Validation Playbook

Core Rule

Never build before you validate. Talk to 10 customers before writing 1 line of code. A week of conversations saves months of wasted development.

Assumption Mapping

Before validating, surface every assumption that could kill the business. Then rank them.

The 5 Critical Assumptions

1. Customer:  Who exactly is this for? Can I name 10 specific people?
2. Problem:   Do they have this problem badly enough to pay? Or just complain?
3. Solution:  Does our approach actually solve it better than alternatives?
4. Channel:   Can we reach them at a viable cost? Where are they?
5. Revenue:   Will they pay? How much? Monthly or one-time?

Assumption Risk Matrix

For each assumption, score two dimensions:

                    HIGH IMPACT IF WRONG
                    |
    Test these      |  TEST THESE FIRST
    second          |  (highest priority)
                    |
LOW UNCERTAINTY ----+---- HIGH UNCERTAINTY
                    |
    Skip these      |  Monitor these
    (safe enough)   |  (uncertain but survivable)
                    |
                    LOW IMPACT IF WRONG

Always test the upper-right quadrant first — high uncertainty AND high impact. If that assumption is wrong, nothing else matters.

Assumption Log Template

ASSUMPTION LOG
==============
Date: [DATE]

#  | Assumption                    | Risk (H/M/L) | Test             | Result    | Action
1  | [ICP] has [problem]           | High          | 10 interviews    | [RESULT]  | [ACTION]
2  | They'll pay $[X]/mo           | High          | Pre-sale offer   | [RESULT]  | [ACTION]
3  | We can reach them via [channel]| Medium       | 50 outreach msgs | [RESULT]  | [ACTION]
4  | [Solution] solves it          | Medium        | Prototype test   | [RESULT]  | [ACTION]
5  | Market is large enough        | Low           | TAM analysis     | [RESULT]  | [ACTION]

Customer Discovery Framework

The Mom Test (Rob Fitzpatrick)

The core rule: Ask about their life, not your idea. People will lie to be nice. Design questions that prevent it.

Bad questions (leading, hypothetical):

• "Would you use this?" → They'll say yes to be polite
• "Do you think this is a good idea?" → Their opinion doesn't predict behavior
• "Would you pay $50/month for this?" → Hypothetical money isn't real
• "Don't you hate when [problem]?" → Leading question invites agreement

Good questions (behavioral, specific, past-tense):

• "Walk me through the last time you dealt with [problem]."
• "What did you actually do about it?"
• "How much did that cost you — in time, money, or frustration?"
• "Have you looked for solutions? What did you find? Why didn't you buy it?"
• "What's the most frustrating part of how you handle this today?"
• "If you could wave a magic wand, what would change?"

The Mom Test Golden Rules:

1. Talk about their life, not your idea
2. Ask about specifics in the past, not generics about the future
3. Talk less, listen more — if you're talking >30% of the time, you're doing it wrong
4. Get commitments, not compliments — "interesting" means nothing; "can I pre-order?" means everything
5. Bad news is good news early — finding out the idea is wrong at interview #5 saves you 6 months

Interview Structure (20-30 min)

SETUP (2 min):
  "Thanks for taking the time. I'm researching [problem area].
   No pitch — just trying to learn from people who deal with this."

PROBLEM EXPLORATION (12-15 min):
  1. "Tell me about your role and what a typical week looks like."
  2. "When was the last time you dealt with [problem area]?"
  3. "Walk me through what happened."
  4. "What was the hardest part?"
  5. "How are you solving it today?"
  6. "How much does that cost you — time, money, or missed opportunities?"

BEHAVIOR & ALTERNATIVES (5-8 min):
  7. "Have you looked for a better solution? What did you find?"
  8. "Why didn't you buy [existing alternative]?"
  9. "What would have to be true for you to switch from your current approach?"

CLOSE (3 min):
  10. "Who else deals with this that I should talk to?"
  11. "Can I follow up in a few weeks with what we've learned?"

After every interview — debrief within 10 minutes:

• Top 3 insights
• Best direct quotes (verbatim — these are gold)
• Did this confirm or challenge my assumptions?
• What should I ask differently next time?

