1-2-3 Quickstart MOM Guide
CODITECT Standard — Customer Validation Interview System For: Any idea. Any industry. Any stage. From inception to validation.
What This Is
A complete system for running Mom Test customer validation interviews that produce truthful, behavioral data instead of polite lies. Three steps. One framework. Works for everything from a napkin sketch to a growth-stage pivot.
What you'll have after completing all 3 steps:
- A set of interview questions custom-fit to your idea that cannot produce false positives
- A predictable conversation flow that's fun for both parties
- A scoring system that separates real signal from noise
- Evidence-backed Go / Pivot / Kill decisions
Time investment: 4–6 hours to set up, then 15–20 minutes per conversation.
Step 1: DECOMPOSE Your Idea Into Testable Beliefs (1–2 hours)
Before you talk to a single person, break your idea into pieces you can actually test. Every idea rests on hidden assumptions. Most founders never surface them. You will.
1A. The Idea Strip-Down
Fill in this template. Be honest — nobody's watching.
MY IDEA IN ONE SENTENCE:
_____________________________________________________________
THE PROBLEM I THINK EXISTS:
_____________________________________________________________
WHO I THINK HAS THIS PROBLEM (be specific — role, industry, situation, NOT demographics):
_____________________________________________________________
WHAT THESE PEOPLE DO TODAY TO HANDLE IT (your real competitor):
_____________________________________________________________
WHY I THINK THEY'D SWITCH TO MY THING:
_____________________________________________________________
HOW MONEY WOULD FLOW (who pays, how much, how often):
_____________________________________________________________
1B. The Assumption Extraction
Now list everything that MUST be true for your idea to work. Force yourself to write at least 7.
ASSUMPTION 1 (Problem exists): _______________________________
ASSUMPTION 2 (Problem is severe): _______________________________
ASSUMPTION 3 (Segment is real): _______________________________
ASSUMPTION 4 (They'd pay): _______________________________
ASSUMPTION 5 (They'd switch): _______________________________
ASSUMPTION 6 (We can reach them): _______________________________
ASSUMPTION 7 (Timing is right): _______________________________
ASSUMPTION 8 (optional): _______________________________
ASSUMPTION 9 (optional): _______________________________
ASSUMPTION 10 (optional): _______________________________
1C. The Three Scary Questions
Pick the 3 assumptions that, if wrong, kill your entire idea. Convert each into a question you can ask a real person WITHOUT mentioning your idea.
The Rules (The Mom Test):
- Talk about THEIR life, not YOUR idea
- Ask about PAST specifics, not FUTURE hypotheticals
- Talk less, listen more
Conversion Formula:
BAD (fails Mom Test):
"Would you buy an app that tracks your quality deviations?"
↓ Mentions your product + asks about hypothetical future
GOOD (passes Mom Test):
"Walk me through what happened the last time you found a quality deviation.
What did you do? Who did you tell? How long did it take to resolve?"
↓ Their life + past specific + invites a story
Write Your Big 3:
SCARY QUESTION 1: _______________________________________________
Tests assumption: ___________________________________________
My idea is DEAD if they say: ________________________________
SCARY QUESTION 2: _______________________________________________
Tests assumption: ___________________________________________
My idea is DEAD if they say: ________________________________
SCARY QUESTION 3: _______________________________________________
Tests assumption: ___________________________________________
My idea is DEAD if they say: ________________________________
1D. Quick Compliance Check
Before moving to Step 2, run every question through this filter:
✓ Could I ask this without them knowing my idea? □ YES □ NO → rewrite
✓ Am I asking about something that already happened? □ YES □ NO → rewrite
✓ Will their answer be a story, not a single word? □ YES □ NO → rewrite
All three YES? You're ready for Step 2.
Step 2: BUILD Your Conversation Kit (1–2 hours)
Now construct the full interview using the 7-Layer Question Architecture. You won't use all layers in every conversation — but having them ready means you're never stuck.
