Case studies
Same instrument, opposite prescriptions. Each engagement opens with a score pair and closes with one, because that is the only claim worth making, and the “what did not work” section is never omitted.
Clients are identified by industry only; names are withheld and surfaces simplified.
B2B SaaS · reconciliation
Reconciliation was costing nine minutes a batch and nobody had ever complained about it.
Before
2.9@0.81After · 9 weeks
3.6@0.87Delta
+0.7
Ecommerce · checkout trust
A checkout that worked, and a checkout nobody trusted. The planned redesign was cancelled; copy shipped instead.
Before
3.1@0.74After · 7 weeks
3.4@0.79Delta
+0.3
Media · activation
Everyone knew onboarding was broken. The research budget went to the thing nobody knew: where readers actually stalled.
Before
2.6@0.68After · 10 weeks
3.2@0.83Delta
+0.6
HR platform · time off and onboarding
Two populations on the same screens, in opposite quadrants. The four per cent who had been asked were describing a different product.
Before
2.7@0.72After · 11 weeks
3.5@0.86Delta
+0.8
Next up
This site is scoring itself in public. The pre-launch baseline is described in Scoring our own product in public; the rescore against real behavioural data will be published here as a further case.