The Default Meta Ads Attribution Window in 2026

The Meta Ads attribution window default in 2026 is 7-day click, 1-day engage-through, 1-day view. Three legs, three different behaviors, and only two of them are yours to change.

That is the answer. The rest of this post is the reference: what each leg actually counts, which parts of the setting you can select and which are fixed, and which number belongs on the report you hand to a client or a CEO.

If your conversions already fell and you are trying to work out whether the drop was real, that is a different job — the attribution-change diagnostic walks through separating a measurement artifact from a genuine regression. This post assumes nothing dropped and you just need to know what the setting is now.

Why this matters

The default moved twice this year. Most reporting templates, agency SOPs, and internal dashboards were built against the old set and nobody rewrote them.

That creates a specific, boring failure: two people pull the same campaign, get different conversion counts, and spend an hour arguing. One is reading the default. The other is reading a saved view with a window that no longer exists. Neither is wrong on purpose.

The second-order problem is worse. Because engage-through was carved out of click-through in March, "clicks" on your report does not mean what it meant in February. Any month-over-month comparison that spans the change is comparing two different definitions, and nothing in the interface warns you.

The three legs of the default

Each leg answers a different question: how did the person interact, and how long after that interaction does a conversion still get credited?

| Leg | Window | What triggers it | Selectable? | |---|---|---|---| | Click-through | 7 days | A link click — the person navigated to your destination | Yes: 7-day or 1-day | | Engage-through | 1 day | A non-link interaction: like, share, save, comment, or a video watched ≥5 seconds | No — fixed at 1 day | | View-through | 1 day | The ad was served and seen, no interaction | Yes: 1-day, or off |

The click-through definition is the part that changed most. According to Meta's own announcement, the company is "changing the definition of click-through attribution for website and in-store conversions to exclusively include link clicks." Before that, a share or a save could land in the same bucket as a click to your site.

Everything displaced by that narrowing went into engage-through. Per PPC Land's reporting on the rollout, the change began rolling out "later this month for campaigns optimizing for website or in-store conversions" in March 2026, campaign by campaign rather than as a single cutover — which is why some accounts saw it weeks apart.

The engaged-view threshold moved in the same release. Meta updated "the definition for an engaged view for a video ad from 10 seconds to 5 seconds," justifying it with the claim that 46% of online purchase conversions with Reels happen within the first two seconds of attention. Halving the threshold means more video views now qualify, so engage-through volume went up for reasons that have nothing to do with performance.

The full mechanics of the click/engage split are in the engage-through explainer. This post only needs you to know which bucket is which.

What you can actually select

Here is the distinction that catches people: the attribution setting in Ads Manager is not a three-way control. You pick a click/view combination. The engage-through leg comes along at 1 day whether you want it or not.

Per JetFuel's breakdown of the 2026 windows, the selectable combinations are:

| Selectable setting | Use it when | |---|---| | 7-day click + 1-day view (default) | Standard ecommerce with a multi-day consideration cycle | | 7-day click only | High-AOV or high-consideration products where view-through credit inflates the picture | | 1-day click only | Same-session, binary conversions — lead forms, app installs, low-ticket impulse buys | | 1-day view only | Rare. Awareness measurement, not performance reporting |

And here is what is no longer on the menu at all. On January 12, 2026, Meta deprecated ten conversion windows. Per Supermetrics' change log, the removed set includes 7d_view, 28d_view, and combinations like 28d_click + 7d_view, all of which "will no longer return results after that date." The actions_7d_view, action_values_7d_view, actions_28d_view, and action_values_28d_view fields went with them.

The same change capped certain unique-count and hourly fields at 13 months of history. If you have a year-over-year dashboard that reaches further back than that, it is going to start erroring rather than quietly returning zeros.

Which one goes on the report

Pick one setting, write it down, and never change it mid-quarter. Which one depends on who reads the report.

Act today — set the standard:

• Report on 7-day click as the headline number. It is the leg tied to someone actually arriving at your site, which is the only definition a finance team or a founder will accept without a footnote.
• Split engage-through into its own column. Do not roll it into total conversions. DataSlayer's guidance is blunt on this point: presenting a combined total "makes it impossible to evaluate the quality of each attribution type and will cause confusion when stakeholders compare the numbers to GA4."
• Label the window on the report itself. Literally print "7-day click / 1-day engage-through / 1-day view" in the header. This single line kills most reporting disputes.

