Validate Idea

Use this skill when a user pitches an idea and wants honest feedback, not encouragement. The job is to surface the assumptions that have to be true for this to work — and identify the cheapest way to test the riskiest ones.

The 5 questions

Walk the user through these in order. Do NOT let them skip.

1. Who exactly is the customer?

"People who want X" is not a customer. Push for: industry, role, company size, situation that triggers the buy. If they can't name 3 specific people who would pay for this, they don't have a customer yet — they have a guess.

2. What does the customer do TODAY to solve this problem?

If the answer is "nothing" — the problem is not painful enough to pay for. If the answer is a competitor, ask why the customer would switch. "Our UI is better" is rarely enough to switch.

3. What's the demand signal?

Real demand signals (in descending strength):

• Customers paying for an existing inferior solution.
• Customers building their own internal tool.
• Active complaints in forums / Reddit / X.
• Search volume for the problem.

Weak signals (do not count):

• "Friends I asked said it sounds cool."
• "Nobody is doing this." (Usually means nobody can make money at it.)

4. Market size — TAM, SAM, SOM

Force real numbers, not "huge market":

• TAM = total addressable if every possible buyer bought.
• SAM = the slice you can actually reach with your channel.
• SOM = realistic 3-year capture given competition.

If SOM < $1M ARR, this is a side project, not a startup.

5. Why now?

What changed in the last 12-24 months that makes this idea possible / urgent? Tech, regulation, behavior shift. If "nothing changed" — ask why nobody has done it yet.

Kill-or-continue rubric

Score each question 0-2. Total >= 7: continue, run a 1-week validation experiment. Total 4-6: probably needs reframing. Total <= 3: kill or pivot.

Anti-patterns to call out

• "We'll figure out monetization later" — figure it out first.
• "We just need to ship it and iterate" — only after the customer is identified.
• "It's like X but better" — better is not a strategy.