Why your ad platforms don't agree with your CRM

Every ad platform marks its own homework. Meta claims a conversion. Google claims the same conversion, on the same day, from the same person. LinkedIn, if it was anywhere near the customer journey, claims a piece of it too. None of them are lying exactly — each is reporting what it saw, inside its own attribution window, using its own definition of a touchpoint. The problem is that "what each platform saw" and "what actually happened" are two different things, and only one of them pays the bills.

Your CRM closed 40 deals but platforms claim 120. Which is right?

Usually the CRM. It records what a sales team actually closed, while each ad platform only records what it claims credit for. If Meta, Google and LinkedIn each claim the same deal, 40 closed deals can show up as 120 platform conversions. The gap is double-counting, not 80 phantom customers, though the CRM still needs checking for duplicates and missing source data.

The double-counting problem

Run a real reconciliation and you'll usually find your ad platforms are overselling their own results (see Meta and Google both claiming the same conversion). Add up every platform's claimed conversions and the total is higher than the business the CRM says actually closed — sometimes by a wide margin. That gap isn't fraud. It's what happens when five sources are each rewarding themselves for a decision that was, at most, made once.

This matters more than it looks like it should, because budget follows whichever platform reports the best numbers. If two platforms are both claiming the same conversion, the one with the more generous attribution window looks like it's winning — and gets more spend — regardless of what it actually contributed.

What blended attribution actually does

Cross-channel attribution isn't a dashboard feature. It's a marketing data warehouse decision: every platform's data, your tracking, and your CRM, landed in one place and reconciled against a single model, so a lead or a sale gets credited once instead of five times. Built properly, it answers a question no single platform can: which channels the closed revenue can be traced back to, not which channel is best at claiming it.

The CRM is the reference point in this, because it records outcomes rather than claiming credit for them. GA4 is no exception: GA4 and the CRM count different things. It needs checking too: duplicate contacts, missing source data and unmatched records all affect it. But it's the closest record of what a sales team actually closed, and blended attribution means running every platform's numbers against it before deciding anything. How the credit gets shared is the job of the attribution model; to size your own overlap first, try the attribution gap calculator.

One caution. Matching a closed sale to a campaign assigns credit under a set of rules; it doesn't prove the campaign caused the sale. The result is only as good as three things: identity coverage (how many CRM records carry a source), the model rules (how credit is shared when several channels touched the deal), and how many records stay unmatched. A reconciliation should report all three.

A worked example

Take an illustrative account running four channels — Google, Meta, LinkedIn and Bing — on a combined $320K of spend over eight weeks. The ad platforms report 451 leads between them. Reconcile that against the CRM and only 388 are unique; 63 were counted more than once across platforms.

Push further, past leads to qualified pipeline and closed revenue, and the picture inverts. Meta brings in the most leads by a wide margin, but ranks third on revenue. LinkedIn brings in the fewest leads and the highest cost per lead, but ranks second on revenue. Judged on ad-platform numbers alone, you'd scale the channel that's actually converting worst and start cutting the one quietly closing the biggest deals. No ad account can show you that comparison — only the CRM, blended against every channel, can.

What to do next

You can't fix what you haven't measured. The starting point is finding out how big your own overlap is — how many of your reported leads are actually unique, and whether your best-performing channel on the platform dashboards is your best-performing channel in the bank. That's a reconciliation exercise before it's a build, which is why it's worth doing as a standalone step rather than assuming.

For the full guide to tying ad spend to closed revenue, read B2B marketing attribution: a practical guide. If you want the full mechanics of how the warehouse and the model get built, that's covered on the Data & BI page. If you just want to know how big your own gap is first, that's what a free marketing data audit is for.

Questions people ask