Skip to content
cccskills
FreePublished on GitHub

nw-persona-jtbd-analysis

Structured persona creation and JTBD analysis methodology - persona templates, ODI job step tables, pain point mapping, success metric quantification, and multi-persona segmentation

The English summary is being prepared. Showing the original description.

See how to install

Files included (1)

  • SKILL.md4.9 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Persona and JTBD Analysis

Use during Phase 1 (GATHER) when the user lacks clear personas or the "Who" section needs rigorous definition. Provides structured alternative to ad-hoc persona descriptions.

Persona Template

For each user type, build a complete persona:

## Persona: {Name}

**Who**: {Role description -- one sentence capturing relationship to product}
**Demographics**:
- {Characteristic 1: e.g., technical proficiency level}
- {Characteristic 2: e.g., frequency of interaction}
- {Characteristic 3: e.g., environment/context of use}
- {Characteristic 4: e.g., primary motivation}

**Jobs-to-be-Done**: (see Job Step Table below)

**Pain Points**:
- {Pain 1} -- maps to Job Step: {step name}
- {Pain 2} -- maps to Job Step: {step name}

**Success Metrics**:
- {Quantified outcome 1: e.g., "Task completed in < 2 minutes"}
- {Quantified outcome 2: e.g., "Zero manual configuration steps"}

Job Step Table

Each persona has job steps describing what they accomplish. Steps follow ODI format.

Job StepGoalDesired Outcome
{Verb}{What the user wants to achieve}Minimize {metric} of {undesirable state}

Rules for Job Steps

  • Job Steps are always verbs: Discover, Validate, Install, Configure, Monitor, Recover
  • Goals describe what the user wants, not what the system does
  • Desired Outcomes use ODI format: "Minimize [time/effort/risk/likelihood] of [undesirable state]"
  • Each step maps to a workflow point where value is created or destroyed

Example Job Step Table

Job StepGoalDesired Outcome
DiscoverFind the right tool for the taskMinimize time to evaluate fit
InstallGet the tool running locallyMinimize steps to working state
ConfigureAdapt to local environmentMinimize likelihood of misconfiguration
VerifyConfirm correct installationMinimize uncertainty about readiness
StartBegin productive workMinimize time from install to first output

Pain Point Mapping

Every pain point maps to a specific job step. Pain points without a corresponding step indicate either a missing step or irrelevant pain point.

Pain Point: "I don't know if the tool supports my OS"
  -> Job Step: Discover
  -> Desired Outcome: Minimize uncertainty about compatibility

Pain Point: "Installation fails silently with no error message"
  -> Job Step: Install
  -> Desired Outcome: Minimize time to diagnose installation failures

Prioritize: pain points on high-frequency job steps deserve attention first.

Success Metric Quantification

Every success metric needs a number or threshold. Qualitative metrics ("easy to use") are not actionable.

QualitativeQuantified
"Easy to install""Install completed in < 2 minutes with zero manual steps"
"Fast startup""First productive output within 30 seconds of launch"
"Reliable""Zero silent failures; all errors produce actionable messages"
"Intuitive""New user completes core task without reading documentation"

Multi-Persona Segmentation

Different users have fundamentally different jobs even when using the same product. Segment by relationship to the product.

Common axes: Frequency (first-time vs returning vs power user) | Role (end user vs admin vs developer) | Context (individual vs team vs CI/CD) | Motivation (exploration vs production vs evaluation)

Example: Same Product, Different Jobs

PersonaPrimary JobKey Difference
ExplorerEvaluate the tool quicklyNeeds fast time-to-value, minimal commitment
ReturnerResume work after absenceNeeds state preservation, quick re-orientation
DeployerInstall for a teamNeeds configuration management, multi-user setup
AutomatorIntegrate into CI/CD pipelineNeeds scriptability, headless operation, exit codes

Each persona gets their own Job Step table because workflows differ. Do not merge personas -- JTBD analysis value comes from surfacing differences.

Integration with Story Crafting

After completing persona analysis, feed results into LeanUX user story template:

  1. Persona Who -> populated from persona template
  2. Persona Pain Points -> inform story Problem section
  3. Job Step Desired Outcomes -> inform AC (ODI outcomes translate to testable criteria)
  4. Success Metrics -> inform NFR requirements in handoff package

Cross-reference: use bdd-requirements skill for Example Mapping once personas established. Use jtbd-workflow-selection skill to determine workflow for resulting stories.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation

The English summary is being prepared. Showing the original description.

nWave-ai/nWave622Updated Sep 16, 2026

Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation

The English summary is being prepared. Showing the original description.

nWave-ai/nWave622Updated Sep 16, 2026

Review dimensions for acceptance test quality - happy path bias, GWT compliance, business language purity, coverage completeness, walking skeleton user-centricity, priority validation, observable behavior assertions, traceability coverage, and walking skeleton boundary proof

The English summary is being prepared. Showing the original description.

nWave-ai/nWave622Updated Sep 16, 2026

Detailed 5-phase workflow for creating agents - from requirements analysis through validation and iterative refinement

The English summary is being prepared. Showing the original description.

nWave-ai/nWave622Updated Sep 16, 2026

5-layer testing approach for agent validation including adversarial testing, security validation, and prompt injection resistance

The English summary is being prepared. Showing the original description.

nWave-ai/nWave622Updated Sep 16, 2026

Architectural style selection decision matrices, trade-off analysis, structural enforcement rules, and combination patterns. Load when choosing or evaluating architecture styles.

The English summary is being prepared. Showing the original description.

nWave-ai/nWave622Updated Sep 16, 2026

See all skills by nWave-ai

Report a problem with this skill