Instrumented Session Quality
Design capture, define a score you can defend, and present findings that survive a sceptical engineering review.
Bangkok studio · since 2015
Session quality is not a dashboard. It is a discipline.
Cloudpacketbase trains product, research, and platform teams to treat Session Quality Analytics as evidence work: instrumentation that can be audited, scores that can be argued, and replay that does not pretend to be the whole truth.
Open the curriculum- Practitioners through the studio
- 1,640
- Live cohort cycles completed
- 29
- Partner product squads in SEA
- 63
From the room
What alumni actually remember
The Instrumented Session Quality cohort forced us to throw out a composite “health” number that mixed latency, rage clicks, and NPS. We now publish three separate traces and a written caveat. Leadership complained for two weeks, then started asking better questions in review.
Latency Budgets for Interactive Products was dry in the best way. I still keep the week-two worksheet taped beside my monitor.
★★★★☆Useful, though the homework in week three landed on our freeze. I skipped two labs and felt the gap later.
Curriculum preview
Courses currently in rotation
Latency Budgets for Interactive Products
Translate percieved slowness into budgets, percentiles, and session-level failure modes rather than a single average.
Cohort Replay Analysis
Sample replays without drowning. Learn when a clip is evidence and when it is just a vivid anecdote.
Why teams enrol
Quality work that can be audited
Most “session health” stacks collapse three different questions into one chart: did the system work, did the person finish, and did they feel respected. Cloudpacketbase keeps those questions separate, then shows how Session Quality Analytics can still produce a score product leadership will fund.
You will practise writing instrumentation notes, scoring rubrics, and limitation statements — the unglamorous artefacts that keep a quality programme honest after the first quarter.
What we refuse to teach
We do not train people to inflate completion rates, hide sampling bias, or treat rage-click heatmaps as a strategy. If your organisation needs a vanity number for a board deck, this is the wrong studio.
We also do not claim that replay vendors are interchangeable. Capture quality, privacy posture, and export fidelity differ, and we name those differences in class.
From the journal
p95 duration is not a quality proxy
Long sessions can mean delight, confusion, or a background tab. Our latest note walks through three Bangkok checkout studies where duration moved in opposite directions to task success.
Read the essayNext step
Sit with the method before you buy a seat
The Method page explains how we construct a session quality score, what we sample, and where the Bangkok Field Lab fits. Fees are listed plainly. Nothing on this site is a checkout.
Study the method