Client stories
These notes come from findings sessions and first Mondays after a wall went live. Names of titles are withheld; the failures are not.
The weekly packs all looked calm. Aisha’s heat table put the shared scanner library in its own row. That was the first time operations and product sat in the same argument, instead of trading crash-free percentages.
Operations lead, Peninsular logistics group · seven live titles · Portfolio Health Review
We commissioned the wall after the review. Lim’s first paper draft was stricter than our old fourteen-chart deck, which was the point. I still think the first workshop ran long — we lost the last hour to a join-key dispute that should have been settled before we arrived — but the Monday view is the one ops actually opens.
Head of digital, retail group with catalogue and loyalty apps · Health Dashboard Commission
The briefing is useful because it has a memory. In Q1 they ranked our homework app above the parent title. In Q2 the parent slipped after a forced update. Nobody could pretend the two readings were unrelated once the heat table sat side by side.
Founder, education apps · Quarterly Health Briefing
Farah would not let us keep ‘user error’ as a closed incident when the tickets clustered on one version. Annoying in the room. Correct in the brief.
Product owner, staff rostering app · Review Room Workshop, then a health review
The group ran driver, warehouse, customer-tracking, two white-label retailer apps, a scanner companion, and an internal roster. Each title had its own vendor console. Crash-free sat above 99.4% everywhere except the scanner, which was blamed on ‘rugged devices’.
The document pack showed the scanner library embedded in four of the seven titles. Time-to-first-failure after a library bump was measured in hours on the warehouse app and in days on the customer-tracking app, because warehouse sessions are short and frequent. Ranking the scanner alone had hidden the library.
The findings session spent its second hour on who owned that library. It was a vendor fork. The brief named the fork, the version window, and the four titles that would move together. The group did not ask us to build a wall until the library owner was written into the release calendar.
We would not have seen this in a single-app crash review. The work was the joint.
A catalogue and a stamp card shared a login. Retention slipped in both after a March library change. Each pack blamed the other directorate. We asked for identity events rather than screen events. The drop-off sat between credential accepted and session handed to the title — an interval neither weekly pack contained.
The heat table grew a shared row. The library group joined the second hour. They had only seen their own sampled crash-free figure. The brief did not rank one consumer app above the other on a failure neither team could ship.