Server-to-server postback vs network pixel: the attribution gap that ate 18% of conversions
The setup: Lead-gen on mobile, lots of in-app browser traffic. Choice for attribution: drop the network's client-side pixel, or wire up server-to-server postbacks.
The move: We ran the pixel for a month, then switched to S2S postbacks and compared recorded conversions against the advertiser's own dashboard.
The numbers (illustrative): Client pixel: under-counted by ~18% — in-app browsers, ad blockers, and cookie loss silently dropped conversions. We were optimizing on a lie and pausing 'losing' campaigns that were actually profitable. S2S postback: matched the advertiser's numbers within 2%, and three 'dead' campaigns turned out to be +20% once tracked correctly.
The lesson: Client-side pixels leak — badly on mobile and in-app traffic. If your attribution under-counts, you don't just lose data, you make WRONG decisions: killing winners that looked like losers. S2S postbacks cost an hour of setup and tell the truth.
What I'd do differently: I ran client pixel for a full month and paused profitable campaigns based on phantom losses — the real cost wasn't the missing data, it was the good campaigns I murdered. Wire S2S before you spend serious money on mobile. The setup friction is nothing next to optimizing on bad numbers.
Arb Files
@ArbFiles