Radio vs narrative
A missing sample is often coverage or battery, not disinterest. IoT App Analytics has to say which, with firmware as a dimension, before anyone writes a churn story.
Practice note
IoT App Analytics is not mobile analytics with extra charts.
It is the craft of measuring a connected product when the hardware, the companion app, and the cloud disagree — and teaching those desks to share identifiers anyway.
What we include
Event contracts across firmware and app. Activation that survives OTA. Retention windows that respect sleep and buffering. Ops uptime and product habit on the same keys. Experiment flags that will not brick a valve.
What we leave out
Clickstream theatre for marketing sites. Warehouse certification. Vendor bake-offs. Real-time feature stores for devices that speak twice a day. If a topic cannot be versioned, we do not visualise it in class.
Three tensions the studio keeps returning to
App screen vs device state
Companion opens are easy to count. Valve state is not. We treat the former as a subplot unless the product truly lives on the phone.
Thailand field time
Installer calendars, holiday weeks, and route coverage change cohort math. Templates from consumer SaaS posts rarely survive a week on the Eastern Seaboard.
Where to practise this
If ops and product already share identifiers and you only need vocabulary, start with Desk Notes. If two firmware lines are leaking a funnel, Fleet Workshop is the right weight. If the desks cannot agree what “online” means, Device Signal Studio is built for that argument.