Meta Rewrote Click Attribution and Dropped the Engaged-View Window to 5 Seconds

Meta rolled out a new default attribution setting in March 2026, and the rollout is continuing progressively through April: 7-day click, 1-day engage-through, 1-day view, applied to campaigns optimizing for website or in-store conversions. Only real link clicks now qualify as click-through. Likes, shares, and saves moved into a renamed "engage-through" bucket. The engaged-view window for video also shortened from 10 seconds to 5.

What shifted under the AI chatbot headline

Jon Loomer's Wednesday roundup led with Meta opening its AI Business Assistant to the rest of the world. Fine. A chatbot that drafts headlines is not the thing that shows up in Monday's reporting review. Social Media Today covered the AI expansion the same day, and that was the headline most of the trade press led with.

The line buried further down is the one that actually changes what your Advantage+ dashboards read next month. Meta redrew the border of "click." For roughly a decade, a like counted as click-through. A save counted. A comment counted. Any interaction on a Reels ad could land in the click column. The new rule: only taps that actually go somewhere. If it does not open a link, it is not a click.

That sounds cleaner on paper. And in most cases I've seen, cleaner measurement is a good trade. But the transition window is where client reports get awkward, because your Q1 numbers and your Q2 numbers are now measuring slightly different things.

Engage-through is where the rest of it goes now

The interactions that used to inflate click-through did not disappear. They got moved. "Engage-through attribution" is Meta's new container for likes, shares, saves, comments, plus the old engaged-view metric from video. Dataslayer's breakdown is the clearest technical walkthrough I have seen on the new math.

Two changes inside engage-through matter more than the renaming itself:

• It applies to every ad format, not just video. So a save on a static carousel now sits in the same attribution bucket as a 5-second Reels view.
• The video view threshold dropped from 10 seconds to 5. Meta's reasoning, from Search Engine Land's coverage, is that "46% of Reels purchase conversions happen within the first two seconds of attention." That is the number they used to justify the shorter window.

Personally, I think 5 seconds is reasonable for Reels and a bit of a stretch for pre-roll. Pre-roll at 5 seconds is basically a skippable impression. But Meta is trying to make one rule that works across Reels, in-feed, and in-stream, and Reels is where the volume lives, so the policy is optimized for that surface.

Your Advantage+ reports are about to look worse (even if nothing is worse)

Here is the part teams keep missing in the rollout notes. Advantage+ Shopping campaigns optimize on purchase signal directly, so the campaign performance itself does not change with the rule. But the reported ROAS inside Ads Manager will change, because the denominator for click-attributed conversions just got smaller and the engage-through window is shorter than the click-through one.

From what I've seen across the trade press reporting, accounts that ran heavily on Reels are seeing the biggest delta. A conversion that used to get credited to a save now lands in engage-through, and engage-through runs on a 1-day window instead of 7. So a conversion that "converted" on day 3 under the old rules may show as zero attribution under the new ones. Anyway, this is the kind of thing that looks like a performance drop when it's actually a measurement change, which is exactly the kind of thing that triggers a panicked Slack from your CMO on a Tuesday morning.

So pull your last 8 weeks of Advantage+ reporting and segment by attribution bucket using the new taxonomy. Establish a fresh baseline now, before your finance team asks why April ROAS dropped 15% against March. It probably did not drop. The measurement changed.

If any of this sounds familiar, it should. Meta's Suite of Truth measurement framework from earlier this year runs on the same principle: Meta publishes the rulebook, and the rulebook is also the grading scale. You can work with it. You just can't pretend you're comparing to a neutral benchmark.

The quieter story: Advantage+ Shopping threshold dropped to 25

There is a second change hiding inside the same rollout that gets less press. The conversion threshold for Advantage+ Shopping campaigns is now 25 conversions per week, down from 50. That is a real expansion of who qualifies. Accounts that were stuck in standard sales campaigns three months ago because they could not hit 50 conversions per week can probably move over now.

If you run agency accounts, this is worth a weekly audit. Pull every client under the 50-conversion threshold who was told they could not run Shopping campaigns, and check whether they now clear the 25-conversion line. For accounts that just barely missed the old bar, switching formats this month could meaningfully improve delivery and reduce the amount of manual ad-set budget work.

Predictive Budget Allocation: claim carefully

Meta also introduced Predictive Budget Allocation, which shifts spend across ad sets based on real-time predicted conversion probability instead of distributing evenly. Meta claims 8 to 15% ROAS improvement over manual ad-set-level budgeting.

I'd hedge that claim, a lot. Meta's own internal comparisons tend to paint their automation in the best light, and "versus manually managing budgets at the ad set level" is the comparison that flatters automation the most. Most paid social managers I know already rely on ABO-then-CBO patterns that are roughly 70% of the way to automated allocation anyway, so the real comparison for an in-house team is "Predictive Budget Allocation versus existing CBO," not "versus fully manual." That comparison is narrower. Run your own A/B before you trust the 8 to 15% number for your account.

The chatbot is fine, the attribution rewrite is the story

To be fair, Meta expanding AI Business Assistant to EMEA, APAC, and Latin America is a real product shipment. Creators in those markets can now use it in their native languages, and that matters for local agencies. It is not nothing.

But for a working paid social manager, the chatbot is a tool you may try this quarter when you have 20 minutes. The attribution rewrite is a measurement change that affects every client report you write this month. The editorial choice most of the trade press made this week was to lead with the AI news and bury the attribution change. I think that choice is wrong, and I think accounts that only read the chatbot coverage are going to be caught off guard when their April ROAS lands.

Four things worth handling this week:

• Re-baseline Advantage+ reporting against the 7-day click / 1-day engage-through / 1-day view windows. Label last month's comparison as "old rules" so nobody builds a narrative on a broken comparison.
• Tell your finance partner the measurement changed before they ask why ROAS moved. The explanation is much easier to deliver in advance than in response to a question.
• Audit which of your sub-50-conversion accounts can now run Advantage+ Shopping under the new 25-per-week threshold. Move the ones that can.
• If you run Reels-heavy campaigns, expect the biggest reported-ROAS compression. Don't cut Reels budget in response. Cut the instinct to react to it.

From what I've seen in the first few weeks of rollouts like this one, the accounts that lose trust with their stakeholders are not the ones with worse performance. They're the ones who get caught explaining a measurement change after the question has already been asked. Get ahead of the question this week and you keep the trust. That is worth more than any 8 to 15% ROAS claim.

Notice Me Senpai Editorial