Privacy-First Mobile Telemetry
Design an event ledger that answers activation and retention questions without storing identifiers you cannot defend.
View syllabusProxy Servicehub · Bangkok studio
Measure what the product does — not who the person is.
We teach product and engineering teams to collect the smallest event graph that still answers a roadmap question. Privacy-First App Analytics is the craft: first-party schemas, on-device aggregation, and metrics that survive a PDPA review without turning the app into a surveillance layer.
73
product teams completed a cohort since 2022
11 weeks
flagship studio with critique, not slide dumps
2 blocks
instructor office hours each teaching week
Programmes
Studios currently in rotation
Each programme is built around a shipping artefact: a schema, a consent map, or a reduced SDK. The flagship is the only live cohort with a full public syllabus.
Cookieless Funnel Architecture
Rebuild web-to-app funnels with first-party keys and server-side joins instead of cross-site trackers.
Ask for dates
PDPA Analytics for Thai Teams
Map purpose limitation, retention windows, and vendor roles onto the metrics your PM already reports.
EnquireHow we teach
Critique rooms, not content libraries
Learners arrive with a live app or a sandbox. Each week they submit a schema diff, a consent copy pass, or a sampling plan. Faculty mark the artefact against a short rubric: necessity, reversibility, and whether a lawyer can explain the purpose in one sentence.
We do not pretend that every metric can be kept. If a question only works with a persistent device graph, we ask you to drop the question or redesign the product surface. That discipline is the point of Privacy-First App Analytics at Proxy Servicehub.
What you leave with
A thinner, more honest measurement stack
- An event dictionary you can defend Names, triggers, and retention for every property — including the ones you decided not to collect.
- Consent that matches the code path Copy, SDK flags, and server gates stay in one map so a reject button actually stops downstream jobs.
- Aggregates that still move a roadmap Daily active, activation, and cohort curves built from hashed or on-device rolls instead of user tables.
From the studio
Notes from people who sat the work
“The on-device aggregation week forced us to kill three vanity events we had been reporting to the board. Painful, but the remaining activation curve finally matched what support was seeing.”
We took Privacy-First Mobile Telemetry while rewriting our Android SDK. The consent map exercise was clearer than any vendor doc. I still wish the sampling homework had used our actual traffic shape instead of the studio dataset — we had to redo the thresholds after week six.
Journal
Recent field notes
First-party events without fingerprints
How to name and store product actions when you refuse device graphs and probabilistic identity.
PDPA retention windows for app metrics
Choosing deletion clocks that still let a PM compare this month with last quarter.
Next cohort
Bring the app you already ship
Enrolment is informational on the pricing page. Write to the studio if you want a seat, a private workshop, or a review of an existing SDK.