continuous-discovery（continuous discovery 搜索）

Continuous Discovery Habits

When This Skill Activates

Claude uses this skill when:

• Setting up discovery processes
• Planning weekly user research
• Creating opportunity solution trees
• Testing assumptions
• Building product trio workflows
• Prioritizing discovery activities

Core Frameworks

1. Continuous Discovery Habits (Source: Teresa Torres)

The Core Principle:

"At a minimum, weekly touchpoints with customers by the team building the product, where they conduct small research activities in pursuit of a desired outcome."

The Three Pillars:

1. Weekly customer contact by the product trio (PM, designer, engineer)
2. Opportunity solution trees to visualize discovery
3. Assumption testing before building

Use when: Establishing discovery processes or improving product decisions

2. The Product Trio

The Team:

• Product Manager: Ensures business viability
• Designer: Ensures usability and desirability
• Engineer: Ensures feasibility

Why Together:

"When a designer, engineer, and PM collaborate on discovery, you get better decisions faster. Each brings a unique lens."

How:

• All three participate in customer interviews
• All three analyze research together
• All three generate solutions together
• All three test assumptions together

3. Opportunity Solution Trees

The Structure:

Outcome (top)
    ↓
Opportunities (customer needs/pain points)
    ↓
Solutions (possible ways to address)
    ↓
Assumptions (what needs to be true)
    ↓
Experiments (how to test)

Use when: Need to visualize the path from outcome to solution

Example:

Outcome: Increase retention to 80%
    ↓
Opportunity: Users forget to use the product
    ↓
Solution: Daily email reminder
    ↓
Assumption: Users check email daily
    ↓
Experiment: Survey 20 users about email habits

4. Interview Snapshot

The One-Pager: After each interview, create a snapshot capturing:

• Date & Participant: Who and when
• Key Insights: 3-5 main takeaways
• Opportunities: Customer needs/pain points discovered
• Quotes: Verbatim customer language
• Next Steps: What to test or explore next

Why: Keeps learning accessible to the whole team

5. Assumption Testing

The Progression:

Story → Assumptions → Tests → Evidence

Question to Ask:

"What needs to be true for this solution to work?"

Test Types (by risk/cost):

1. One-question surveys (lowest risk)
2. Customer interviews
3. Prototypes/mockups
4. Concierge tests (manual behind-the-scenes)
5. Wizard of Oz (fake the feature)
6. Live data tests (build and measure)

Rule: Test highest-risk assumptions first with lowest-cost method

Decision Trees

Should I Build This Feature?

Do we have a clear outcome?
├─ No → Define outcome first
└─ Yes → Have we interviewed 6+ customers?
    ├─ No → Do discovery first
    └─ Yes → Have we identified opportunities?
        ├─ No → Map opportunities
        └─ Yes → Have we tested key assumptions?
            ├─ No → Test assumptions first
            └─ Yes → Build it!

How Should I Test This Assumption?

What's the risk if we're wrong?
├─ Low risk → Build and ship (reversible)
└─ High risk → How much does testing cost?
    ├─ Low cost → Interview 5 users
    ├─ Medium cost → Prototype test
    └─ High cost → Still cheaper than building wrong thing

Action Templates

Template: Weekly Discovery Plan

# Weekly Discovery Plan - Week of [Date]

## Outcome We're Pursuing
[e.g., Increase activation rate to 50%]

## This Week's Focus
**Opportunity:** [Which pain point are we exploring?]
**Solution:** [Which solution are we considering?]
**Key Assumption:** [What needs to be true?]

## Discovery Activities (Minimum 1 per week)

### Monday-Wednesday: Research
- [ ] Interview 1: [Participant profile] - [PM/Designer/Engineer attending]
- [ ] Interview 2: [Participant profile] - [PM/Designer/Engineer attending]
- [ ] Interview 3: [Participant profile] - [PM/Designer/Engineer attending]

### Thursday: Synthesis
- [ ] Product trio synthesis session (30 min)
- [ ] Create/update interview snapshots
- [ ] Update opportunity solution tree
- [ ] Identify new assumptions to test

### Friday: Planning
- [ ] Review evidence collected
- [ ] Decide: build, test more, or pivot?
- [ ] Plan next week's discovery activities

