| name | validating-ideas |
| description | Validates startup ideas through a sequenced workflow of Lean Canvas, Riskiest Assumption Testing, Mom Test interviews, and Pretotyping. Use when exploring a new product idea, pivoting an existing product, testing a B2B or B2C hypothesis, or when someone says "should we build X?"
|
Skill: Idea Validation
When to Use
Invoke this skill when:
- A new product idea or feature bet needs structured validation
- The team is considering a pivot or major direction change
- Someone asks "should we build X?" and the answer requires evidence, not opinion
- The company is evaluating a B2B infrastructure/platform play OR a B2C consumer product
- A strategic bet from the DIBB framework (see: shaping-product-strategy skill) needs ground-truth testing
Context Detection: B2B vs B2C
Before starting Phase 1, determine the idea's context by checking these signals:
B2B signals (API/platform infrastructure for business customers):
- The customer is a company (enterprise, SMB, agency, platform)
- Revenue comes from API usage, platform fees, or transaction volume
- Integration, compliance, and developer experience are central concerns
- Interview targets: CTOs, VP Engineering, Product Heads, Compliance Officers at target companies
B2C signals (consumer-facing product):
- The end user is an individual consumer or retail customer
- Revenue comes from consumer subscriptions, spreads, or direct transaction fees
- UX simplicity, trust, and onboarding friction are central concerns
- Interview targets: End users matching your ideal customer profile in [PRIMARY_MARKET]
Label the idea as B2B, B2C, or BOTH and carry this label through all phases. Templates below include variant sections for each.
Optional Pre-Filter: Kevin Hale (YC) Startup Ideas Framework
Before entering the main workflow, optionally run this 10-minute diagnostic to assess whether the idea is worth validating at all.
Score the idea on these five dimensions (1-5 each):
| Dimension | Question | Score |
|---|
| Problem frequency | How often does the target user face this problem? | [1-5] |
| Problem intensity | How painful is this problem when it occurs? | [1-5] |
| Willingness to pay | Would they pay to solve it today? | [1-5] |
| Founder-market fit | Does the founding team have unfair insight or access? | [1-5] |
| Market timing | Why now? What changed that makes this solvable today? | [1-5] |
Threshold: Total >= 18 to proceed. Below 15: challenge whether to validate at all. 15-17: proceed with caution, flag weak dimensions.
Phase 1: Lean Canvas (~1 hour)
Goal: Articulate the full hypothesis on one page so the team has a shared starting point.
Instructions: Fill in all 9 blocks. Be specific to your company's domain and market context. Write falsifiable statements, not aspirations.
Lean Canvas Template
LEAN CANVAS: [IDEA NAME]
Date: [DATE]
Author: [NAME]
Context: [B2B / B2C / BOTH]
1. PROBLEM (Top 3)
P1: [Specific problem statement -- who has it, how often, how painful]
P2: [Second problem]
P3: [Third problem]
Existing alternatives: [How do they solve it today? Name competitors or workarounds]
2. CUSTOMER SEGMENTS
Early adopter: [Specific profile -- e.g., "Mid-size companies with 50-500 employees
wanting to add [CAPABILITY] without building it in-house"]
Broader market: [Who comes after early adopters]
3. UNIQUE VALUE PROPOSITION
Single sentence: [Clear, jargon-free statement of why this is different]
High-level concept: [X for Y analogy -- e.g., "Stripe for [YOUR_DOMAIN]"]
4. SOLUTION (Map to each problem)
S1 -> P1: [How your product solves problem 1]
S2 -> P2: [How your product solves problem 2]
S3 -> P3: [How your product solves problem 3]
5. CHANNELS
Awareness: [How do prospects learn about this? -- e.g., industry conferences,
developer communities, LinkedIn, local community groups]
Evaluation: [How do they evaluate? -- e.g., sandbox, documentation,
case studies]
Conversion: [What triggers signup? -- e.g., demo call, self-serve API key]
6. REVENUE STREAMS
Model: [Transaction fee / platform fee / subscription / hybrid]
Pricing: [Specific numbers -- e.g., "0.2% per transaction + $2K/month platform fee"]
Target unit economics: [Average ARR per client: $__K]
7. COST STRUCTURE
Fixed: [Team, infrastructure, compliance costs]
Variable: [Provider fees per transaction, cloud costs, support costs]
Gross margin target: [__%]
8. KEY METRICS
North star: [Single metric that captures value delivery]
Leading indicators: [3-4 metrics that predict the north star]
Lagging indicators: [Revenue, churn, NPS]
9. UNFAIR ADVANTAGE
[What cannot be easily copied or bought? -- e.g., [FOUNDER]'s industry network,
proprietary aggregation layer, regulatory positioning, team's deep
domain experience]
Output: Save the completed canvas. Identify the 3-5 blocks where you are least confident -- these feed directly into Phase 2.