2A. The 7 Layers
Layer 1 — World Discovery (Start here. Always.)
Open wide. Understand their universe before you zoom into your problem.
| Question | Purpose |
|---|---|
| "What are the biggest challenges you're dealing with right now in [domain]?" | See if your problem area is top-of-mind |
| "Walk me through a typical [day/week/sprint]. Where does your time go?" | Real priorities, not stated priorities |
| "What's changed about how you handle [domain] in the last year?" | Openness to change |
KEY RULE: If they don't mention your problem area unprompted, it's probably not their top priority. That's data. Don't force it.
Layer 2 — Problem Excavation (Only enter when THEY raise the topic)
| Question | Purpose |
|---|---|
| "Tell me about the last time [problem] happened." | Past-specific behavioral data |
| "What did you do about it?" | Current behavior = gold |
| "What did that cost you — in time, money, or stress?" | Pain quantification |
| "Why do you put up with that?" / "What are the implications?" | Severity test |
Layer 3 — Current Solution Mapping
| Question | Purpose |
|---|---|
| "How are you dealing with [problem] right now?" | Your real competitor |
| "What do you pay for that — in money and time?" | Price anchor |
| "What do you love about it? What do you hate?" | Switching cost + gaps |
| "Have you tried anything else before this?" | Solution-seeking history |
Layer 4 — Active Search Behavior (The Buyer-vs-Complainer Test)
| Question | Purpose |
|---|---|
| "Have you searched for a better way to handle this?" | Active intent |
| "What did you find? Why didn't those work?" | Differentiation opportunity |
| "What would have to be true for you to switch?" | Feature requirements |
THE DOUBLE-KNOT TEST: If they claim it's a huge problem but have never searched for a solution, they're not a buyer. They're a complainer. Move on.
Layer 5 — Decision Architecture (Essential for B2B)
| Question | Purpose |
|---|---|
| "Where does the money come from for something like this?" | Budget owner |
| "Walk me through how a purchase like this actually happens at your org." | Decision process |
| "Who else would need to be involved?" | Stakeholder map |
Layer 6 — Commitment Extraction (Never skip this)
End every conversation by asking for something concrete:
| Stage | Ask For |
|---|---|
| Very early | "Can we chat again in 2 weeks? I'd love to share what I'm learning." |
| Forming hypothesis | "Who else should I talk to? Would you make an intro?" |
| Have prototype | "Would you test this with your team for 2 weeks?" |
| Ready to sell | "Can we start a pilot on [date]?" |
THE COMMITMENT LADDER:
Time → Reputation → Effort → Money → Contract
(weakest) (strongest)
Track where every prospect sits. If they don't move up after 3 touchpoints, they're a zombie. Let them go.
Layer 7 — Network Expansion (Last question, every time)
| Question | Purpose |
|---|---|
| "Who else deals with this? Could you intro me?" | Next conversations |
| "Is there anything else I should have asked?" | Blind spot insurance |
2B. The Deflection Cheat Sheet
When conversations go sideways — and they will — use these:
| They Give You | You Say |
|---|---|
| A compliment ("That's a great idea!") | "Thanks — but honestly, I'm more interested in how YOU deal with [X] today. Walk me through it." |
| Fluff ("I would totally use that!") | "When's the last time [problem] actually came up? What did you do?" |
| A feature request ("You should add [X]!") | "Interesting — what's happening right now that makes [X] important to you?" |
| Silence / short answers | Count to 5. Wait. People fill silence with truth. |
| You want to pitch (you feel it rising) | STOP. Write down what you wanted to say. Ask another question instead. |
2C. The Conversation Flow (15–20 Minutes)
OPEN (2 min)
→ "I'm trying to understand [domain]. No pitch — just learning."
→ Layer 1: Wide-open world discovery
DIG (8-10 min)
→ If they mention your area: Layer 2 (Problem) → Layer 3 (Solutions)
→ If they don't: Probe once with "Where does [domain] fit?"
