Siloed platform reports vs. unified attribution: diagnosing the double-counting tax

Siloed platform reports vs. unified attribution: diagnosing the double-counting tax

Add up the conversions every platform claims. The total exceeds your actual sales — sometimes by 30-50%. The gap is the most common, least-discussed attribution failure, and the fix is a tooling choice.

Why the numbers don't reconcile
Each ad platform runs its own attribution in its own walled garden, with its own window, crediting any conversion it touched. When a user sees a Meta ad, clicks a Google ad, and opens an email, all three platforms independently claim that single conversion. Sum them and you've triple-counted. This isn't error — each is correctly reporting its own contribution; the error is summing self-reported, overlapping credits.

The two reconciliation approaches
Unified MTA ingests all touches into one identity graph and assigns each conversion's credit once, splitting it across channels — eliminating double-counting by construction, but inheriting all of MTA's tracking-coverage and correlation limits. MMM sidesteps the problem entirely by working at aggregate spend level, where overlapping claims never arise because it never tracks individuals.

The diagnostic ratio
Sum of platform-claimed conversions divided by actual total conversions. A ratio of 1.4 means your platforms are collectively over-claiming 40%. Track it; if it's rising, your channels are increasingly cannibalizing credit.

Bottom line for practitioners: Never sum platform-reported conversions for budgeting — it structurally over-credits and inflates every channel's apparent ROAS. Use a unified model (MTA or MMM) that credits each conversion exactly once. And compute the over-claim ratio quarterly: it's the cheapest single number that tells you how much your siloed reports are lying, and it requires no new tooling — just your platform exports and your real order count.