---

name: assumption-testing description: Identify, prioritize, and test product assumptions systematically. Reduce risk by validating riskiest assumptions first through structured experiments. allowed-tools: Read, Write, Glob, Grep, Task, WebSearch, WebFetch

Assumption Testing

When to Use This Skill

Use this skill when:

• Assumption Testing tasks - Working on identify, prioritize, and test product assumptions systematically. reduce risk by validating riskiest assumptions first through structured experiments
• Planning or design - Need guidance on Assumption Testing approaches
• Best practices - Want to follow established patterns and standards

Overview

Every product idea rests on unvalidated assumptions. Assumption testing systematically identifies these assumptions, prioritizes by risk, and designs experiments to validate or invalidate them before significant investment.

The DVFUE Framework

Five categories of assumptions to test:

Category	Question	Failure Mode
Desirability	Will customers want this?	Nobody uses it
Viability	Does the business model work?	Unsustainable economics
Feasibility	Can we build this?	Technical blockers
Usability	Can users figure it out?	Too confusing to use
Ethical	Should we build this?	Harm to users/society

Assumption Identification

Step 1: List All Assumptions

For your product/feature, brainstorm assumptions across categories:

Desirability Assumptions:

• Users have this problem
• The problem is painful enough to solve
• Users will pay for a solution
• Users prefer our approach over alternatives
• The timing is right

Viability Assumptions:

• We can acquire customers affordably
• Customers will pay our target price
• Customer lifetime value exceeds acquisition cost
• The market is large enough
• We can scale the business model

Feasibility Assumptions:

• We can build the core technology
• We have the required expertise
• We can deliver within time/budget constraints
• Third-party dependencies will work
• Performance requirements are achievable

Usability Assumptions:

• Users will understand the interface
• Users can complete key tasks
• Learning curve is acceptable
• Accessibility requirements can be met
• Users will adopt new workflows

Ethical Assumptions:

• The solution doesn't cause harm
• Data usage is acceptable to users
• No unintended negative consequences
• Aligns with our values
• Regulatory compliance is achievable

Step 2: Identify Leap of Faith Assumptions

Leap of Faith Assumptions are the assumptions that:

1. If wrong, would sink the entire project
2. You have no data to support
3. You're taking on faith

Prioritization Questions:

• If this is wrong, does the project fail? → High priority
• Do we have evidence for this? → If no, higher priority
• How easily can we test this? → Easier = test sooner

Assumption Mapping

Risk × Certainty Matrix

         High Certainty
              │
   ┌──────────┼──────────┐
   │   Known  │  Known   │
   │   Safe   │   Risk   │
   │ (Ignore) │ (Mitigate)│
Low├──────────┼──────────┤High
Risk│ Unknown │ Unknown  │Risk
   │   Safe   │   Risk   │
   │ (Defer)  │  (TEST!) │
   └──────────┼──────────┘
              │
         Low Certainty

Priority Order:

1. Unknown Risk (High risk, low certainty) → Test immediately
2. Known Risk (High risk, high certainty) → Mitigate
3. Unknown Safe (Low risk, low certainty) → Defer testing
4. Known Safe (Low risk, high certainty) → Ignore

Assumption Card Template

┌─────────────────────────────────────────────────────────────────┐
│  ASSUMPTION: [Statement of what we believe to be true]          │
├─────────────────────────────────────────────────────────────────┤
│  Category: □ Desirability □ Viability □ Feasibility             │
│            □ Usability    □ Ethical                             │
├─────────────────────────────────────────────────────────────────┤
│  Risk Level:     □ High  □ Medium  □ Low                        │
│  Certainty:      □ High  □ Medium  □ Low                        │
├─────────────────────────────────────────────────────────────────┤
│  Evidence (current):                                            │
│  [What evidence do we have? None = "Leap of Faith"]             │
├─────────────────────────────────────────────────────────────────┤
│  Test Method:    [How could we test this?]                      │
│  Success Signal: [What would validate it?]                      │
│  Failure Signal: [What would invalidate it?]                    │
└─────────────────────────────────────────────────────────────────┘

Experiment Design

Experiment Types by Category

Category	Low-Effort Experiments	High-Effort Experiments
Desirability	Customer interviews, surveys, landing page	Concierge MVP, pilot program
Viability	Pricing surveys, market research	Pre-sales, crowdfunding
Feasibility	Technical spikes, prototypes	POC with real data
Usability	Paper prototype tests	Clickable prototype tests
Ethical	Expert review, user interviews	Pilot with monitoring

Experiment Card Template

┌─────────────────────────────────────────────────────────────────┐
│  EXPERIMENT: [Name]                                             │
├─────────────────────────────────────────────────────────────────┤
│  Assumption Being Tested:                                       │
│  [Link to assumption]                                           │
├─────────────────────────────────────────────────────────────────┤
│  Hypothesis:                                                    │
│  We believe [specific outcome] will happen because [reason].    │
├─────────────────────────────────────────────────────────────────┤
│  Experiment Method:                                             │
│  [Detailed description of what we'll do]                        │
├─────────────────────────────────────────────────────────────────┤
│  Metrics:                                                       │
│  • Success: [Threshold for validation]                          │
│  • Failure: [Threshold for invalidation]                        │
│  • Sample Size: [Number of data points needed]                  │
├─────────────────────────────────────────────────────────────────┤
│  Duration: [Time to run experiment]                             │
│  Cost: [Resources required]                                     │
├─────────────────────────────────────────────────────────────────┤
│  Result: □ Validated □ Invalidated □ Inconclusive               │
│  Evidence: [What we learned]                                    │
│  Next Step: [Decision based on result]                          │
└─────────────────────────────────────────────────────────────────┘

Writing Falsifiable Hypotheses

Good hypotheses are falsifiable - you can definitively prove them wrong.

