For most sales campaigns, Meta's default attribution setting of 7-day click plus 1-day view is the right choice, because it feeds the delivery algorithm the most conversion signal and that usually improves optimization. The catch is reporting: the 1-day view portion credits people who merely saw your ad, so the ROAS you read includes conversions you might have earned anyway. Optimize on the default, but judge performance with click-only comparisons and incrementality checks.
What the attribution setting actually controls
The attribution setting does two jobs at once, and confusing them is where most bad decisions start. First, it defines which conversions Meta's delivery system learns from: a wider setting means more conversion events flow back as training signal, and the algorithm finds buyers faster. Second, it defines which conversions appear in your reports for that ad set.
People argue about this setting as if it were only a reporting preference. It is not. Narrowing it changes what the machine optimizes toward, which changes who your ads are shown to. That is why you cannot treat attribution as a cosmetic reporting toggle you flip after launch; it is a delivery lever with reporting side effects.
The four options compared
| Setting | What it counts | Best for |
|---|---|---|
| 7-day click + 1-day view | Purchases within 7 days of a click or 1 day of a view | Default for most sales campaigns; maximum signal |
| 7-day click only | Purchases within 7 days of a click | Cleaner reporting; comparing against GA4 or backend |
| 1-day click | Purchases within 1 day of a click | Impulse products, flash sales, very short cycles |
| 1-day click + 1-day view | Adds same-day view conversions to 1-day click | High-frequency retargeting where views plausibly matter |
Meta also exposes a 28-day click window for reporting comparisons in some views, but the ad set optimization choices are the four above. The wider the window, the more conversions Meta can attribute to itself, and the more flattering the ROAS looks on screen.
When to move away from the default
The default is a strong starting point, not a law. Three situations justify a narrower window:
Short consideration cycles
If nearly all your purchases happen the same day as the click, 1-day click concentrates the signal on your real buying moment and can tighten delivery around people who act fast. Impulse categories and flash sales are the clearest cases.
Heavy retargeting overlap
Retargeting audiences see many impressions, so 1-day view attribution vacuums up conversions from people already on their way to purchase. Click-only settings keep retargeting honest and stop it from claiming demand it did not create.
Strict measurement periods
When you are running a lift test or reconciling against backend data, click-only windows make the comparison cleaner because they remove the view-through credit that no other system can see.
Longer cycles do not need a longer optimization window as much as people assume; the 7-day click component already captures most considered ecommerce purchases. B2B with multi-week cycles is a different measurement problem that no Meta window solves, and stretching the window will not fix it.
How the setting distorts reported ROAS
Widening attribution always raises reported ROAS without changing a single sale in your backend. Moving from 1-day click to 7-day click plus 1-day view can make the same campaign look dramatically better, which is why comparing ROAS across accounts or time periods with different settings is meaningless. Before you celebrate a jump in performance, check whether someone changed the window.
View-through credit is the slipperiest part. A 1-day view conversion means someone was served an impression and bought within 24 hours. Some of those were genuinely influenced; many were existing customers or people arriving through email, search, or plain habit. Meta cannot distinguish influence from coincidence at the user level, and neither can you from the report alone. Only holdout-based lift testing separates the two, which is why blended metrics like MER and backend revenue belong next to every platform ROAS number you look at.
A practical setup that stays honest
Run prospecting on 7-day click plus 1-day view for maximum signal, so the algorithm has the fuel it needs to find new buyers. Run retargeting on click-based settings so it cannot inflate itself on warm audiences. Build a reporting view that shows click-only numbers next to the default, and track backend revenue as the denominator that platform settings cannot touch. When a big decision rests on view-through value, test it with a holdout instead of debating it in a meeting.
Where a unified view helps
Attribution settings decide what Meta shows you; they do not decide what actually happened. Admira layers multi-touch attribution, marketing mix modeling, and lift testing over your platform and backend data, so you can see how much of that view-through ROAS is incremental and how much is Meta grading its own work. If you are tired of defending a Meta ROAS your backend cannot find, book a demo and see the same campaigns scored against real revenue.
FAQ
Does changing the attribution setting reset learning?
Changing it on an existing ad set counts as a significant edit and can send the ad set back into the learning phase. Decide the setting before launch when possible, and avoid flipping it back and forth on live campaigns, because each change costs you delivery stability and clean data while the algorithm recalibrates around the new signal.
Why did my conversions drop when I switched to 1-day click?
Reported conversions drop because the counting window narrowed, not necessarily because performance fell. Delivery may also shift as the algorithm optimizes toward faster converters. Judge the change against backend sales and blended ROAS, not the in-platform number, or you will mistake a reporting change for a real decline and overreact.
Which setting matches what GA4 shows?
None exactly, but 7-day click only is the closest comparison, since GA4 never sees view-through conversions from Meta impressions. Expect a gap anyway, because of tracking loss, consent modeling, and the fact that GA4 counts on the conversion date while Meta counts on the interaction date. Compare weekly totals rather than daily.
Is 1-day view worthless then?
No. As optimization signal it often helps delivery and speeds up learning. As proof of value it is weak evidence, because it bundles real influence with conversions that would have happened regardless. Treat it as signal for the algorithm, not as a verdict on incremental performance, and let lift tests settle the value question.
.svg)
%20(2).png)
