So, Pinterest's official line is that their "enhanced" iOS 15+ conversion tracking is a robust workaround. My own data, and I suspect the data of anyone who actually checks, suggests otherwise. Our reported conversions from Pinterest traffic have dropped by roughly 60% since we implemented their recommended CAPI (Conversion API) setup alongside the pixel.
We followed their documentation to the letter: the Pinterest tag is firing, our server-side events are being sent via their CAPI, and the two are supposedly being deduplicated. Yet the attribution in the Pinterest Ads Manager looks like a ghost town. Off-platform analytics, where we can still see a somewhat sane world, show Pinterest traffic converting at a much higher rate than Pinterest itself claims.
This isn't just a discrepancy; it's a systemic failure. I'm curious if anyone else has torn apart the actual plumbing here and found the leak.
* Are you seeing a similar under-reporting gap?
* More importantly, has anyone done a methodical, hold-everything-else-constant test to see what, if anything, in their new setup is actually capturing conversions?
* I'm particularly suspicious of their deduplication logic between the browser pixel (now largely crippled by ITP) and the server-side events. The timestamps and IDs would have to be perfectly synchronized, and I have zero faith in that being the case.
Before I get another generic support response about "allowing 72 hours for data to populate," I'd like to hear from people who've done the forensic work. What's actually broken? Is it the client-side tag's inability to set a proper `event_id`? Is the CAPI integration silently dropping events? Or is this just the new normal of massively under-reported performance we're supposed to blindly optimize against?
Data skeptic, not a data cynic.