Questions and beliefs

The evidence ladder, the challenger, and the riskiest belief

Updated 1 Nov 2026 · 2 min read

On this page
  1. Beliefs to test
  2. A belief's page
  3. The challenger
  4. Questions, and comparing segments

Learn › Questions is what you don't know yet, in one list with two views.

Beliefs to test

A belief is something that must be true for your product to work ("Solo designers chase late invoices every month"). Each starts at Belief, with no evidence, and sits on a board whose columns are the evidence ladder:

Praise (they said it) · Past pain (they described a specific time it hurt, unprompted) · Asks for access (they asked to get in) · Gives time · Makes an intro · Pays.

A card moves right only when someone did something. The rules are strict on purpose:

  • Compliments never move a belief past Praise. "I'd pay for that" is Praise.
  • A pain only counts as Past pain when it was unprompted: an answer to your own leading question is Praise.
  • Your own lines, and your teammates', never count.
  • A conversation can move a belief as far as Asks for access. Time and intros need something you log by hand ("from you"), and Pays is reached only through a test whose line you locked before it started.
  • Contradicting evidence is shown but never lowers a belief: calling it refuted is your decision.

The summary reads compliments against commitments, and one belief is flagged Riskiest (if it is false, the rest does not matter): Mind's pick, or yours. At most five beliefs can be open per objective. On day one, Mind can draft three beliefs from your product description (2 credits): accept the ones that sound right.

A belief's page

Its Evidence (all, unprompted, prompted, contradicting), each quote with who said it, where and when, and the question that prompted it, greyed ("your question, not counted"). That's wrong corrects how a quote counts, with a reason. History shows every change. Prepare drafts an interview script (1 credit: questions about past behaviour, one last question that asks for a commitment) and drafts invites for the three best people to ask (1 credit; each carries the consent-to-record line; you send them).

Problem, under About, is the problem the belief rests on. An editor picks it from the product's problems (or None), and the change shows in History. The people who raised that problem then show on Learn › People › Audience, ready to ask about the belief.

The challenger

When you accept a belief, start a test or decide while the evidence is thin, the challenger asks the one question that would most change your mind, quoting your own counts and quotes. Answer it, overrule it with a reason (always allowed), or Not now (it comes back on Friday). How often and how bluntly it asks is yours to set in Settings › You › How hard Mind pushes back (Gentle, Direct, Relentless).

Questions, and comparing segments

The Questions view lists what you would look up before a decision, by what each blocks. Research outside asks Mind to search public sources; what it finds is labelled public and never moves a belief. Compare segments (in the menu) sets two groups of customers side by side, row by row, never as one score; Pick logs the beachhead as a decision, and Run the deciding test drafts the test that would settle it.