## Interview Snapshots
[Link to snapshots folder]

## Opportunity Solution Tree
[Link to latest tree]

Template: Interview Snapshot

# Interview Snapshot - [Date]

## Participant
- **Name/ID:** [Anonymized if needed]
- **Role:** [Job title/persona]
- **Context:** [Relevant background]

## Interview Focus
[What we were trying to learn]

## Key Insights
1. [First major insight]
2. [Second major insight]
3. [Third major insight]

## Opportunities Discovered
- 📍 [Pain point or unmet need #1]
- 📍 [Pain point or unmet need #2]
- 📍 [Pain point or unmet need #3]

## Memorable Quotes
> "[Exact customer words that capture key point]"

> "[Another powerful quote]"

## Updated Assumptions
- ✅ Validated: [What we confirmed]
- ❌ Invalidated: [What we disproved]
- ❓ New: [New assumptions to test]

## Next Steps
- [ ] [Specific action based on learning]
- [ ] [Another action]

## Attending
- [PM name]
- [Designer name]
- [Engineer name]

Template: Opportunity Solution Tree

# Opportunity Solution Tree - [Product/Feature Name]

## Outcome
🎯 **[Business outcome we're driving]**
[Specific, measurable, time-bound]

---

## Opportunities (Customer Needs/Pain Points)

### Opportunity 1: [Customer problem]
**Evidence:** [3-5 customer interviews, usage data, etc.]
**Impact:** [How big is this problem?]

#### Solutions Being Considered:
1. **[Solution A]**
   - Assumptions:
     - [ ] Assumption 1
     - [ ] Assumption 2
   - Tests: [How we'll validate]
   - Status: [Testing/Building/Shipped]

2. **[Solution B]**
   - Assumptions:
     - [ ] Assumption 1
     - [ ] Assumption 2
   - Tests: [How we'll validate]
   - Status: [Testing/Building/Shipped]

### Opportunity 2: [Another customer problem]
**Evidence:** [3-5 customer interviews, usage data, etc.]
**Impact:** [How big is this problem?]

[Continue for each opportunity...]

---

## Decision Log
- **[Date]:** Chose Solution A for Opportunity 1 because [evidence]
- **[Date]:** Decided to test Assumption X before building
- **[Date]:** Pivoted from Solution B to Solution C based on [learning]

Template: Assumption Test Plan

# Assumption Test Plan - [Feature/Solution Name]

## Solution Statement
[Brief description of what we're considering building]

## Key Assumptions

### Assumption 1: [High Risk]
**Statement:** [What needs to be true]
**If wrong:** [What's the impact?]
**Confidence:** [Low/Medium/High]

**Test Method:** [Interview/Survey/Prototype/etc.]
**Success Criteria:** [What would validate this?]
**Timeline:** [When we'll test]
**Owner:** [Who's running the test]

---

### Assumption 2: [Medium Risk]
**Statement:** [What needs to be true]
**If wrong:** [What's the impact?]
**Confidence:** [Low/Medium/High]

**Test Method:** [Interview/Survey/Prototype/etc.]
**Success Criteria:** [What would validate this?]
**Timeline:** [When we'll test]
**Owner:** [Who's running the test]

---

## Test Results

### Assumption 1 Results
**Date Tested:** [Date]
**Method Used:** [What we did]
**Sample Size:** [How many participants]

**Findings:**
- [Key finding 1]
- [Key finding 2]
- [Key finding 3]

**Decision:** ✅ Validated / ❌ Invalidated / ❓ Needs more testing
**Next Steps:** [What we'll do based on results]

---

### Assumption 2 Results
[Same structure as above]

Quick Reference

📅 Weekly Discovery Cadence

Every Week Minimum:

• 3-5 customer touchpoints (interviews, observation, etc.)
• Product trio participates together
• Create interview snapshots
• Update opportunity solution tree
• Test at least 1 assumption

Every Month:

• Review all evidence collected
• Update outcomes if needed
• Celebrate learning (not just building)

🎯 Discovery vs Delivery Balance

Good Discovery Practice:

• ✅ Discovery happens weekly (not just quarterly)
• ✅ Product trio does discovery together
• ✅ Small tests before big builds
• ✅ Evidence-based decisions
• ✅ Comfortable saying "we learned that won't work"

Signs of Insufficient Discovery:

• ❌ Only talking to customers after shipping
• ❌ PM does all research alone
• ❌ Building first, validating later
• ❌ Opinion-based decisions
• ❌ Fear of "wasting time" on research

