The Horizontal Validation Canvas, module 04

Seven questions under every claim.

They are the difference between a canvas that records your thinking and one that improves it.

The evidence layer

The interrogation never varies

Claim

What do we believe?

Basis

Why do we believe it?

Evidence

What have we actually observed?

Counter-evidence

What does not fit the claim?

Criticality

What happens if this is wrong?

Next test

What would resolve the uncertainty?

Threshold

What result would justify proceeding?

Two of the seven do most of the damage, usefully

The questions people skip

Counter-evidenceAn empty line rarely means nothing exists. Usually it means nobody looked, so say which: searched and none found, or not searched.
CriticalityWhat breaks if the claim is wrong. That single classification is what the gate will read later, so it is worth an argument now.

Criticality

Classified by consequence, not by confidence

Load-bearingThe model fails if this is wrong.
MaterialThe model would need significant modification.
SupportingImportant, but not decisive at the present stage.

Most claims feel load-bearing to their authors. Most are not.

At Halden Systems, an illustrative firm

One claim through the layer

Claim

Clients will pay 850 a month.

Basis

Six clients expressed interest at quarterly reviews.

Evidence

Anecdotal only, level 2. Interest is a sentence, not a behaviour.

Counter-evidence

Two clients declined a comparable add-on last year, at half the price.

Criticality

Load-bearing, without argument.

Next test

A full-price offer to the eleven clients whose insurance renews inside sixty days.

Threshold

Four signed, written before a single call.

The rule

The one that is load-bearing deserves your next test, not your next slide.

The next test is always the cheapest test of the most load-bearing unknown, not the easiest one.