Validation Tests (Ordered by Effort)

Tier 1: Low Effort (1-3 days)

Smoke Test Landing Page

• Build a simple page describing the value proposition
• Include a clear CTA ("Join waitlist" / "Get early access")
• Drive traffic from relevant communities
• Validates: Does the positioning resonate? Will people opt in?
• Green signal: >5% visitor-to-signup conversion
• Red signal: <1% conversion or zero signups after 200+ visitors

Fake Door / Feature Test

• Add a button or link for a feature you haven't built
• Track clicks. Show "coming soon" message on click.
• Validates: Do users want this specific feature?
• Green signal: >10% click rate from users who see the button

Community Validation

• Post your problem statement (not solution) in 3-5 relevant communities
• Track responses, DMs, "me too" reactions
• Validates: Do real people in the wild confirm the problem?
• Green signal: Strangers DM you asking for the solution

Tier 2: Medium Effort (1-2 weeks)

Pre-Sell Offer

• Describe the product. Set a price. Ask for payment before building.
• Offer early-bird pricing (30-50% off) and a delivery date
• Validates: Will people pay real money before the product exists?
• Green signal: 3+ pre-sales from people who aren't friends/family
• See: first-revenue.md Path 1 for the full pre-sale playbook

Concierge MVP

• Deliver the result manually to 3-5 customers
• Charge full price. Track your time per customer.
• Validates: Is the outcome valuable? Can you deliver it?
• Green signal: Customers want to continue after the manual period
• See: first-revenue.md Path 5 for the services wrapper approach

Wizard of Oz MVP

• Front-end looks automated. Back-end is you doing it manually.
• Customer doesn't know (or doesn't care) that it's not automated.
• Validates: Does the experience work? Would automation be worth building?
• Green signal: Users treat it as a real product and get real value

Tier 3: High Effort (2-4 weeks)

Working Prototype

• Build the minimum possible version that tests the core hypothesis
• Not feature-complete — just the one thing that matters most
• Validates: Does the technology solve the problem?
• Green signal: Users complete the core workflow without hand-holding

Paid Pilot

• 5-10 customers at 50-70% of target price for 30-90 days
• Weekly check-ins. Structured feedback.
• Validates: Will they pay, use it, and stay?
• Green signal: >60% convert to full price at pilot end
• See: first-revenue.md Path 3 for the full pilot playbook

Minimum Viable Tests by Business Type

Different business models need different validation approaches:

PMF Signals

How to Know You Have Product-Market Fit

The Sean Ellis Test

Ask existing users: "How would you feel if you could no longer use [product]?"

• Very disappointed
• Somewhat disappointed
• Not disappointed
• I no longer use [product]

>40% "very disappointed" = PMF. Below that, keep iterating.

When to run it: After you have 30+ active users who have used the product at least twice. Don't run it on day 1 — wait until they've experienced the core value.

When You DON'T Have PMF Yet

No PMF signals: High acquisition but high churn. Users sign up but don't stay. This means your marketing works but your product doesn't.

What to do:

1. Stop spending on acquisition — it's waste until retention improves
2. Interview churned users: "What would have made you stay?"
3. Interview power users: "What keeps you coming back?"
4. Double down on whatever the power users love
5. Kill features that nobody uses

Jobs To Be Done (JTBD)

Frame every product decision around the job the customer is hiring your product to do.

When [situation], I want to [motivation], so I can [outcome].

Examples:

• "When I'm trying to track co-parenting communications for court, I want a neutral documentation tool, so I can protect my child and my legal position."
• "When I'm onboarding a new enterprise customer, I want automated setup workflows, so I can reduce time-to-value from 2 weeks to 2 days."
• "When I'm preparing for an investor meeting, I want a binder readiness score, so I can prioritize what to fix before the meeting."

Define your top 3 JTBD statements before building anything. Every feature should map to a job.

Validation Sprint (1 Week)

Friday Decision Framework:

Kill Criteria

Consider killing or pivoting if:

• Fewer than 3 of 10 interviews confirm the problem is painful
• No one has tried to solve it themselves (low pain = low willingness to pay)
• They say "yes" but won't pre-pay, join a waitlist, or give a referral
• The willingness-to-pay is below your cost to serve
• You've tested 3 different audiences and none care enough
• After 20+ interviews, you still can't describe the ICP in one sentence

Pivoting is not failing. It's the whole point of validation — to learn what works before you invest months building the wrong thing.

See also: resilience.md for the emotional side of pivoting | first-revenue.md for monetizing what you've learned | customer-discovery.md for detailed interview scripts

> Disclaimer: This playbook provides educational frameworks for startup validation. Customer discovery methods should be adapted to your specific industry and context. This is not professional business advice.