The situation
Meta says it drove the signup. Google says it drove the signup. Your affiliate network says it drove the signup - and you only had one new user. Add up the three platform dashboards and you'd think growth tripled overnight, right up until finance reconciles it against actual accounts opened.
This isn't any one platform lying - each is reporting honestly on what it can see, which is only its own touchpoint in a journey that crossed several. The overcounting is structural, built into how self-reported attribution works everywhere, not a bug in one vendor's dashboard.
The pain
Every platform's self-reported attribution is built to make that platform look good, so stacking their numbers together always overcounts, and nobody can say by how much. Budget decisions built on that overcount are decisions built on nothing real - a channel that looks like it's driving strong growth might be taking credit that two other channels are also claiming for the same user.
The pattern compounds as spend grows across more platforms: the more channels in the mix, the more overlap, and the further the reported total drifts from the actual number of people who signed up.
What we implement
We route conversion events through one first-party system that assigns each signup to a single, rules-based attribution path before any platform sees it, instead of letting each platform claim the same event independently - this is server-side tracking with deduplicated, first-party attribution. Every platform still gets a conversion signal to optimize against; it just gets the true, deduplicated one instead of a shared claim on the same user. The rule set for assigning credit is documented and applied consistently, so the same signup is never counted twice and the same rule doesn't quietly change from one reporting period to the next.
What you get
- One true count of signups instead of the sum of every platform's inflated claim.
- A consistent attribution rule applied the same way every month, so trends are actually comparable.
- Platform bidding still gets a real signal, because they're fed the deduplicated outcome, not raw self-reported credit.
- A growth number you can put in an investor update without a footnote explaining why it doesn't match any single platform's dashboard.
Illustrative example
A fintech that had been adding up conversions reported by three platforms might find its true signup count is meaningfully lower than the sum, once duplicates are removed - and that the channel mix behind real signups looks different from the one behind the inflated total, changing where budget should actually go next. Illustrative - the actual overcounting rate depends entirely on your channel overlap and audience mix.