Skip to content
AI Customer Research & ValidationLesson 9 of 13

9 · Reading the real signal (sign-ups and payments, not likes)

Now you have data coming in — survey answers, landing-page visits, comments, sign-ups. The skill is telling a real signal of demand from a vanity one that just feels good. Founders fool themselves by celebrating the metrics that flatter them and ignoring the ones that decide whether they have a business.

Weak signals (nice, but not proof):

  • Likes, follows, and "great idea!" comments. Approval is cheap and polite. It costs the person nothing, so it proves nothing about demand.
  • "I would totally use that." A hypothetical promise (Lesson 3) — words about a future that usually never happens.
  • An AI saying your numbers look promising. Not a customer. Not evidence.

Strong signals (cost something, so they count):

  • Email sign-ups for a real waitlist. A small cost (their attention + address) — a real, if early, vote.
  • Pre-orders and payments. The strongest signal short of a repeat sale. Money is the most honest survey there is. People reaching for their wallet is the willingness-to-pay that no persona can fake.
  • Repeat use and referrals. People coming back, or bringing a friend, is demand you can build on.

The rule of thumb: weight any signal by what it cost the person. A like costs nothing and means little. A payment costs real money and means a lot. The further someone goes — attention → email → money → repeat → referral — the more you can trust it.

Where AI helps (carefully): AI can tally and chart your results and compute simple rates (e.g., sign-ups ÷ visitors). Useful — but don't let a clean dashboard launder a weak signal. Automation bias (Lesson 6) makes an AI-generated chart feel like proof; the chart is only as good as whether the underlying action cost the person anything (NIST, 2022). A beautiful graph of likes is still a graph of likes.

Check yourself. Rank these by how much you'd trust them as proof of demand, and say why: 100 likes, 10 email sign-ups, 3 pre-orders that paid money.

Sources

Sign in to track your progress