Voices
Notes from people who sat the grid
We keep the unflattering lines. A studio that only collects praise is running a marketing funnel, not a school. Names appear as the writer asked.
★★★★☆
“Cohort Signal Lab made our D30 chart look less heroic and more useful. The survival-curve session in week four is the piece I still quote in planning. I would not run the live studios again at 19:30 after a full on-call week — the recordings were fine, the pace on week three was not.”
Helen Crowe · Head of Product, Leeds
João, Lisbon (remote seat): the Event Grammar worksheets “stopped our Android and iOS streams using two words for the same checkout.” He asked for more examples from marketplace apps; we had mostly media and fintech that term.
A client in grocery delivery wrote that Funnel Forensics “caught a ghost step we had been reporting as ‘address confirmed’ when the API had already failed.” Attribution was listed as first name only: “Ravi, Birmingham.”
8.4 / 10 · alumni form, winter 2025
“Clear on what the programme will not cover. The dbt disclaimer saved us buying the wrong seat. Taxonomy Atelier notes were drier than the live room.”
Anonymous · client in subscription audio
Case note · a payments SDK, Glasgow
A three-person product trio arrived with 214 event names and no version field. After a Grid Fellowship quarter they shipped a 61-name grammar, killed a “session_ping” that had been inflating DAU, and took a two-page retention brief to their board. Install-to-second-open moved from 41% to 48% over eleven weeks. They were careful to say the SDK’s onboarding copy changed in the same window; we are not claiming the studio did that work alone.
The reservation they filed: private tutorials slipped twice because their engineer was in a freeze. We now ask Fellowship teams to name a deputy before week one.
Case note · a museum app, Edinburgh
The team sat Packet Studio on Funnel Forensics while preparing a new audio-guide flow. They learned the “start tour” event fired on prefetch, which made the funnel look generous. After the fix, the honest conversion was lower and the exhibition team finally funded a shorter intro. Visitor-services staff, not the analysts, noticed the tour completion rise the following month.
Eloise M., visitor experience lead, allowed her role and city to be printed. She declined a star rating: “It was a method, not a hotel.”