Why ad platform numbers never match on the same campaign.

Open Google Ads and it reports one conversion count. Open Meta and it reports another. Open the analytics tool and it reports a third, usually the lowest. The instinct is to work out which one is right, usually because a client has asked.

None of them is wrong. Each is answering a different question, and each is answering it correctly. Google Ads reports the conversions Google believes its clicks caused. Meta reports the conversions Meta believes its ads caused. The analytics tool reports what it could observe in the browser. Three questions, three answers.

This matters because you cannot fix it by picking a winner. You fix it by knowing which question you are asking before you look at a number.

In rough order of how much damage each one does. The first three move a number by a lot and are almost never checked.

• The attribution window

  Each platform counts a conversion if it happens within a set period after someone interacted with the ad. Those periods are configurable, and they are rarely set to the same length on two platforms. Check the setting on each client account before comparing anything.

• Clicks against views

  Some platforms will credit a conversion to someone who saw an ad and never clicked it. Others count clicks only. A platform counting views will always report more conversions, and the difference is not an error.

• The account time zone

  A platform reports days in the time zone set on the account, not the one you report in. If those differ, spend and conversions near midnight land on the wrong day, and a daily comparison is off by a slice of every day.

• Restatement

  Platforms revise figures for days that have already closed, as late conversions arrive and as invalid clicks are refunded. The number you read on Monday for Sunday is not the number that will be there on Friday.

• Currency and billed cost

  Spend and billed cost are different figures on most platforms, and a converted currency depends on the rate and the date used. Two totals can differ purely on which of these you pulled.

• The same person, counted twice

  Somebody who saw an ad on one platform and clicked one on another gets credited by both. Add the two platforms together and that person becomes two conversions. This is why a total across platforms is almost always higher than reality.

It does not mean one conversion number that every platform agrees with. That number does not exist, because the platforms are not measuring one thing.

It means the difference is explainable and stable. You know why one platform reports more than analytics, you know the window and the time zone behind every figure, and the gap does not move for reasons nobody can name. A gap you can explain is fine. A gap that changes every month is a broken pipeline.

For an agency, that explanation is the thing to give a client, once and in writing: which figure the report uses, why the platforms show more, and that the gap is expected.

Four rules, and they work whether you build this yourself or buy it.

• Reconcile spend, not conversions

  Spend does reconcile. It is a figure the platform charged the client, and it can be checked against an invoice. Make spend the number that has to agree everywhere, and treat conversions as platform-reported.

• Store the window and the time zone as columns

  Not in a document, and not in somebody's memory. If the window is a field on the row, anyone reading that number later can see what it means.

• Keep history, do not overwrite it

  Add each sync to the table rather than replacing yesterday. When a platform restates a figure you then hold both the original and the correction, which is the only way to tell a restatement from a bug.

• Never total conversions across platforms silently

  If you must show one figure, label which definition it uses. An unlabelled cross-platform conversion total is the most common wrong number in marketing reporting.

Every client's platforms are pulled into a warehouse you own and modelled to one name, one type, one grain and one currency per metric, so the figures that can be compared are comparable and the ones that cannot are labelled.

Syncs add to the tables rather than replacing them, so a restated figure leaves both numbers on record. Freshness and threshold checks run against those tables, which is what tells you a gap moved because a platform changed and not because a sync failed.