How a national pest control company’s Android team went from finding out about issues two days late to knowing before their route managers do.

| 99.5% | ~0 | <70% | <5 sec |
|---|---|---|---|
| Crash-free last 90 days | Serious tickets per week | Crash-free at launch | Time to alert the team |
A leading Canadian pest control company working at national scale. More than 500 route managers use their Android app every day. They use it to manage jobs, routes, customer visits, and service records.
This is not a nice-to-have app. When it breaks, route managers can’t finish jobs. Customers don’t get served. Jobs get missed. Revenue takes a hit.
When the new Android app launched in early 2024, it was rough. The crash-free rate came in under 70%. The team jumped straight into firefighting mode, finding crashes, ranking them by how bad they were, and shipping fixes as fast as they could.
Over the months that followed, the team worked through the list. The numbers got better. But there was a second problem that fixing crashes alone couldn’t touch.
Even as the app got more stable, the alerting process was broken. There were no real-time crash alerts. No pipeline. Just silence until someone complained.
And when complaints came, they didn’t come quietly.
What happened every time a crash hit: