7-Day Click, 1-Day View: What Changing Your Attribution Window Actually Does

The attribution setting is one of the few controls in Meta Ads that changes your reported results without changing anything about your business. Switch from 7-day click to 1-day click and your ROAS will fall, immediately, while exactly the same number of people buy exactly the same things. Understanding what the setting does is the difference between using it deliberately and being confused by it for a quarter.

What the setting actually controls

The attribution setting answers one question: how long after someone interacts with your ad will Meta still take credit for a resulting conversion, and does merely seeing the ad count.

The common options combine two things.

• Click window. Usually 1 day or 7 days. Someone clicks the ad, then converts within that period, and the conversion is credited to the ad.
• View window. Usually off or 1 day. Someone is served the ad, does not click it, and converts within that period anyway. Meta claims some credit.

Two things follow from this that are worth stating plainly.

A wider window always reports more conversions. It cannot report fewer. Every conversion captured by 1-day click is also captured by 7-day click, plus the ones that took longer. So 7-day click with 1-day view is the most generous common setting and will always show the best ROAS.

None of this changes your bank balance. The orders were the same. You have changed the accounting rule, not the result.

The number that matters most: your real purchase delay

The right window is not a matter of taste. It is a property of how your customers actually buy, and you can measure it.

If you sell a £20 impulse product, most people who are going to buy do so within hours. A 7-day window is mostly capturing people who would have bought anyway, and the extra credit is partly illusory. If you sell a £900 considered purchase, a 1-day window is throwing away most of your genuine influence, because nobody buys that on the first visit.

The practical way to find out: look at your own order data and work out the time between first touch and purchase for a representative month. Most ecommerce backends or analytics tools can show you time-to-purchase or days-to-conversion. If 90% of your orders happen within 24 hours, a long click window is adding noise. If half your orders take more than three days, a 1-day window is systematically understating what your ads did.

This is the same underlying discipline as everything else worth doing in measurement: anchor on what your own systems record rather than arguing about what the platform claims. The general case is in ad platform revenue versus actual revenue.

View-through attribution deserves its own scepticism

Click attribution has a defensible logic: the person did something deliberate. View-through is softer, and it is where most overstated-results arguments live.

A view-through conversion means someone was served your ad, did not click it, and later bought. Sometimes that is genuine influence. Sometimes the person was going to buy regardless and happened to be shown an ad in the preceding day, which is especially likely for retargeting audiences made up of people who already visited your product page.

Retargeting with 1-day view enabled is the classic way to produce a spectacular ROAS that does not correspond to incremental revenue. You are largely taking credit for buyers you already had.

That does not mean view-through is worthless, and turning it off everywhere is a blunt response. It means view-through numbers should be read as a softer claim than click numbers, and that comparing a retargeting campaign on view-through against a prospecting campaign on click is comparing two different standards of evidence.

The part almost nobody accounts for: it changes optimisation too

Here is the consequence that gets missed, because the setting looks like a reporting preference.

The attribution setting also defines what the delivery system treats as success. On a 1-day click setting, the algorithm learns from people who converted within a day, and goes looking for more people who behave like them. On 7-day click, it learns from a broader and slower-converting population.

So changing the window does two things at once. It restates your history, and it changes who the system targets going forward. Those effects arrive on different timescales, which makes the change genuinely hard to read: the reporting shift is instant, and the delivery shift takes days to work through.

There is a volume consequence as well. A narrower window means fewer attributed conversions, which means fewer signals for the algorithm to learn from. If you were already close to the threshold where campaigns struggle to exit learning, tightening the window can push you under it. The mechanics of that threshold are in the 50 conversions per week rule.

Measuring your actual purchase delay in ten minutes

Everything above depends on knowing how long your customers really take. Most people guess. It is measurable, and the answer is often not what the team assumed.

1. Take a month of orders from your backend, ideally a few hundred.
2. For each order, find the time between the customer's first recorded session and the purchase. Most analytics tools expose this as days to conversion or time lag. If yours does not, a rough version using first-seen date from your email platform or CRM works.
3. Bucket them: same day, 1 to 3 days, 4 to 7 days, more than 7 days.
4. Look at where the mass sits. If 80% of orders land in the first bucket, a 7-day window is mostly padding your reports rather than capturing genuine influence. If a third are beyond 7 days, even your widest setting is understating what the ads did, and you should be reasoning about incrementality rather than attribution.

Two practical notes. Do this per product line if you sell things at very different price points, because a £15 accessory and a £900 sofa do not share a decision curve and a blended average describes neither. And redo it after any major pricing or product change, since the delay shifts with consideration.

Attribution is not incrementality

Worth naming directly, because widening the window can feel like finding more value when it is really just claiming more credit.

Attribution asks which touchpoint gets credit for a sale that happened. Incrementality asks whether the sale would have happened anyway. Those are different questions, and no attribution setting answers the second one. A conversion attributed on a 1-day view to a retargeting ad shown to someone who had the product in their basket was probably going to happen regardless.

You do not need a formal study to keep this in view. The cheap version is to watch total backend orders and total spend across a window where you meaningfully changed budget. If spend rose 40% and total orders did not move, your attributed conversions are redistributing credit rather than creating sales, whatever the window says. That is the number that pays rent, and it is immune to every setting discussed here.

How to change it without losing a quarter of readable data

If you have decided to switch, the process matters as much as the choice.

1. Record the before, from your backend. Spend, orders and cost per order from your own systems for a normal month. This is your invariant. It does not care what attribution setting you use, which is exactly why it is the thing to anchor on.
2. Change it once, and write down the date. Annotate every report. Someone will compare across this boundary in three months if you do not.
3. Expect the reported numbers to move immediately and ignore it. The instant change is arithmetic, not performance. Do not react to it.
4. Wait at least two full attribution windows before judging. You need enough settled data to see the delivery effect separately from the reporting effect.
5. Judge on the backend numbers. If cost per real order improved, the change was good, whatever Ads Manager says about ROAS.

And do not change it in the middle of a campaign test you care about. Changing the success definition mid-test invalidates the test, because the two arms are no longer being measured the same way.

When to leave it alone

Most accounts should not be adjusting this setting often, and there are situations where it is a clear mistake.

• To make a report look better. Widening the window before a client review produces a nicer chart and a worse business, because you have just made your own numbers less comparable to everything before them. It also tends to be noticed.
• In response to one bad week. Attribution settings are a structural choice about how you measure. Changing them reactively means you will never have a stable series long enough to learn from.
• While you have a tracking problem. If your events are not arriving reliably, no attribution setting will fix that, and changing it adds a second variable to an investigation that already has one. Establish that the plumbing works first, using how to tell if your pixel is broken.

The short version

The attribution window is a lens, not a lever. It changes what you can see, and it changes what the algorithm chases, but it does not change how many people bought from you.

Pick the setting that matches how long your customers genuinely take to decide. Write down when you picked it. Then stop touching it, and judge performance on the number of real orders your own systems recorded against what you spent to get them. That figure is true under every attribution setting, which is the only reason it is worth building a reporting habit around.