The Testing Framework: Separate Ego from Evidence

1) Start with the riskiest assumption

• Don’t test everything. Test the one thing that, if false, kills the business.

• Example: Airbnb’s core risk—“Will strangers host strangers?”

• Do this:

  • List assumptions → rank by risk/uncertainty → test #1 first.

2) Build a minimum testable hypothesis (MTH)

Before an MVP, validate with the smallest possible test:

• Landing page with “Get Early Access”

• Manual “concierge” version of the service

• Clickable prototype (looks real, no backend)

• Short survey or interviews to confirm the problem

• Example: Dropbox validated demand with a simple video.

  
  

3) Strong opinions, weakly held

• Argue your hypothesis.

• If user data contradicts it, change your mind fast.

• Think like a scientist: being wrong = progress.

4) Make feedback a ritual

• Weekly: 3 user interviews

• Bi-weekly: usability tests (watch, don’t help)

• Monthly: cohort metrics (retention, engagement, conversion)

• Quarterly: assumption review (what changed and why)

  
  

The Founder’s Emotional Toolkit

Reframe criticism

• “This isn’t for me” = market data.

• Try a “rejection journal”: log negative feedback → note what you learned.

Separate identity from idea

• You ≠ your startup.

• Say: “That idea/version didn’t work,” not “I failed.”

Get an objective sounding board

• Advisor/mentor/founder who can challenge your assumptions.

• Give explicit permission to be blunt.

Celebrate pivots

• A pivot is a learning win: “We saved 6 months by learning this now.”

The Counterintuitive Truth

• The best founders love the problem, not any single solution.

• Commit to the mission; stay flexible on the path.

• As Reid Hoffman says: if you’re not embarrassed by v1, you launched too late. Early imperfection → feedback → improvement.

"Create a pic of one being afraid taking steps"

Your first step (tomorrow):

1. Show your product to a true target user (not a friendly insider).

2. Don’t explain first. Watch them use it.

3. Take notes where they hesitate or get confused.

4. Turn those notes into your next test.

Conclusion: Love the problem

• Be loyal to user needs, not your assumptions.

• Markets test ideas anyway—do it early when it’s cheap.

• Ship → watch → listen (no defending) → iterate.