How a verdict is made
No black box. Every number in your report comes from your testers' answers and a few simple rules you can check yourself.
Would buy
Testers answer on a five point scale from "definitely would not" to "definitely would buy". We count the top two answers. This is the classic top-two-box read.
Concept score
With several concepts: 40% buying intent, 30% liking, 30% first-pick share. With one: 55% intent, 45% liking. Highest score wins, and if two are within five points we say there's no clear winner.
Best price
For each price on your ladder we take the share of testers willing to pay it, then multiply by the price. The highest result is the price that earns the most per tester.
The verdict
Nothing is called under 20 responses. After that: 50% or more would buy is promising, 30 to 49% is iterate, under 30% is rethink. We show the margin of error too.
Build, Modify, Retest, Drop
Fewer than 20 answers: Retest. Top concept 50% or more would buy and most testers fit your target: Build (meaning "worth a small real-world test", not proof of demand). 20 to 49%: Modify, fixing the top hesitation. Under 20%: Drop. The Before Roast is one-liners written from your actual numbers, never a replacement for the analysis. No AI is connected yet, so questions and roasts come from built-in templates and rules.
The honest caveat
People overstate what they'll buy. These thresholds are rules of thumb, not promises. Use BEFORE to rule out weak ideas, choose between designs and set a price to test, then confirm with a pre-order or a small run.