Phase 2: Riskiest Assumption Test (~30 minutes)
Goal: Identify the single assumption that, if wrong, kills the entire idea. Focus validation effort there.
Step 2.1: Extract Assumptions
Review the Lean Canvas and extract every implicit assumption. Categorize each one:
| # | Assumption | Category | Confidence | Evidence |
|---|
| A1 | [Statement phrased as testable belief] | Desirability / Viability / Feasibility | Low / Med / High | [What you know today] |
| A2 | ... | ... | ... | ... |
| A3 | ... | ... | ... | ... |
Categories defined:
- Desirability: Do customers want this? Will they use it?
- Viability: Can we make money? Does the business model work?
- Feasibility: Can we build it? Do we have the tech, team, and regulatory path?
Step 2.2: Rank by Risk
Plot assumptions on a 2x2 matrix:
HIGH importance (kills the idea if wrong)
|
|
VALIDATE | VALIDATE
SECOND | FIRST
(important | (important AND
but we have | we don't know)
some signal) |
HIGH -------------------|------------------- LOW
confidence | confidence
MONITOR | INVESTIGATE
(important | (uncertain but
and we know) | not fatal)
|
LOW importance
Step 2.3: Design the RAT Experiment
For the #1 riskiest assumption, design a test:
RAT EXPERIMENT DESIGN
Riskiest Assumption: [Statement]
Category: [Desirability / Viability / Feasibility]
Hypothesis: "We believe that [target user] will [expected behavior]
because [reasoning]."
Test Method: [Interview / Landing page / Fake door / Concierge / Prototype]
Sample Size: [N -- minimum viable for signal, typically 5-12 for interviews,
100-500 for landing pages]
Timeline: [X days/weeks]
Budget: [$X or $0 if pure interviews]
Pass Criteria: [Specific threshold -- e.g., "6 of 10 interviewees describe this
problem unprompted" or "3%+ landing page conversion"]
Fail Criteria: [What result means KILL or PIVOT -- e.g., "fewer than 3 of 10
mention the problem" or "<1% conversion"]
Owner: [Team member]
Start Date: [DATE]
Review Date: [DATE]
Output: One clearly defined experiment. Do NOT test multiple assumptions simultaneously.
Phase 3: Mom Test Interviews (~1-2 weeks)
Goal: Talk to real potential users/buyers without leading them. Extract truth about their problems, current behavior, and willingness to act.
Mom Test Core Rules
- Talk about THEIR life, not YOUR idea
- Ask about the PAST, not the FUTURE ("Tell me about the last time..." not "Would you use...")
- Ask about SPECIFICS, not GENERICS ("How much did that cost you?" not "Is cost a problem?")
- Listen for emotion, workarounds, and money already spent
- Never pitch. Never explain. Never seek validation.
Interview Guide: B2B Variant
Use when the idea targets businesses integrating your platform or API.
Target interviewees: CTO, VP Engineering, Head of Product, Compliance/Operations leads at companies matching your ideal customer profile.
| # | Question | What You're Learning |
|---|
| 1 | "Walk me through how you decided to add [CAPABILITY] to your product." | Purchase process, triggers, timeline |
| 2 | "What providers did you evaluate? What did that process look like?" | Competitive landscape from buyer's POV |
| 3 | "What was the hardest part of that integration?" | Pain points, feasibility concerns |
| 4 | "How long did it take from decision to live in production?" | Time-to-market pain (speed as value prop) |
| 5 | "If you could wave a magic wand and fix one thing about your current provider, what would it be?" | Unmet need, switching triggers |
| 6 | "How many different providers do you use for [DOMAIN] today?" | Multi-provider pain (aggregation thesis) |
| 7 | "Tell me about a time a compliance or regulatory issue blocked a product launch." | Regulatory friction as buying trigger |
| 8 | "What's your budget for [DOMAIN] infrastructure annually?" | Willingness to pay, budget validation |
| 9 | "If this problem disappeared tomorrow, what would you build next?" | Aspirational use case, expansion potential |
| 10 | "Who else in your org would need to approve a provider switch?" | Decision-making unit, sales cycle complexity |
Interview Guide: B2C Variant
Use when the idea targets individual consumers or end users.
