I came for Sequence Atlas because our “activation” metric was just first login. Module 05 forced us to look at days 8–14. We still have not fixed invite copy — that is on us — but the quiet-week split is now on the Monday board.
Flagship · 8 live sessions
Sequence Atlas: Usage Paths for SaaS Teams
A working method for reading how tenants actually move through a product. You will not leave with a vendor dashboard. You will leave with a labelled path library, a written event contract, and a clear view of where pattern work stops helping.
Modules
- 01 · Event contracts that survive a rename Name, trigger, required properties, owner. No SQL until the page is boring enough to defend in a hallway.
- 02 · Grain: tenant, seat, and session Why B2B SaaS invents “power users” when you only count clicks. Shared-login cases from Thai workspaces.
- 03 · Sequence without the junk drawer How far n-grams are useful, when to collapse, and when a path is just noise from a tooltip experiment.
- 04 · Feature gravity after week two Return rates by family, empty-state traps, and the difference between discovery and habit.
- 05 · The quiet week Days 8–14 after invite: silence that still pays, silence that precedes churn, and why DAU cannot see either.
- 06 · Expansion cues in the clickstream Admin invites, seat adds, and export spikes — which ones precede an invoice change in SEA billing cycles.
- 07 · Lab: anonymised Khlong San archive You label twelve journeys from a synthetic invoicing product. Faculty score the atlas, not the chart polish.
- 08 · Briefing the room A ten-minute path story a PM can argue with. If it needs a legend, it is not ready.
What you should be able to do
- Write an event contract another team can implement without a call.
- Separate tenant rhythm from individual clicking in a weekly review.
- Produce a path library of no more than twenty named journeys.
- Explain one real limitation of pattern work to finance without apologising.
Informational fee
THB 39,800 / seat
Includes recordings for three weeks, the paper atlas template, and one office hour in the following month. VAT and bank fees are quoted on the invoice. This page does not process payment.
Ask for the next intakeInstructor
Nalinee Charoensuk
Nalinee ran product analytics for two Bangkok SaaS companies before opening the Khlong San desk in 2018. She still refuses to teach a “retention score” that cannot be unpacked into named paths. She teaches in English; Thai office hours are available on request.
From recent seats
★★★★☆
Paper maps before SQL felt slow in week one. By week five I was glad. Would have liked more warehouse-specific SQL; they stay tool-agnostic on purpose.
Questions we actually get
Do I need a data warehouse already?
You need exports you are allowed to share in class, even CSV. We do not require BigQuery, Snowflake, or a particular CDP. If you have nothing but Mixpanel screenshots, Feature Gravity Lab is a better first desk.
Is this useful for consumer apps?
The method travels, but the archive and examples are B2B SaaS. Consumer teams often find the tenant-grain module less relevant. We say so on day one rather than stretching the material.
What will this course not do?
It will not rescue a product with no reason to exist, and it will not produce a model that “predicts expansion” from three weeks of noisy events. Pattern analysis describes how people already move. If nobody returns after invite, the atlas will show an empty room — not a growth lever. That limitation is part of the teaching, not a footnote.
Can my whole team sit together?
Two seats on the public intake is fine. Larger groups should look at House Pattern so the taxonomy work happens on your floor. See fees.