How We Ensure Credible Analytics on Dropbox Mobile Apps
Article Summary
Marianna Budnikova from Dropbox reveals why your mobile analytics are probably lying to you. And it's costing you two weeks every time you get it wrong.
Dropbox Mobile invested in fixing their analytics integrity problem after realizing last-minute, poorly-tested analytics were leading to faulty business decisions. The team developed a systematic approach combining intentional design, snapshot testing, and stakeholder collaboration to ensure their data actually tells the truth.
Key Takeaways
- Three pillars: Intentional (answers real questions), Credible (well-tested), Discoverable (easy to use)
- Snapshot testing catches analytics bugs by comparing JSON outputs in UI tests
- Found double-logged events and unusable localized error messages across 22 languages
- Biweekly releases mean analytics mistakes cost two weeks to fix
By treating analytics as code with snapshot tests and engaging stakeholders early, Dropbox eliminated double-logging bugs and made their mobile data trustworthy enough to drive decisions.
About This Article
Dropbox Mobile's analytics code had poor communication between engineers and product stakeholders. This led to implementations that were rushed and poorly designed, and they became unreliable as new features shipped without proper test coverage.
Budnikova's team added snapshot testing to their UI tests. They serialized analytics events into JSON and compared them against expected outputs. This caught field tagging changes and kept analytics accurate during complex user workflows across multiple screens.
The snapshot testing found specific bugs, like navigation events being logged twice. It also revealed that sign-in error messages in 22 languages were broken because they needed enum-based error state mapping. This helped stakeholders focus on problems that actually mattered.