→ If still nothing: Graceful exit. Not your customer.
→ Layer 4: Search behavior test
TEST (3-5 min)
→ Layer 5: Decision architecture (B2B)
→ Push with "Why do you bother?" and "What are the implications?"
→ Scary Questions (the ones you prepped in Step 1)
CLOSE (2 min)
→ Layer 6: Commitment ask
→ Layer 7: "Who else? Anything I should have asked?"
→ "This was incredibly helpful. Thank you."
2D. Practice Before You Go Live
Before your first real conversation, role-play Script 2: The Polite Liar with a teammate.
Have them play a person who:
- Doesn't really have the problem
- Is too polite to say so
- Gives compliments and vague enthusiasm
- Says "I would totally buy that!"
Your job: Detect every false signal and redirect to behavioral facts.
If you can't catch the polite liar in practice, you won't catch them in the wild.
Step 3: EXECUTE, Score, and Decide (Ongoing)
3A. Run Your Conversations
Batch size: 5–10 conversations per batch. Pace: 3–5 conversations per person per week. Format: 2 people per conversation (interviewer + notetaker) whenever possible.
3B. Score Every Conversation
Immediately after each conversation (within 2 minutes), fill in:
┌─────────────────────────────────────────────────────┐
│ QUICK CAPTURE CARD │
│ │
│ Date: ________ Person: _________________________ │
│ Role: ________ Segment: ________________________ │
│ │
│ TOP 3 QUOTES (their exact words): │
│ 1. "____________________________________________" │
│ 2. "____________________________________________" │
│ 3. "____________________________________________" │
│ │
│ BEHAVIORAL FACTS (things they DID, not SAID): │
│ 🎯 ____________________________________________ │
│ 🎯 ____________________________________________ │
│ 🎯 ____________________________________________ │
│ │
│ COMMITMENT RECEIVED: │
│ □ None □ Time □ Reputation □ Effort □ Money │
│ Details: ____________________________________ │
│ │
│ BELIEF UPDATE: │
│ Before: ____________________________________ │
│ After: ____________________________________ │
│ │
│ NEXT: Talk to _____________ (intro? □ Y □ N) │
│ │
│ QUALITY SCORE: │
│ Listen ratio: ___% them Mom Test compliance: __/5 │
│ Tier 1 facts: ___ Zombie? □ Yes □ No │
└─────────────────────────────────────────────────────┘
3C. Classify Your Data
Every data point from every conversation goes into one of three tiers:
| Tier | Definition | Examples | Trust Level |
|---|---|---|---|
| Tier 1: Behavioral Facts | Things they actually DID | "We spent $50K on consultants last quarter for audit prep." / "I searched for QMS tools on G2 last month." | HIGH — use for decisions |
| Tier 2: Contextual Signals | Emotional reactions, stated priorities, comparative judgments | "That process drives me crazy." / "It's our #2 priority this year." | MEDIUM — use for direction |
| Tier 3: Noise | Compliments, hypotheticals, feature requests, fluff | "That sounds cool!" / "I would totally use that." / "You should add AI." | LOW — discard |
The rule: No decisions based on Tier 3 data. Ever.
3D. Run the Batch Review (90 Minutes, Full Team)
After every 5–10 conversations:
Part 1 — Pattern Detection (30 min):
- Read all capture cards aloud
- What themes repeat (≥3 mentions)?
- What contradictions exist?
- What surprised everyone?
Part 2 — Belief Update (30 min):
- For each Scary Question: what do we now believe?
- Which assumptions survived? Which died?
- What new assumptions emerged?