🌳 Opportunity Solution Tree Checklist

Before Creating:

• Clear outcome defined
• Conducted 6+ customer interviews
• Identified multiple opportunities

When Building Tree:

• Start with ONE outcome (top)
• Map opportunities (not solutions)
• Generate multiple solutions per opportunity
• List assumptions for each solution
• Plan tests for assumptions

Using the Tree:

• Update weekly with new learning
• Share with stakeholders
• Use to explain why you're building what
• Reference when prioritizing work

🧪 Assumption Testing Hierarchy

Test in This Order:

1. Desirability - Do customers want this?
2. Usability - Can they use it?
3. Feasibility - Can we build it?
4. Viability - Should we build it?

Use Cheapest Test First:

Interview < Survey < Prototype < Concierge < Build

Real-World Examples

Example: Spotify's Discovery Process

Outcome: Increase music discovery engagement

Opportunity: Users don't know what to listen to

• Evidence: Interviews showed decision fatigue
• Solution considered: Algorithmic playlists
• Assumption: Users trust algorithmic recommendations
• Test: Created Discover Weekly, measured engagement
• Result: Massive success, became core feature

Key Learning: They tested the algorithm assumption before building fancy UX

Example: Netflix's Continue Watching

Outcome: Reduce time to content consumption

Opportunity: Users forget what they were watching

• Evidence: Drop-off analysis + customer interviews
• Solution: "Continue Watching" row
• Assumption: Users want to resume (not restart)
• Test: A/B test with 5% of users
• Result: Validated, rolled to 100%

Key Learning: Small test before full build saved months of work

Common Pitfalls

❌ Discovery Theater

Problem: Doing research but not changing decisions Solution: Explicitly decide what you'll do if assumptions are wrong

❌ Outsourcing Discovery

Problem: PM does research, then "throws it over the wall" Solution: Product trio interviews together

❌ Building Multiple Solutions at Once

Problem: Spreading resources too thin Solution: Test assumptions first, build one at a time

❌ Skipping Discovery "To Move Fast"

Problem: Building wrong thing is slowest path Solution: Small tests are faster than big rebuilds

❌ Only Talking to Happy Customers

Problem: Missing problems and churn reasons Solution: Interview across the spectrum (new, power, churned users)

Key Quotes

Teresa Torres on Weekly Contact:

"If you're not talking to customers every week, you're not doing continuous discovery."

On Product Trios:

"The best product decisions come from diverse perspectives. A PM, designer, and engineer will see different things in the same customer interview."

On Opportunity Solution Trees:

"The tree makes your thinking visible. It shows how you got from an outcome to a solution, which builds stakeholder trust."

On Assumption Testing:

"Don't ask customers what to build. Test assumptions about what will work."

On Discovery vs Delivery:

"Discovery and delivery should happen continuously. Discovery doesn't end when you start building."

Related Skills

Use together with:

• user-feedback-system - For ongoing feedback collection
• jtbd-building - For understanding customer motivations
• exp-driven-dev - For testing assumptions with data
• metrics-frameworks - For defining outcomes
• strategic-build - For deciding what's worth discovering

Comes before:

• zero-to-launch - Discover before building
• design-first-dev - Design based on discovery

Comes after:

• strategy-frameworks - Define strategy, then discover how

Quick Start Guide

Week 1: Set Up Discovery Process

1. Form product trio (PM, designer, engineer)
2. Define one clear outcome to pursue
3. Schedule first 3 customer interviews
4. Create interview snapshot template

Week 2: Start Discovery Habit

1. Conduct 3 interviews together
2. Create interview snapshots
3. Begin opportunity solution tree
4. Identify opportunities from interviews

Week 3: Map Solutions

1. Generate 3+ solutions per opportunity
2. List assumptions for each solution
3. Prioritize which assumptions to test
4. Plan assumption tests

Week 4: Test Assumptions

1. Run first assumption tests
2. Update opportunity solution tree
3. Decide: build, test more, or pivot
4. Make discovery routine sustainable

Remember: Continuous discovery isn't a phase. It's a habit. The product trio that talks to customers weekly makes better product decisions.

Guest: Teresa Torres Book: Continuous Discovery Habits (2021) Website: producttalk.org Known for: Opportunity Solution Trees, Product Trios, Weekly Touchpoints