Target interviewees: End users matching your ideal customer profile in [PRIMARY_MARKET].
| # | Question | What You're Learning |
|---|
| 1 | "Tell me about the last time you [DID RELEVANT ACTIVITY]. Walk me through the process." | Current behavior, tools used, friction points |
| 2 | "What apps or platforms do you use for [ACTIVITY] today?" | Competitive landscape from user's POV |
| 3 | "What's the most frustrating thing about [ACTIVITY] right now?" | Core pain, emotional intensity |
| 4 | "Have you ever wanted to [DO SOMETHING] but couldn't figure out how?" | Access gaps, unmet needs |
| 5 | "Tell me about a time you had a bad experience with [EXISTING SOLUTION]." | Trust barriers, UX pain |
| 6 | "How do you decide what [PRODUCT/SERVICE] to use? Who or what influences you?" | Discovery channels, trust signals |
| 7 | "What would make you switch from [current solution] to something new?" | Switching cost and triggers |
| 8 | "How much do you spend on [ACTIVITY] per month, roughly?" | Wallet size, unit economics viability |
| 9 | "Have you ever tried [ADJACENT SOLUTION]? What was that like?" | Adjacent product perception, education needs |
| 10 | "If a product could give you [KEY BENEFITS] in one place, what questions would you have?" | Objections, trust barriers, feature priority |
Interview Analysis Template
After completing 8-12 interviews, synthesize:
MOM TEST SYNTHESIS
Interviews completed: [N]
Date range: [START] to [END]
Interviewee profile: [Summary of who you talked to]
PATTERN 1: [Theme]
- Frequency: [X of N interviewees mentioned this]
- Representative quote: "[Direct quote]"
- Implication for hypothesis: [Supports / Challenges / Neutral]
PATTERN 2: [Theme]
- Frequency: [X of N]
- Representative quote: "[Direct quote]"
- Implication: [Supports / Challenges / Neutral]
PATTERN 3: [Theme]
- Frequency: [X of N]
- Representative quote: "[Direct quote]"
- Implication: [Supports / Challenges / Neutral]
SURPRISES (things you did NOT expect):
- [Finding 1]
- [Finding 2]
RISKIEST ASSUMPTION UPDATE:
- Original assumption: [From Phase 2]
- Interview evidence: [Supported / Challenged / Inconclusive]
- Confidence shift: [Low->Med / Med->High / etc.]
- Recommended action: [Proceed to Phase 4 / Pivot assumption / Re-interview]
Phase 4: Pretotype / Fake Door Test (~1-2 weeks)
Goal: Measure actual behavior, not stated intent. People lie in interviews (not maliciously); behavior reveals truth.
Step 4.1: Select Test Type
Choose based on what you are validating:
| Test Type | Best For | Cost | Timeline | Signal |
|---|
| Landing page + waitlist | Desirability, messaging | $200-500 ads | 5-7 days | Signup rate |
| Fake door (button in existing product) | Feature demand | $0 | 3-5 days | Click rate |
| Concierge (manual delivery) | Solution quality | $0 (time only) | 1-2 weeks | Retention, WTP |
| Explainer video + CTA | Complex product understanding | $500-1K | 7-10 days | Watch rate, CTA clicks |
| Wizard of Oz (fake automation) | Feasibility perception | $0 (time only) | 1-2 weeks | Completion rate |
Step 4.2: Design the Test
PRETOTYPE SPEC
Test type: [From table above]
What we're measuring: [Desirability / Viability / Feasibility]
Hypothesis: "If we show [target audience] a [test artifact], at least [X%]
will [desired action]."
TARGET AUDIENCE
- Profile: [Specific description]
- Source: [How you'll reach them -- ads, email list, community, direct outreach]
- Sample size target: [Minimum 100 for landing pages, 200+ for statistical
significance on conversion]
TEST ARTIFACT
- Landing page headline: "[Compelling headline that states the value prop]"
- Subheadline: "[One sentence elaborating the benefit]"
- CTA: "[Button text -- e.g., 'Get Early Access', 'Request API Key']"
- Social proof element: [If available -- logos, quotes, numbers]
TRAFFIC PLAN
- Channel: [LinkedIn ads / Google ads / direct email / community post]
- Budget: $[AMOUNT]
- Duration: [X days]
- Target impressions: [N]
SUCCESS CRITERIA
- Primary metric: [e.g., landing page -> CTA conversion rate]
- Pass: [>= X% -- benchmark: 3-5% for cold traffic, 10-15% for warm]
- Inconclusive: [X-Y% -- need more data or different audience]
- Fail: [< X% -- strong signal to pivot or kill]
ANTI-GAMING MEASURES
- [How you prevent false positives -- e.g., require email, not just clicks]
- [How you ensure traffic quality -- e.g., geographic targeting, interest filters]
Output: A fully specified test that can be launched within 48 hours.
