Flagship course
Feature Uptake Measurement
Twelve weeks, one live feature, no toy warehouse. You will write an uptake map, a cohort contract, and a retirement brief — then defend them in critique.
Learning outcomes
- Define exposure without pretending a home-tab impression equals intent.
- Name first value as an observable job-to-be-done, not a screen view.
- Set a repeat window that matches how often the job actually occurs.
- Write a quiet-death rule your leadership can disagree with in public.
- Split curves by surface and language when the product ships in Thai and English.
- Produce a retirement brief that names political cost, not only the metric.
Modules
-
01
Bring the dirt
Inventory the event list for one shipped feature. We do not clean it for you. Missing properties become part of the uptake map’s limitation note.
-
02
Exposure without theatre
Who could have seen it, on which surface, after which release. Home-tab inflation is the usual villain.
-
03
First value, in sentences
Write the job. Bind it to one event. If PM and analyst disagree, the week repeats until they do not.
-
04
Habit windows
Daily is often wrong. We match repeat use to the real cadence of the job — weekly payroll, monthly close, occasional transfer.
-
05
Cohort contract
Stop mixing launch week with everyone else. Draft the contract; critique another team’s.
-
06
Language and surface splits
Thai versus English copy, mobile versus desktop, new versus returning. Adoption that only works in QA English is not adoption.
-
07
Quiet death
Write the rule for calling a feature failed while taps still exist. Practice saying it out loud.
-
08
Retirement brief
Assemble evidence, options (kill, shrink, hide, relabel), and the political cost. Guest reader from alumni.
-
09
Leadership readout
A 12-minute brief. No appendix dump. Instructors play a sceptical CPO.
-
10–12
Revise and close
Second drafts only. 90-day follow-up prompt after the cohort ends.
Instructor
Nalinee Prasert
Lead instructor. Former head of product analytics at a Bangkok fintech, now teaching Feature Adoption Analytics full time at Cloudeventgrid. Nalinee still sits in on Adoption Atelier diagnostics when a warehouse is especially tangled. She does not ghost-write your SQL.
FAQ
Do I need SQL?
You should already be able to join an events table and filter by a property. This programme will not teach SQL from zero. If you cannot yet do that, wait, or send an analyst teammate with you. That limitation is real: we have turned people away in week one rather than pretend otherwise.
Can I switch features mid-cohort?
No. The point is to live with one feature long enough to see quiet death. If your feature is deprecated by legal, write to the desk and we will discuss a rare exception.
Is the Bangkok room required?
No. Remote seats follow the same artefacts. In-person seats meet at Street: 222/6 Gp 1 Onnuch-Ladkrabang Lardkrabang Lat Krabang; City: Bangkok; State/province/area: Bangkok; Zip code: 10520; Country: Thailand.
Will you instrument our app?
No. Teaching only. If tracking is empty, we will say so immediately.
Reviews from this syllabus
“Module three forced us to split opened-the-screen from finished-the-job. That argument alone paid for the seat.”
Krittin Cho, Senior PM, payments
The cohort moved faster than our analytics stack could keep up; I spent week four waiting on event naming cleanup before the dashboards made sense. Once that landed, the uptake maps were the first artefact our CPO used in roadmap review.
verified seat · Signal Bench · March 2026 · ★★★★☆
Nalinee refused our first ‘adoption %’ slide in the leadership readout. Embarrassing. Correct.
client in regional e-commerce