The shape of this problem
Too much arrives to check it all, so you sample
The problem
Product teams needed to understand mobile-store feedback without reading every review manually. App-store reviews arrive continuously and unstructured, mixing one-off complaints with the early signal of a real regression. Read manually, they get read when someone has time - which is rarely the week a release actually broke something, so the feedback loop closed too late to influence the release that caused the problem.
What we built
We built a service that collected App Store and Play Store feedback, classified sentiment, grouped related issues, and presented the trends through a dashboard.
What moved
| Measure | Before | After |
|---|---|---|
| Lag from a review appearing to a grouped, scored issue reaching the product team | ~3 weeks | under 24 hours |
| Review coverage | a read sample | 100% of App Store and Play Store feedback |
| Differently-worded reviews correctly consolidated to one issue | not measured | 87% |
Lag from a review appearing to a grouped, scored issue reaching the product team
Differently-worded reviews correctly consolidated to one issue
87%
Differently-worded reviews correctly consolidated to one issue
The ring is the whole of that measure. The grey arc is the other 13%.
Figures are drawn from the practice's own delivery records for the engagement named, measured against the process that preceded it. They have not been through third-party audit, and none is presented as an average across clients.
What it turned on
Grouping issues rather than scoring sentiment alone is what made the output actionable - a product team needs to know which specific thing twenty people are describing in twenty different ways, not that satisfaction moved. Relevant to any team whose release feedback currently arrives too late to act on.
- Service line
- Report & Reconciliation Automation
Partial. No case yet where a specific recurring finance report was replaced end to end. The Gulf fuel-retail operator shape - continuous unstructured input becoming a grouped, scored dashboard - is the honest analogue, and should be offered as one.
Start here
Which of the four is yours?
Tell us the documents and the monthly volume and we will send the two closest records, with the proof behind each and the honest note where the match is partial.
You get a reply within one working day, from the engineer who would do the work - not a sales sequence.