Phase 5: GO / PIVOT / KILL Decision (~1 session)
Goal: Make an evidence-based decision. No gut feelings. Score every dimension and let the data speak.
GO/PIVOT/KILL Scorecard
Score each dimension 1-5 based on evidence gathered in Phases 1-4.
| # | Dimension | Score (1-5) | Evidence Summary | Key Data Point |
|---|
| 1 | Problem-Solution Fit -- Do real people have this problem and does our solution address it? | [SCORE] | [2-3 sentences from Mom Test + Pretotype data] | [Single strongest data point] |
| 2 | Willingness to Pay -- Have prospects indicated they would pay, or shown purchasing behavior? | [SCORE] | [Evidence of budget, current spend, stated WTP] | [Specific $$ or conversion %] |
| 3 | Market Size -- Is the addressable market large enough for the company's financial targets ($[X]M ARR Y1)? | [SCORE] | [TAM/SAM estimate with sources] | [SAM number] |
| 4 | Competitive Advantage -- Can we win against alternatives (including "do nothing" and "build in-house")? | [SCORE] | [Key differentiators validated in interviews] | [Most cited advantage] |
| 5 | Feasibility -- Can we build this with the current team and resources within 3-6 months? | [SCORE] | [Technical assessment, provider readiness, regulatory path] | [Estimated time to MVP] |
| 6 | Strategic Fit -- Does this align with the company's vision and long-term strategy? | [SCORE] | [Alignment with strategic plan, product roadmap, team strengths] | [Which OKR does this serve?] |
Scoring Guide
- 5: Strong evidence supporting (multiple data points, quantitative signal)
- 4: Moderate evidence supporting (qualitative signal, some data)
- 3: Mixed or insufficient evidence (need more data, conflicting signals)
- 2: Moderate evidence against (warning signs, weak demand signals)
- 1: Strong evidence against (clear rejection, no market, blocked path)
Decision Rules
| Total Score | Decision | Action |
|---|
| 25-30 | GO | Proceed to business model design. Hand off to designing-business-models skill. |
| 19-24 | CONDITIONAL GO | Proceed, but address dimensions scored <= 2 before investing heavily. Set kill criteria for 90 days. |
| 13-18 | PIVOT | The core insight has value but the current form does not work. Identify which dimension(s) to change and re-enter at Phase 1 with adjusted hypothesis. |
| 7-12 | KILL | Insufficient evidence across too many dimensions. Archive learnings and move on. |
| 6 | KILL | Nothing validated. Archive and move on immediately. |
Decision Record
DECISION: [GO / CONDITIONAL GO / PIVOT / KILL]
Idea: [Name]
Date: [DATE]
Total Score: [X/30]
Decided by: [Names]
Score Breakdown:
Problem-Solution Fit: [X] -- [One-line justification]
Willingness to Pay: [X] -- [One-line justification]
Market Size: [X] -- [One-line justification]
Competitive Advantage: [X] -- [One-line justification]
Feasibility: [X] -- [One-line justification]
Strategic Fit: [X] -- [One-line justification]
If GO: Next step is designing-business-models skill.
If CONDITIONAL GO: Must resolve [DIMENSION] within [TIMEFRAME].
Kill criteria: [What result in 90 days triggers a KILL]
If PIVOT: New hypothesis is [RESTATED HYPOTHESIS]. Re-enter Phase 1.
If KILL: Key learning to archive: [INSIGHT].
Save this record to your decision log.
Cross-References
| Decision | Next Skill |
|---|
| GO or CONDITIONAL GO | designing-business-models -- Build the Business Model Canvas, pricing architecture, and unit economics |
| Need strategic framing before validation | shaping-product-strategy -- Use DIBB framework to structure the bet before entering this workflow |
| Market-specific validation needed | planning-market-entry -- Assess regulatory readiness and market attractiveness for a specific geography |
Optional Post-Launch Add-On: Sean Ellis PMF Survey
Not part of core validation. Use only AFTER the product is live with real users (minimum 40 responses needed for signal).
The question: "How would you feel if you could no longer use [PRODUCT]?"
| Response | Threshold | Interpretation |
|---|
| "Very disappointed" | >= 40% | Strong PMF signal |
| "Somewhat disappointed" | -- | Moderate engagement, not PMF |
| "Not disappointed" | >= 40% | No PMF -- revisit value prop |
If below 40% "very disappointed": Segment responses by user type. PMF may exist in a subsegment. Narrow focus to that segment before expanding.
Checklist: Full Workflow at a Glance