This week — reconcile:

• Compare 7-day-click conversions against your backend (Shopify, your CRM, GA4) for the same period. You are looking for a stable ratio, not a match. The ratio is the useful artifact.
• Re-save every dashboard view built before January. Anything referencing a removed window is returning empty, and empty often renders as zero.

Monitor:

• Watch engage-through as a share of total conversions over time. A rising share on flat spend usually means creative mix shifted toward video, not that demand improved.
• Re-check the ratio each month. If the backend ratio drifts while the setting is unchanged, something upstream moved — pixel, CAPI, or delivery.

What this looks like on a real report

Take a DTC account spending $40,000 a month, mostly on Reels and video. Numbers here are illustrative, but the shape shows up repeatedly in video-heavy accounts.

| Column | Before March | After March | Why | |---|---|---|---| | Click-through conversions | 820 | 690 | Shares, saves and likes moved out | | Engage-through conversions | — | 165 | New bucket, includes 5-second video views | | Total (if you combine them) | 820 | 855 | Looks like growth. It isn't | | View-through conversions | 95 | 95 | 1-day view, unchanged |

Read the combined row and you report a 4% conversion lift. Read the click-through row and you report a 16% decline. Both come from the same export. The account did the same thing all month.

The honest read is the third option: click-through fell because its definition narrowed, engage-through appeared because it was created, and neither number tells you anything about performance until you compare them to the backend. That is why the label in the report header matters more than the number under it.

This is the class of problem worth removing from the human entirely. Good Morning's Meta Ads reporting software pins the attribution setting, keeps the click and engage-through legs separate, and hands back an urgency-tiered action list — no analysis required, because the analysis is pre-done.

Common mistakes

1. Combining click-through and engage-through into one "conversions" number. The most common error, and it manufactures growth that did not happen.
2. Comparing across the March boundary without a note. Any period that spans the rollout is comparing two definitions. Flag it or don't run it.
3. Assuming a 7-day engage-through window exists. It doesn't. Engage-through is fixed at 1 day, and asking for a longer one wastes an afternoon.
4. Leaving pre-January saved views in place. A request for 28d_view returns nothing, and a nothing that renders as 0 looks like a catastrophe rather than a deprecation.
5. Changing the window to make a bad month look better. Switching from 7-day-click-only to the default mid-quarter adds view-through credit and breaks every prior comparison. If you change it, restate the history.
6. Treating the setting as a performance lever. It changes what gets counted, not what happens. Billing is unaffected — Meta states plainly that "there will be no change to how advertisers are billed."

FAQ

What is the default Meta Ads attribution window in 2026? 7-day click, 1-day engage-through, 1-day view, for campaigns optimizing toward website or in-store conversions. Engage-through arrived in March 2026 as a separate leg.

Can I still use a 28-day attribution window? No. The 28-day-view and 7-day-view windows, plus eight related combinations, were removed on January 12, 2026 and no longer return data. There is no workaround at the API level.

Is engage-through the same as the old engaged-view? Engage-through absorbed engaged-view and extended it. It now covers likes, shares, saves, comments and video views of at least 5 seconds — down from the previous 10-second threshold.

Should I turn view-through off? For high-AOV or high-consideration products, 7-day-click-only removes a credit source you cannot verify against your backend. For standard ecommerce, leave the default. Decide once and hold it for a full quarter.

Why don't my Meta numbers match GA4? They are not supposed to match, but the gap should be narrower than it was — the point of restricting click-through to link clicks was closer alignment with analytics platforms. Track the ratio rather than chasing a match. If you want a starting structure, the Meta Ads report template is already built around the current metric set.

My conversions dropped after the change. Is it real? Possibly, and possibly not. Work through the conversion-drop diagnostic before touching a budget, and run an audit of the account if the front of the funnel looks unchanged.

Lock the setting, then stop thinking about it

The attribution window is a reporting decision, not an optimization decision. Pick 7-day click as your headline, keep engage-through in its own column, print the setting on the report, and re-derive the backend ratio monthly.

Then get out of the spreadsheet. If you manage more than a couple of accounts, agency Meta Ads reporting that pins the window and returns a pre-diagnosed action list per account is the version of this that scales — a list of what to do today, not another dashboard to interpret.