Open three ad accounts on the first of the month and you will find the same order counted three times. Meta reports it. Google reports it. TikTok reports it. Add the dashboards together and you have three thousand dollars of revenue from a single thousand-dollar customer.
Nobody is lying. Each platform is following the rule it was built to follow, and each rule is defensible on its own. The trouble starts when you try to use those three numbers to decide where next month’s budget goes.
Why every platform is right and the total is wrong
Each ad platform measures conversions inside its own attribution window, using its own definition of a contributing touch. Meta counts a sale if it served an impression within its window — an impression, not a click. Google counts the same sale because the buyer searched your brand name before checkout. TikTok counts it because the video that started the whole thing played eleven seconds two weeks earlier.
All three claims are true. None of them is the whole story. And critically, no platform can see the other two, so none of them can tell you which touch was doing the heavy lifting and which was taking credit for a decision that had already been made.
Reported revenue says $3,000. Your bank says $1,000. Until something arbitrates between the platforms, you are allocating next month on a number that does not exist.
The three habits this produces, all of them expensive
1. Over-funding the best claimer
Retargeting and brand search are structurally excellent at claiming credit, because they touch people who were already going to buy. Budget drifts toward them because the dashboard rewards them, and drifts away from the channels that actually created the demand in the first place. Six months later the account is efficient and the business has stopped growing.
2. Killing the wrong campaign
Upper-funnel work rarely gets last-click credit. Judged inside a single ad account it looks like waste, so it gets switched off — and eight weeks later the “efficient” channels quietly get more expensive, because nobody is feeding them any new demand.
3. Arguing instead of deciding
When three sources disagree, meetings become debates about whose number is right. The debate is unwinnable, because all three are right by their own definitions. What is missing is not more reporting. It is a decision about which number the business is going to be run on.
What to do instead, in the order that works
Pick one number and make it the referee
Before touching any campaign, decide what a result is. For a store it is usually gross profit after cost of goods; for a service business it is closed revenue, or qualified opportunities if the sales cycle is long. Write it down. Everything else — clicks, ROAS, video views, reach — becomes a diagnostic, useful for finding where a campaign is broken, never for deciding whether it deserves funding.
Fix the plumbing before you argue about the numbers
Most attribution arguments are actually tracking arguments. Server-side tracking, a conversions API on each platform, consistent event naming, and one analytics property that sees everything. If the events firing into each platform are not the same events, no amount of modelling will reconcile them. This is unglamorous work and it is always the highest-leverage week in an engagement.
Judge channels on blended economics, not platform ROAS
Total spend against total revenue, at the business level, month over month. It is a blunter instrument than a platform dashboard and it has one enormous advantage: it cannot double-count. If blended cost per acquisition is falling while a platform swears its ROAS is falling too, the platform is losing visibility, not losing performance.
Test by holding something out, not by reading a report
The only honest way to know what a channel contributes is to turn it off in a region or an audience and watch what happens to the blended number. It is uncomfortable and it is the closest thing to truth available since the privacy changes. Run it deliberately, for long enough to mean something, and write down the expected result before you start.
This is the entire job of media buying
One strategy governing every platform, one measurement standard agreed before a dollar is spent, and budget that moves to whatever is genuinely producing revenue rather than to whatever is best at claiming credit for it.
The uncomfortable part
Doing this properly usually means your reported numbers get worse before your business gets better. Three inflated dashboards become one honest one, and the honest one is smaller. That first month is where most companies lose their nerve and go back to the comfortable version.
The businesses that push through it get something the comfortable version never provides: the ability to answer, with evidence, which dollar earned the most — and to move the next one there before a competitor does.
If the tracking underneath all of this is the part you are least sure about, that is usually where we start too. Our CRO & Analytics work exists precisely because measurement that survives a cookieless browser is now a prerequisite for everything else, and for online stores in particular the e-commerce version of this problem has its own specific shape.