Part 3 — Decision (30 min):
| Signal | Decision |
|---|---|
| 3+ people describe the same behavioral pattern around the problem AND 30%+ show commitment advancement | Continue — you're onto something |
| Mixed signals, some strong patterns but contradictions | Narrow segment — you're talking to too broad a group |
| Problem exists but nobody's tried to solve it or willing to commit | Pivot — either the problem isn't severe enough or your segment is wrong |
| After 15+ conversations, no consistent pattern and no commitment | Kill the idea — this isn't a real problem worth solving |
3E. The Validation Scorecard
Track these numbers across your entire batch:
| Metric | Target | Your Number |
|---|---|---|
| Total conversations this batch | 5–10 | _____ |
| % that mentioned problem unprompted | ≥ 50% | _____ |
| Average Tier 1 facts per conversation | ≥ 3 | _____ |
| % with commitment advancement | ≥ 30% | _____ |
| Zombie leads detected | Track it | _____ |
| New contacts generated | ≥ 1 per convo | _____ |
| Big 3 Questions updated? | Yes/No | _____ |
| Team members who participated | All core | _____ |
| Go/Pivot/Kill decision made | Yes/No | _____ |
The AI Acceleration Layer (After 15+ Manual Conversations)
Once your team has internalized the methodology through manual practice, accelerate with AI tools:
Phase 2: AI-Assisted (Conversations 15–50)
| Tool Category | Examples | Use For |
|---|---|---|
| Transcription | Otter.ai, Fireflies | Auto-capture conversation text |
| Question Generation | Claude, GPT + Prompt Engine | Generate and refine question sets |
| Note Analysis | Dovetail, Marvin | Tag and cluster insights across conversations |
| Pattern Detection | Claude analysis | Cross-conversation pattern synthesis |
Phase 3: AI-Scaled (Conversations 50+)
| Platform | What It Does |
|---|---|
| Prelaunch AI Interviewer | Runs Mom Test-aligned AI conversations at scale |
| Marvin AI Moderator | Voice AI conducting adaptive interviews |
| Cusmos AI | Structured AI research conversations |
CRITICAL GATE: Do NOT skip to Phase 3 without completing Phase 1. AI amplifies your methodology — good or bad. If your questions are wrong, AI will scale wrong faster.
Using the CODITECT Prompt Engine
To generate custom interview questions for ANY idea, use the prompts in idea-to-questions-prompt-engine.md. The one-shot version:
- Describe your idea to an AI assistant
- Tell it your target customer
- Tell it what you're most uncertain about
- Ask it to run the full CODITECT validation pipeline
The AI produces: assumption decomposition → scary questions → 30-45 question library → conversation flow → 3 practice dialogs → anti-pattern audit → interviewer prep card.
Review the output with your team. Customize for your domain. Then go talk to humans.
Quick Reference: The Mom Test in 30 Seconds
RULES:
1. Their life, not your idea
2. Past specifics, not future hypotheticals
3. Listen more, talk less
ASK ABOUT: NEVER ASK:
✓ "Tell me about last time…" ✗ "Do you think it's a good idea?"
✓ "What did you do?" ✗ "Would you buy this?"
✓ "How much did that cost?" ✗ "How much would you pay?"
✓ "Have you searched for…?" ✗ "Do you like this feature?"
✓ "Who else should I ask?" ✗ "What features would you want?"
AFTER EVERY CONVERSATION:
□ Captured 3+ Tier 1 behavioral facts
□ Asked for a commitment
□ Got a referral
□ Updated my beliefs
Companion Documents
| Document | What It Contains |
|---|---|
coditect-mom-test-question-framework.md | Complete 7-layer question architecture with industry adaptations |
idea-to-questions-prompt-engine.md | 8 AI prompts that convert any idea into interview scripts |
interactive-interview-playbook.md | Dialog scripts, deflection toolkit, and fun-factor techniques |
the-mom-test-full.md | Full text of The Mom Test by Rob Fitzpatrick (reference) |
CODITECT 1-2-3 Quickstart MOM Guide v1.0 Three steps. Any idea. Any industry. From inception to validation.