

examples/what-people-hire-an-ai-assistant-to-do-2023-2026.md

# Method in Action: What People Hire an AI Assistant to Do (2023–2026)

> *Example for the [jobs-to-be-done](../SKILL.md) skill.*

In the years after the launch of ChatGPT in late 2022, many "AI wrapper" products shipped: an interface, a system prompt, and a call to an underlying model such as OpenAI's GPT family, Anthropic's Claude, or Google's Gemini. A large share of them reportedly struggled to retain users, even when the model underneath was excellent. This case runs the JTBD process on a recurring, non-obvious question: **what job does a knowledge worker actually hire an AI assistant/agent to do — and why do so many wrappers misread it?**

## Step 1 — State product + assumed customer

Product: an AI assistant/agent (a chat product or an agentic coding/writing tool) built on top of a frontier model. Assumed customer, as most 2023-era wrappers framed it: "people who want AI-generated text / answers." The implicit assumption baked into most products was that the job is *"produce good output"* and that the winner is whoever wraps the smartest model with the most features. This is the starting point we will dismantle.

## Step 2 — Switch Interviews

Reconstructing the switching moment from widely-reported adoption patterns across 2023–2025 (developers adopting AI coding assistants; writers, analysts, and support teams adopting chat assistants):

1. **First thought:** "I'm stuck / this will take me all afternoon / I dread starting this." The trigger is rarely "I wish I had AI-written text" — it's a moment of being blocked, behind, or facing drudgery.
2. **Circumstance:** a blank page, an unfamiliar codebase, a deadline, a repetitive task (reformatting, boilerplate, first-draft email), or a question whose answer is buried in docs.
3. **What else they considered:** searching the web, asking a colleague, copying an old template, reading documentation, or simply grinding through it manually.
4. **Push (what was wrong with the old way):** searching returns generic results that still need synthesis; the colleague is busy; the manual path is slow and boring; getting started is the hardest part.
5. **Pull (what attracted the new hire):** it gets me *unblocked in seconds*, produces a *good-enough starting point* I can edit, and handles the parts I don't want to do myself.
6. **Anxiety:** "Will it hallucinate and embarrass me? Will I ship something wrong? Do I have to check every line — and if so, did it even save me time?" Trust and verification cost are the dominant anxieties.
7. **Habit:** having to phrase things as prompts, learning to paste in context, changing where the work starts (in the assistant vs in the IDE/doc).
8. **First use:** the products that retained users delivered a *"whoa, that unblocked me"* moment fast; the ones that didn't left users with plausible-looking output they still had to redo.

## Step 3 — Extract job statement

> **When** I'm blocked, behind, or facing tedious work I don't want to do, **I want to** make concrete progress

Github ReposUpdated 1d agoRank 70

AionUi

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!

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

activepieces

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

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.