Name the unit
Before a single span is emitted, we force a written definition of “a use”: call, unique device, billable operation, or something your finance team will recognise.
Backup Cloudhub · Chatuchak studio
Measure what the app actually spends — not what the dashboard invents.
We teach API Usage Analytics for Apps to product, platform, and mobile teams who already ship, and who are tired of cardinality explosions dressed up as insight. One curriculum. Real traces. No vanity charts.
Read the flagship syllabus →
11 weeksFlagship studio length
38 cohortsRun since late 2018
1.7MRequests in the shared lab day
27Thai product teams hosted
From the studio floor
What recent participants actually said
The sampling week stopped us from shipping a client SDK that would have doubled our billable units overnight. I still disagree with their default histogram buckets, but the argument was worth the tuition.
★★★★☆Quota Design Intensive is dry in the best way. We left with alert copy our on-call would actually keep enabled.
Flagship
The Usage Ledger
An eleven-week studio on API Usage Analytics for Apps: defining a billable unit, instrumenting without drowning the binary, sampling rare mobile paths, and reading dashboards with commercial intent. Taught by Arthit Khemkhaeng with live traces from a Bangkok ride-hailing archive.
Evenings in Chatuchak, plus a recorded recap for anyone on a late shift. Informational tuition from ฿24,800 — see the pricing page.
Open the syllabusHow we teach
Restraint first, then instrumentation
Sample on purpose
Head sampling, tail sampling, and client-side thinning are treated as product decisions, not library defaults. Rare checkout paths stay visible.
Read for money
Dashboards are graded on whether a PM can explain cost movement to a CFO, not on how many panels they contain.
Catalogue
Other programmes this season
Mobile Sampling Clinic
Client telemetry that survives flaky networks and OS background limits.
Quota Design Intensive
Budgets, burst behaviour, and alerts that do not get muted in week two.
Regional Latency Workshop
SEA routing, Bangkok peak hours, and what they do to usage graphs.
Bangkok
A studio that notices peak hour
Chatuchak is not a backdrop. Evening traffic, festival weekends, and mobile-first habits in Thailand show up in every usage graph we study. If your app’s “normal” was modelled on a US lunch break, the ledger will look wrong by 18:00 here.
How the studio works →Next conversation
Bring a trace, not a slide deck.
Tell us which endpoint is lying to finance. We will say whether The Usage Ledger, a clinic, or a private desk is the honest next step.