Every platform claims the same sale
Platapode · 
Someone sees an ad on social media, searches for the brand a few days later, clicks a search ad, gets a retargeting ad that evening, and finally buys after opening an email.
One sale.
The social platform counts it, because the person saw its ad within its attribution window. The search platform counts it, because they clicked its ad. The retargeting platform counts it. The email tool counts it. Each one, in its own dashboard, reports a conversion.
Add them together and one sale has become four.
Nobody's lying
None of the platforms are doing anything dishonest. Each one is counting by its own rules, using its own tracking, looking at the part of the journey it can see. And each set of rules is reasonable on its own terms. The person did see the ad. They did click. The email was opened.
The problem is that each platform can only see its own part, and each one is built to report its own contribution as fully as it can. They don't check with each other. They can't really. So they each claim the sale, in good faith, and nobody's dashboard shows the overlap.
Across a whole business, this adds up. If you total what every channel reports it drove, you can easily get more conversions than the business actually recorded. Anyone who's done that sum has seen it happen. It's not an edge case. It's what you get when every system is marking its own work.
The number everyone uses anyway
And everybody in the industry knows about it. It's not a secret. Ask anyone who manages spend across several channels and they'll tell you the platform numbers don't add up to reality, and that each channel looks better in its own reporting than it would anywhere else.
And still, those numbers are what budgets get argued over. Each channel's owner brings their platform's figures to the meeting. Each set of figures looks good. The decision gets made somewhere in between, based partly on the numbers and partly on who's most convincing.
Some businesses try to fix this with a single model on top: data-driven attribution, a separate analytics tool, media mix modelling. Each of those has its own assumptions, and its own people who'll explain, if pushed, where it's likely to be wrong. The overlap gets smaller. It doesn't go away.
So the business ends up choosing a number, not because it's right, but because everyone can work with it. A shared figure that's roughly defensible and gets the budget decided.
Attribution, most of the time, is a way of agreeing on a version of events.
