Studio
A teaching room for people who ship mobile products, not a content mill.
Identity Flowcore started because too many “predictive” decks in Bangkok were really just last week’s funnel with a trendline. We built a studio that treats identity, time, and decision rights as the curriculum.
Mission
We exist to make predictive mobile analytics boring in the right way: documented identities, defensible clocks, and scores that a growth lead can explain to finance without blushing.
The studio is deliberately small. Faculty rotate between teaching and client work so the labs stay close to the messy event streams coming out of Thai and regional super-apps, not to a textbook example from 2016.
Why Thailand
Pathum Wan is not a backdrop. Festival calendars, LINE as a second operating system, PromptPay retry loops, and multi-app wallets change how retention actually looks. Teaching from Bangkok keeps those textures in the room instead of flattening every market into a US SaaS template.
Pedagogy
Every live lab follows the same spine: a short reading, a SQL or spreadsheet drill, a critique, then a one-page operating note. We do not lecture for ninety minutes and call it a workshop.
Students work on a provided anonymised event extract first. Only after they can name leakage in that extract do they bring a slice of their own product data — and only if their employer allows it. We would rather a clean public drill than a confidential dump that nobody can discuss.
Feedback is written. Verbal praise fades; a marked notebook with “this join looks into next week” does not.
Community
Alumni stay in a quiet mailing list. It is not a Slack full of job ads. Once a quarter we host an evening critique in the Pathum Wan rooms: three people present a scoring change, the rest ask unkind, useful questions.
If you want a networking mixer, this is the wrong studio. If you want someone to notice that your “predicted LTV” still conditions on a post-purchase flag, you will be among friends.
Student journey
Most people arrive after a painful quarterly review: a churn model that looked brilliant in a notebook and did nothing in CRM. They spend the first week unlearning dashboard folklore.
By week three they can describe identity states on their app without waving at “the user.” By the final critique they can defend a threshold, name what the model cannot see, and write the email that tells lifecycle marketing when not to trust the score.
Some continue into the Atelier retainership. Others disappear back into their companies, which we count as success. The point was never to keep you in our orbit.