Two replica studies — a B2B channel survey (67 vs 149 synthetic) and a concept test (n=220) — scored on distribution distance. Worst question moved from TVD 0.881 to 0.113 after calibration.
144 US commercial-refrigeration distributors: reliability is the entry ticket, yet 88% will take on a new brand within a year. Five findings, five personas, and why the distributor rep decides.
AI synthesis is fast, but unsupported. Here is the discipline we use to keep every cluster traceable to clips, quotes, and segments.
Synthetic focus groups are not a replacement for participants. Here is the decision matrix we use across 200+ studies.
Across 14 SaaS pricing studies, the same micro-pattern: cursor stalls between tiers, then drifts to support.
A B2B SaaS team rebuilt onboarding off research evidence — every change tied back to a specific clip, transcript, and behavioral signal, auditable instead of trusted.
Source clip, transcript moment, behavioral signal, segment pattern, confidence — five layers that make a recommendation defensible.
Navigational dead-ends, cognitive load spikes, signposting gaps, trust failures, and mismatched expectations — each with a sample diagnostic.
Across 22 product studies, the latency between intent and action predicted churn better than any survey instrument.
Activation dashboards say onboarding is done. Research shows where users actually stall, recover, or quietly disengage — and how to close the gap between the metric and the felt experience.
The phrase 'enterprise plan' triggers a specific hesitation in mid-market buyers — opaque pricing, forced sales calls, the fear of being upsold. The recurring language, and what reduces it.
259 buyers, 5 findings, 5 personas. Comfort beats the logo for 92%, and 73% trust peers over big-brand ads — the data, the verbatims, the strategy.
Parents named the gentler of two screen properties as their favorite. The attention and durability signals — the things that actually drove willingness to pay — overturned the stated ranking.
A consumer brand ran several value-prop and messaging directions through research before launch. Most died on contact with real reactions; one survived and tested strongly.
n=112. 89 of 112 rated interest 8-plus, but faces stayed mild — and the top blocker was price ambiguity. Where attention held, where it dropped, and the said-versus-felt gap.
n=133 active retail investors. 93% say management quality is critical; about half actually check it. The say-do gaps that move confidence — and the language that closes them.
A national food brand A/B-tested two pre-roll cuts with video emotion capture (n=244). The predominant expression was flat — the real signal was a Surprise spike invisible in the aggregate.