Hypothesis card

Start from one sentence you believe about your product. You get a card that says what you believe, why it is risky, the cheapest test that could prove it wrong, what to measure and the number that decides it. The bar locks the day you start testing. When the results are in, the verdict — right, wrong or can’t tell — is counted in code from the bar, and Piko writes what you learned and the next belief to test.

On the Free planCanvasesStart free
Asks for
  • One sentence you believe about your product or its users
  • For closing: how many took part and how many showed what you measured, with your notes
Gives back
  • The belief written so it can be proven wrong, and why it is risky
  • The cheapest test that could prove it wrong, with two alternatives
  • A numeric bar that locks the day testing starts
  • A verdict counted from your results, what you learned, and the next belief to test
Plan
Free

What it gives you

We believe   PMs stop updating their roadmap mainly because each update takes over an hour.
Test         Interviews · 8 PMs who stopped in the last six months
Right if     at least 6 of 8 name the effort first          (locked 9 November)

Observed     3 of 8                                          Wrong.
Learned      Roadmaps go stale when nobody reads them (4 of 8).
Next belief  PMs keep a roadmap current when someone above them reads it weekly.
A closed card, shortened.

How it works

Most hypotheses are written so they cannot fail. “Users want a better onboarding” survives every result, so testing it teaches nothing. The card is checked before you see it: one belief rather than two, a result that could prove it wrong, a test that can actually see what is measured, a bar that is a number, and enough people for the test to mean something.

The commonest way a team “validates” everything is moving the goalposts after the results arrive. Here the bar locks when testing starts. You can still change it, but the change stays on the card for good, and the verdict is shown against the original bar as well as the new one.

The test is chosen for what it can see. A reason has to be heard, so a belief about why people do something gets interviews; whether many people would act gets a fake door, a concierge trial or the data you have. Swap it for another and the checks run again on the spot.

Who it is for

A PM about to build on an assumption nobody has checked, or one whose team keeps saying a test “validated” something without ever saying what would have counted as failing.

Where it stops

It cannot run the test for you, and the verdict is only as honest as the numbers you enter. Below the minimum for a test — five interviews, thirty survey answers, a hundred visitors to a fake door — it says can’t tell, whatever the numbers look like.

Questions people ask

What makes a good product hypothesis?

One belief, about named people, with a result that would prove it wrong and a number decided before the test. “PMs at mid-size software companies stop updating their roadmap within two months because each update takes over an hour” can fail; “users want a better roadmap” cannot.

Why does the bar lock?

Because the easiest way to pass a test is to decide what counts after seeing the results. Locking the number the day testing starts keeps the verdict honest. A change is still allowed, and the card keeps it for good with the verdict against both bars, so anyone reading it can see what happened.

Can a PikoPM survey fill in the results?

Yes. Link the survey to the card and say which answers count; the two numbers are then counted from its answers rather than typed in.

Run it on your own work

Hypothesis card needs an account, because it writes against what you give it and remembers what you have told it before. It runs on the Free plan. Or see what the plans cost first.

Hypothesis card — turn a belief into a test · PikoPM