✅ Falsifiable:

• "At least 40% of interviewed developers will mention code review as a top-3 pain point"
• "5% of landing page visitors will click 'Request Demo'"
• "Users can complete onboarding in under 5 minutes"

❌ Not Falsifiable:

• "Users will like our product"
• "The market wants this"
• "It will be successful"

Sample Size Guidelines

Experiment Type	Minimum Sample	Ideal Sample
Qualitative interviews	5-8	12-15
Usability tests	5	8-12
Landing page tests	100 visitors	1000+ visitors
Survey validation	50 responses	200+ responses
A/B tests	Depends on effect size	Use calculator

Common Experiment Patterns

1. Customer Problem Interviews

Tests: Desirability (problem exists)

Method:

1. Recruit 5-8 target users
2. Ask about past behavior (not future intent)
3. Explore specific recent experiences
4. Listen for pain intensity

Success Signal: 80%+ confirm problem exists and actively seek solutions Failure Signal: <50% recognize the problem as significant

2. Landing Page Test

Tests: Desirability (solution interest)

Method:

1. Create simple landing page describing solution
2. Include clear CTA (signup, request demo, etc.)
3. Drive traffic via ads or direct outreach
4. Measure conversion rate

Success Signal: 5%+ conversion to waitlist Failure Signal: <1% conversion after 200+ visitors

3. Fake Door Test

Tests: Desirability (feature interest)

Method:

1. Add button/link for non-existent feature
2. When clicked, show "coming soon" message
3. Offer to notify when available
4. Measure click rate and signup rate

Success Signal: 2%+ of users click, 30%+ of clickers signup Failure Signal: <0.5% click rate

4. Concierge MVP

Tests: Desirability + Viability

Method:

1. Manually deliver the service
2. Charge real money (or time commitment)
3. Observe customer behavior
4. Gather feedback

Success Signal: Customers pay, return, and refer Failure Signal: Customers churn quickly or won't pay

5. Wizard of Oz

Tests: Desirability + Usability

Method:

1. Create realistic front-end
2. Manually process behind the scenes
3. Users don't know it's manual
4. Measure engagement and satisfaction

Success Signal: Users complete tasks and express satisfaction Failure Signal: Users confused or abandon workflow

6. Technical Spike

Tests: Feasibility

Method:

1. Time-box exploration (1-3 days)
2. Build smallest possible proof of concept
3. Document unknowns discovered
4. Assess technical risk

Success Signal: Core mechanism works, risks understood Failure Signal: Fundamental blockers discovered

7. Pricing Survey

Tests: Viability

Method: Van Westendorp Price Sensitivity Meter

• At what price is this too cheap (suspect quality)?
• At what price is this a bargain?
• At what price is this getting expensive?
• At what price is this too expensive?

Success Signal: Acceptable price range aligns with business model Failure Signal: Willingness to pay below cost

8. Expert Review

Tests: Ethical + Feasibility

Method:

1. Engage domain experts / ethicists
2. Present product concept
3. Structured review against principles
4. Document concerns

Success Signal: No critical concerns, minor issues addressable Failure Signal: Fundamental ethical or regulatory issues

Experiment Prioritization

ICE Score

Rate each experiment 1-10 on:

• Impact: How much risk reduction if we learn the answer?
• Confidence: How confident are we in the experiment design?
• Ease: How easy/cheap is the experiment to run?

ICE Score = I × C × E ÷ 3

Prioritize highest ICE scores first.

Learning vs. Validation

Purpose	Approach	Sample Size	Rigor
Learning	Exploratory, qualitative	Small (5-8)	Lower
Validation	Confirmatory, quantitative	Larger (50+)	Higher

Start with learning experiments, then validate promising directions.

Results Documentation

After Each Experiment

Document:

1. What we tested: Assumption + hypothesis
2. What we did: Experiment method
3. What we learned: Results + insights
4. What we decided: Pivot, persevere, or more testing
5. What's next: Follow-up actions

Assumption Board

Track all assumptions and their status:

Assumption	Category	Risk	Status	Evidence	Decision
Users have problem	D	High	Validated	8/10 interviews	Proceed
Will pay $50/mo	V	High	Testing	Survey live	-
Can build in 4 weeks	F	Med	Validated	Spike done	Proceed
AI accuracy 90%+	F	High	Invalidated	POC: 75%	Pivot

AI-Assisted Assumption Testing

Assumption Identification

Given a product concept:

1. Generate comprehensive assumption list (DVFUE)
2. Categorize by type
3. Estimate risk and certainty
4. Prioritize for testing

Experiment Design

For each risky assumption:

1. Suggest appropriate experiment types
2. Draft falsifiable hypothesis
3. Define success/failure criteria
4. Estimate effort and timeline

Results Analysis

Given experiment data:

1. Assess against success criteria
2. Determine validation status
3. Identify follow-up questions
4. Recommend next steps

Integration Points

Inputs from:

• lean-startup skill: Leap of faith assumptions
• opportunity-mapping skill: Solution assumptions
• design-sprint skill: Prototype test results

Outputs to:

• Product decisions: Build/don't build
• lean-startup skill: Pivot/persevere recommendations
• Engineering backlog: Validated features

References

For additional Assumption Testing resources, see: