Where Thai booking flows actually lose people
When a Bangkok-built travel app says “checkout conversion is 41%,” that number is usually an average of three different products sharing one screen. Travel App Analytics starts by un-averaging them.
We review event maps, not marketing decks. Across recent Destination Funnel Studio homework, three drop-off windows appear often enough that we now teach them before any dashboard cosmetics.
Window one: the THB wallet pause
Thai baht checkout frequently offers a local wallet alongside cards. Travelers tap the wallet, leave the app, and either complete payment in another surface or abandon there. If your analytics only fire checkout_complete inside the travel app, wallet successes look like failures and wallet failures look identical to card failures.
The fix is not a new vanity event. It is a return handshake: wallet intent, wallet return, wallet result. Until those three exist, do not compare THB conversion to foreign-card conversion. You are not measuring the same journey.
Window two: foreign-card 3-D Secure
Foreign cards on Thai acquiring often hit an extra authentication step. That step can last long enough for mobile OSs to suspend your app. When the traveler returns, your session may have expired, your cart token may be stale, and your funnel may record a new session starting at home.
Teams that ignore this window “optimise” the home screen when the leak is token lifetime. In class we ask for the median time between pay_attempt and pay_result split by card region. If that median jumps above two minutes for foreign cards, treat session restoration as a conversion problem, not a growth experiment.
Window three: address and tax fields that only appear for some passports
A guest checkout that suddenly asks for a billing address in a Latin-only field will lose travelers whose cards expect a different script or a shorter form. This is not a localisation essay. It is a measurement essay: if the extra fields are conditional, your funnel steps must be conditional too. A single “checkout step 3” rate will hide the leak inside the smaller passport group.
If you cannot say which window moved, you do not yet have Travel App Analytics. You have a blended checkout chart.
What we do not claim
These windows are patterns from apps we have taught, not a census of Thailand. Your leak may be partner inventory latency instead — that is a different module. We also do not claim that adding events raises conversion by itself. Events only make the argument possible.
If you want the live critique, the flagship is Destination Funnel Studio. If you only need the recorded version, Field Notes covers the same windows without homework review.