User story mapping

User Story Mapping

Before you start

Gather the following from the user:

1. What product/feature area are we mapping?
2. Who are the users? (List specific personas or segments, not "users")
3. What is the goal? (Are we mapping for an MVP? A new feature? A redesign?)
4. What constraints exist? (Timeline, team size, technical limitations)

If the user gives you a vague scope ("map our product"), narrow it: "Which user journey should we focus on first?"

Step 1: Identify the backbone (Activities)

Activities are the high-level things users do. They read left-to-right as a narrative.

Rules for activities:

• Use verb phrases: "Discover agents", "Configure workspace", "Monitor performance"
• Keep to 4-8 activities for a single feature area
• Order them chronologically as the user would experience them
• Each activity should be completable in one session

[Discover agents] → [Evaluate agent] → [Install agent] → [Configure agent] → [Monitor usage]

Step 2: Break activities into tasks

Each activity contains 2-5 tasks. Tasks are the steps a user takes within an activity.

Rules for tasks:

• Tasks are smaller verb phrases: "Search catalog", "Read reviews", "Compare options"
• Order them top-to-bottom by typical sequence
• Every task should map to an observable user action

Activity: Discover agents
├── Search catalog
├── Browse categories
├── View trending
└── Read recommendations

Step 3: Generate stories under each task

Stories are the specific, implementable items. Write them in standard format:

As a [persona], I want [action] so that [outcome].

Rules for stories:

• Each story must be independently deliverable
• Each story must be testable (you can write an acceptance criterion)
• Avoid technical stories at this stage — frame everything from the user's perspective
• It's okay to have 3-10 stories per task

Activity: Discover agents
├── Task: Search catalog
│   ├── As a developer, I want to search agents by keyword so that I can find relevant tools quickly
│   ├── As a developer, I want to filter search results by category so that I can narrow down options
│   └── As a developer, I want to see search results ranked by relevance so that the best matches appear first

Step 4: Draw the release slices

Slice the map horizontally. Each slice is a shippable release.

Slice 1 (Walking Skeleton / MVP):

• Pick the ONE story from each task that delivers the minimum end-to-end experience
• The user should be able to complete the entire journey, even if each step is bare-bones
• This slice should be shippable in 1-2 sprints

Slice 2 (v1.0):

• Add the stories that make the experience good, not just functional
• Better search, richer detail pages, smoother flows

Slice 3+ (Future):

• Everything else. Nice-to-haves, power user features, optimizations

Rules for slicing:

• Every slice must be independently shippable and valuable
• The walking skeleton must include at least one story from each critical activity
• Don't put all the "easy" stories in slice 1 — put the most valuable ones

Output format

Present the final story map in this structure:

# Story Map: [Feature/Product Name]

## Personas: [List them]

---

### Activity: [Name]

#### Task: [Name]
- [MVP] Story description
- [v1.0] Story description
- [Future] Story description

#### Task: [Name]
- [MVP] Story description
- [v1.0] Story description

---

### Activity: [Name]
...

Quality checklist

Before delivering a story map, verify:

• Activities read left-to-right as a coherent user narrative
• Every activity has 2-5 tasks
• Every story is independently deliverable and testable
• The MVP slice delivers a complete (if minimal) end-to-end journey
• No slice contains only backend/technical work with no user-facing value
• Stories are framed from the user's perspective, not technical tasks
• The map has been reviewed against the stated personas — does each persona have a clear path?

Common mistakes to avoid

• Starting with stories instead of activities. Always build the backbone first. Stories without a backbone are just a flat backlog.
• MVP that's too thick. The walking skeleton should be embarrassingly thin. If your MVP has more than 2 stories per task, you're not cutting enough.
• Confusing tasks with stories. "Search catalog" is a task. "As a developer, I want to search by keyword" is a story. Tasks are categories; stories are deliverables.
• Leaving out the "return" journey. Users don't just complete a flow once. Map what happens when they come back: re-evaluate, update, remove, monitor.