DRANSAY · LEADERSHIP

Mobile app: App Store rating from 3.3 to 4.8

Company
DrAnsay
Period
Nov 2025 to present
Result
3.3 to 4.8

Results

  • App Store rating: 3.3 to 4.8. The rating was 3.3 when I joined in November 2025 and stayed there until early 2026. It was 4.8 on the live listing in October 2026, and my notes place it at 4.8 from the third quarter of 2026.
  • Backend migration finished. The mobile app moved off Firebase and Cloud Functions onto the platform’s tRPC API. We started the migration in the first quarter of 2026 and finished it in the second.
  • Rebuilt app in gradual rollout. An external partner began a ground-up React Native rebuild in June 2026. The new app went through internal production testing from 30 September to 7 October 2026 and has been in gradual rollout since mid-October 2026.

Situation

I joined DrAnsay in November 2025 to manage two teams, one of them the mobile team. The app was built in Flutter and covered one of the marketplace’s use cases. It sat outside the React stack the rest of the platform runs on, and its backend ran on Firebase and Cloud Functions. Its rating was low for three reasons: constant crashes, slow performance, and how little attention the app had been getting.

A backend migration of the app to REST APIs had been prioritised for security reasons.

What I owned and what the team did

The rating rose from 3.3 to 4.8 over the same months in which we finished the backend migration.

I led the migration to tRPC and the mobile team carried it out. The team has three engineers today. I cancelled the earlier REST migration and planned and started its replacement.

I ran the React Native rebuild end to end. I wrote a map of the current product, with its user journeys and business rules, and a register of the traps the team had already hit, what had failed and what had worked. I got our product managers and team leads to write a specification for every feature before the partner built it. I designed the acceptance model the partner works against. I set up plans with our internal teams for the APIs, login, payments and tracking the new app needs, scheduled around their quarterly goals, and I run the weekly sync with the partner.

The partner builds the app. Our internal teams build the services it depends on. The mobile team keeps direction, acceptance and the integrations, and takes over the app itself at handover.

Decisions

Cancel the REST migration (first quarter of 2026). The agreed plan was not being followed, scope kept changing, complexity grew and the team was working without a clear picture of the end state. Costs and uncertainty were rising, and we were still discovering things in the legacy code as we went. The security concerns that had prompted the migration had already been addressed. I declared the project a failure and stopped it. Developers had put real time into it, and continuing because of that investment would have been the sunk-cost trap.

Restart on tRPC with a clear plan. I planned and started a migration to the platform’s tRPC API in the same quarter, and the team finished it in the next. The platform already had the endpoints the app needed, so the work was to write adapters and migrate.

Run the React Native rebuild with an external partner. Discontinuing the Flutter app was a company decision, not mine. Flutter sat outside the React ecosystem the rest of the platform runs on, and the app had to become a multi-service platform with prescription journeys. Using a partner meant the rebuild would not land on the mobile team’s roadmap. My decisions were about how to run it.

Accept features against their specification. Each feature is checked against its specification in four lanes: code, design, QA and behaviour. It is graded by its worst finding, with a stricter rule for regulated logic.

People and operating model

When a decision is hard, I tell the whole team at once and give the reasons, then follow up with each person individually. I did this for the REST cancellation and for the Flutter discontinuation. For the discontinuation the reasons had to be concrete, because the team had just finished the migration, so I also said what stayed with them.

Internal work for the rebuild is scheduled around each team’s quarterly goals. The partner works to a weekly sync, where I clear blockers before they turn into waiting time.

What I would do differently

I would do the tRPC migration first. It was the easier solution, and I would commit to it before anything larger, then align capacity and scope over time for the proper migration to REST API services. That is the order we are following with the new app.

All work