Mobile Analytics
Mobile Analytics
Mobile Analytics tests Mixpanel implementation on iOS and Android: SDK lifecycle, automatic events, offline delivery, identity, device context, and mobile-specific analysis. The exam focuses on why mobile data differs from web data.
Core rules
- Mobile SDKs track events with a name and properties and attach device and app context such as OS, app version, and carrier where available.
- Automatic events use the
$ae_prefix and cover lifecycle moments such as first open, updates, and sessions. They are useful baselines but do not replace product events. - The SDK queues events offline and sends them in batches.
flushforces delivery; without it, backgrounding or killing the app can delay or lose queued events. - Call
identifywhen the user logs in. Usealiasonly for a genuine merge between an anonymous ID and a known ID, not as a routine login step. - Deep links and push notifications should carry campaign or source properties so acquisition and re-engagement are measurable.
- Mobile Session Replay can record app sessions when enabled, but sampling and privacy settings determine what is available.
Worked example
You want onboarding completion by app version. Track Onboarding Step Viewed with step and Onboarding Completed, then break down a funnel by $os and app_version. If the latest release shows a drop at step 3, compare event volume with automatic session events to separate a tracking regression from a real UX problem. The reasoning: version properties must be attached automatically by the SDK or consistently set, otherwise the breakdown is meaningless.
Traps
Tracking before identify and splitting the user; calling alias on every login; forgetting to flush before app termination; treating background/foreground cycles as clean sessions; ignoring ATT/IDFA limits on advertising identifiers; and assuming offline events arrive in real time.