What the setting counts
An attribution setting is a rule deciding which conversions Meta will claim. This one claims a conversion if the person clicked the ad and converted within seven days, or if they were shown the ad without clicking and converted within one day of seeing it.
Two things follow from that. First, the window is measured backwards from the conversion, and the conversion is reported against the day of the ad interaction rather than the day of the sale, so yesterday’s figures keep rising for several days afterwards. Second, the click window is far more generous than the view window, which is Meta’s own way of saying a click is stronger evidence than an impression.
The setting sits at ad set level, and it does more than shape reporting. It also tells the delivery system which conversions to chase, so two otherwise identical ad sets on different settings will be shown to different people.
Why the attribution setting matters
Every conversion number you look at is a claim made under a rule, and this rule decides how large the claim can be. A shorter window reports fewer conversions and a lower return on ad spend; a longer one reports more of both. Nothing about the business has changed — only the accounting.
It matters most when you compare things. Meta on this setting and GA4 on last-click will disagree about the same week, and neither of them is lying. Adding the two together double-counts. Comparing this month on one setting against last month on another produces a trend that does not exist.
For a considered purchase — an education consultancy, a property, a remittance service someone signs up for after asking family — a seven-day click window can still be too short to see the whole effect, so the ads look weaker than they are.
Where it goes wrong
The most damaging mistake is changing the setting and then comparing periods across the change. The numbers move, someone concludes the campaign improved or collapsed, and a budget decision follows from what was only an accounting change.
The second is trusting view credit uncritically. A view-through conversion counted because someone was shown an ad the day before may well have happened anyway. It is a signal, not proof, and it deserves less weight in your decisions than a click.
The third is reading a report too early. Because conversions are attributed back to the day of the interaction, the most recent days are always incomplete. Judging yesterday’s cost per result as final makes every campaign look worse than it is.
How to act on it
Pick one setting, write it down, and hold it across the account so periods stay comparable. Put the setting on the front of any report you send, so nobody has to guess what the number means.
Use the comparing windows view in Ads Manager to see the same period under several settings at once. If the click-only figure sits close to the reported one, view credit is not doing much work and you can rely on the number. If the two are far apart, most of the claim rests on impressions, and you should check the campaign against your own sales records before scaling it.
Above all, keep an independent count. Orders in your shop admin, leads in your CRM, enquiries in your inbox — none of these care about attribution rules. Platform figures are for steering the campaign; your own records tell you whether the business grew. Making those two line up, and knowing why they differ, is the ordinary work of analytics and tracking.