Meta attribution setting change: why your results moved and how to report it

What is the attribution setting in Meta Ads?

The attribution setting is the rule Meta uses to decide which conversions count for an ad set: how many days after a click, an engagement or a view a conversion can still be credited. It sits at ad set level, near the performance goal, and it applies to conversions that happen off Meta, such as purchases or leads on your website.

For ad sets that optimize for website conversions, the default in 2026 is 7-day click-through, 1-day engage-through and 1-day view-through. You can narrow each part:

Two more choices sit next to it. The attribution model can be Standard or Incremental; Incremental, rolled out in 2025, locks the windows. The conversion count can be all conversions or first conversion. A 28-day click window is no longer something you can optimize for, but you can still see it in reports.

The setting is only one of the reasons Meta and other tools report different numbers. The overview of marketing attribution models covers the wider picture; this article is about what happens when the Meta setting itself changes.

Why do my Meta results change when the attribution setting changes?

Because the setting changes both what is counted and what Meta optimizes for. Switching from 7-day click to 1-day click removes from the Results column every conversion that happened between day 2 and day 7 after the click, and it also tells the delivery system to look for people who convert faster.

So one change can show up in three ways:

• Reporting only: the same sales happen, but fewer are credited to the ad set. Store or CRM numbers stay flat while Meta's numbers drop.
• Delivery: Meta starts showing ads to a different mix of people, so real results can go up or down after a few days.
• Mixed comparisons: a campaign whose ad sets use different settings adds up numbers that follow different rules, and a month-over-month chart quietly compares two definitions of a conversion.

Meta has also changed the rules on its side. From June 10, 2025, Meta's developer documentation says the Insights API returns attributed values based on the ad set's attribution settings, and that the API default can differ from Ads Manager. On January 12, 2026, the 7-day view and 28-day view windows stopped being returned by the Insights API. In March 2026, click-through started counting only link clicks; conversions after other clicks, such as likes, comments or profile taps, moved to engage-through. Many accounts saw fewer click-through conversions without any edit by the team, and billing was not affected.

A drop like this in the Results column can be a real loss or only a new rule for counting. Before you look for causes in creatives or audiences, find out which of the two it is.

How do I compare attribution settings in Meta Ads Manager?

Use "Compare attribution settings" in the "Columns" menu of Ads Manager. It adds one column per window (for example 1-day click, 7-day click, 28-day click, 1-day view and 1-day engaged-view) without changing how the ad set is optimized, so you can recount any period under the old rule.

1. Open Ads Manager at the level you report on (campaign or ad set) and set the same date range as your report.
2. Check the "Attribution setting" column next to Results to see which rule each row uses. Different values inside one campaign are a warning sign.
3. Open "Columns", choose "Compare attribution settings" and tick the windows you need, including the one used before the change.
4. Compare the old-window totals before and after the change. If the old-window number stays flat while Results fell, the drop is mostly reporting.
5. For a split by type, use the breakdown by attribution setting. It shows rows for 1-day click, 2–7 day click, 1-day engagement and 1-day view, based on the ad set's own setting.

The breakdown only splits what the current setting already counts. To see conversions outside it, such as 28-day click, you need the compare columns. If you pull Meta data into Looker Studio or a spreadsheet through a connector, check which windows the connector requests, since the 7-day view and 28-day view windows are no longer returned by the API.

How do I prove which change caused the drop: the attribution setting or real performance?

Put every change from the days before the drop on one timeline, check the attribution setting before anything else, then test one hypothesis at a time. An attribution change is a measurement change, so it belongs in the same step as ruling out tracking.

1. Same window, same columns
2. List every change
3. Recount under the old rule
4. Test one cause at a time

Start with the account's activity history (see the Meta Ads activity history guide): look for duplicated or new ad sets, edits to the performance goal and anything that touched attribution. A duplicated ad set is a common trap, because the copy is a new ad set and someone may pick a different window while setting it up. Then add what Meta does not record: a new checkout, a consent banner, a CRM change, a price change, and the dates Meta changed its own rules.

Next, recount the period with "Compare attribution settings" under the old window. If the gap closes, the attribution change explains the drop. If it doesn't, look at what else moved in the same days; the guide to explaining a conversion drop to a client covers that second step.

With every change on one timeline, you would see the duplicated ad set with a 1-day window sitting a day before the drop, instead of reading the fall as worse performance. A shared marketing change log in Changeline keeps that timeline per client: Meta Ads edits, attribution switches included, are logged by the team with date, owner and reason (Meta Ads import is coming soon), and Google Ads changes are imported automatically.

Example: on the monthly call, a client asks why Meta purchases fell 22% while store orders held steady. Lining up the changes showed that on Tuesday the team duplicated the main ad set to test a new audience, and the copy was set to 1-day click instead of 7-day click. On Wednesday the original was paused. From then on, most purchases that came two or more days after the click were no longer credited. Recounted at 7-day click with the compare columns, purchases were within 4% of the old level. The report got a note, and the next test kept the original setting.

How should I note an attribution setting change in client reports?

Write it down the day it happens, and show the old and the new rule side by side for at least one full reporting period. A report that changes its definition without saying so will be read as a change in results.

• Add the setting to the report footer or a notes slide, for example "Meta results: 7-day click, 1-day engage-through, 1-day view", with the date it last changed.
• For the month of the change, show Results under the new setting and the old window from the compare columns.
• Never compare months that use different settings as if they were the same metric. Recount the earlier month or mark the break on the chart.
• Treat March 2026 as a break of its own: from then on click-through means link clicks only, so year-over-year comparisons cross two definitions.
• Explain why platform numbers differ from GA4 or the store: GA4 cannot see Meta's view-through conversions.

A dated note in GA4 helps the next person who opens a chart understand why Meta-sourced numbers moved (see the GA4 annotations guide). For the full structure of a client report, see agency client reporting.

When should you change the Meta attribution setting?

Change it when the business question changes, not to make a report look better. A shorter window can suit offers people buy on impulse; a longer one suits higher-priced products that people consider for a few days.

A few rules keep the change readable:

• Change it at the start of a new campaign or test, not in the middle of a flight. Because the setting also affects delivery, results after a mid-flight switch mix two causes. The post on Meta learning phase and significant edits explains why big edits make the following days harder to read.
• Keep one setting per campaign so totals add up under one rule.
• Before turning off view-through or engage-through, check in the compare columns how much volume they bring. If many conversions come from them, the change will look like a drop.
• If you move to the Incremental model, log it as a change of its own: it replaces the windows and counts conversions differently.

Whatever you choose, the setting is a decision someone made on a date. Treat it like a budget or bid change: record who changed it, when and why, so the next drop in the Results column doesn't start a week of guessing.