Store and OS clocks
SKAN postbacks, Play install referrers, and first-party web timestamps do not share a timezone or a privacy window. Alignment starts by writing those clocks down, including Thailand’s ICT offset for campaign reporting.
Custom map
A working geography of cross-platform analytics alignment: the surfaces that claim to describe one product, the places they quietly invent their own users, and the desk rituals that keep Thai-market teams honest.
The Atlas is not a vendor comparison. It is a map of disagreement. Cohorts use it in week one before they touch a single dashboard theme.
SKAN postbacks, Play install referrers, and first-party web timestamps do not share a timezone or a privacy window. Alignment starts by writing those clocks down, including Thailand’s ICT offset for campaign reporting.
Device IDs, hashed emails, in-app account IDs, and guest checkouts. Most ‘fraud’ spikes in hybrid retail apps are two islands claiming the same order.
Firebase, Adjust, a warehouse dbt model, and a spreadsheet named ‘final_v7’. If the name purchase means three different moments, the board is reading fiction.
Songkran, 11.11, Chinese New Year travel, and payday Fridays change volume and mix. An alignment desk that ignores the Thai calendar will call seasonality a tracking outage.
During Cross-Platform Analytics Alignment, each team pins their live stack onto these four territories. Pins that land in two territories at once — a mini-app checkout that is both an identity island and an event dialect — become the week-three critique.
Field Notes readers get a static version. Atelier teams annotate a private copy with their actual table names. Continuum Seat alumni return when a new surface appears, usually a second iOS listing or a LINE mini-app.
It will not pick an MMP for you. It will not promise blended ROAS. It will not certify that two platforms can be forced into a 100% match under current privacy rules.
If you want a seat that walks this map with critique, start with the flagship modules or write to the desk with the two reports that already disagree.