Reviews
These are unsolicited write-ups and interviews after cohorts. We do not pay for testimonials. One of them includes a reservation on purpose — the JS-bridge labs are not equally useful if your vendor wraps native APIs.
★★★★☆
Taxonomy week on Event Taxonomy for Dual Runtimes stopped our team inventing a third “legacy_v2” prefix. I still think the sandbox’s Flutter module is thinner than a real production shell, so treat that lab as a sketch.
Arthit Phong — measurement lead, ride-hailing
“Consent Architecture for PDPA” was the first time legal and analytics sat in the same document. We now fire the AppsFlyer init after the Thai consent screen, not before the splash.
Ploy, Chiang Mai
I came in expecting Attribution Windows After ATT to hand me a magic lookback. Instead Kanya made us write the window in the measurement contract and defend it to a fake CFO. Our install-to-first-purchase number dropped. Finance prefers the smaller number because it no longer includes WebView ghosts.
Client in consumer banking · Singapore desk, remote seat
★★★★★
Debugging the JS Bridge saved our checkout QA cycle. We found the WebView firing purchase on page load whenever the native shell injected a stale session token.
Verified learner · Hybrid App Measurement Strategy, 2025-B
Quiet forum. People post diffs, not pep talks. I used the aftercare to ask about GAID when our Android WebView started stripping the advertising ID on a WebView update.
Niran, product manager, Bangkok
Case study · 14 weeks
A regional wallet team enrolled two analysts and one Android engineer in Hybrid App Measurement Strategy, then extended into a short Measurement Residence. Before the studio, purchase events existed in Firebase (native) and in a custom pixel (WebView). Weekly board packs showed two conversion rates, 1.7 points apart.
The contract they left with assigned purchase ownership to the native SDK after the WebView posted a signed receipt. The pixel became a diagnostic stream only. Duplicate purchases fell out of the finance extract within three release cycles. They still cannot attribute mini-app traffic that never touches the shell — that limitation is written into the contract rather than hidden in a footnote.
Case study · taxonomy only
Ploy’s team (quoted above) took Consent Architecture for PDPA plus Event Taxonomy for Dual Runtimes. Their native app was a thin wrapper. Almost every meaningful action happened in a WebView. The mistake was treating screen_view from the wrapper as equivalent to page_view from the HTML.
After the desk they versioned events with a runtime prefix, delayed SDK init until PDPA consent, and stopped reporting a blended “activation” metric that mixed splash impressions with policy-form starts. Activation looks smaller. Retention conversations got shorter.