What changed
Meta's click-through attribution window used to credit a conversion to an ad if the user interacted with it in any way before converting. A like, a share, a save, all of those counted as a "click." Starting with the late-March 2026 rollout for website and in-store conversion campaigns, click-through attribution requires an actual link click: the user has to leave Meta's surface and land somewhere (a website, app, lead form, or other destination).
The engagements that used to get counted didn't disappear from reporting. They moved into "engage-through attribution," which is what Meta renamed from the older "engaged-view attribution" label. That category is on by default with a 1-day conversion window, meaning most accounts now have it enabled without any deliberate action taken.
Alongside this, Meta shortened the video engaged-view qualification window from 10 seconds to 5 seconds. The stated rationale: a large share of Reels purchase conversions register within the first two seconds of a user's attention, so 10 seconds was measuring something too far downstream of the actual behavior.
Billing is untouched. This is purely a reporting change.
Who it affects
If your primary KPI is Ads Manager CPA or ROAS, you're looking at a data series break. Accounts that run heavy video on Reels, or that have high organic engagement (lots of likes/shares on ads), will see the largest shift in how conversions are distributed across attribution buckets. Anyone doing weekly or monthly performance reviews against prior-period benchmarks needs to flag the rollout date as a methodology change, not a performance change.
Accounts that have already been reconciling Ads Manager against GA4 or a mobile measurement partner (MMP) will feel this less, because those tools never credited engagement interactions as conversions in the first place.
Why it matters
The practical risk here is misdiagnosis. A media buyer who checks Ads Manager on April 1 and sees click-through conversions down 20% versus March has three possible explanations: creative fatigue, audience exhaustion, or a reclassification of conversions that were real but are now reported differently. Without knowing which it is, the wrong response (pulling budget, switching creatives, killing ad sets) costs money.
The subtler issue is that engage-through attribution is now its own visible, default-on signal. That 1-day engage-through window captures people who watched a video, didn't click, but converted later within the same day. That behavior is real and worth understanding, but it doesn't mean the same thing as a click-through conversion for budget planning. Engage-through conversions tend to skew toward users already in-market or already familiar with the brand. They represent assisted influence, not direct response. Treating them identically to click-through conversions in a CPA target will tend to over-credit brand-aware audiences and potentially inflate bids on prospecting campaigns where the actual click-to-purchase funnel matters more.
The 5-second video view window change has a cleaner implication: Meta is now attributing more weight to fast attention. A viewer who watches 5 seconds and converts the same day counts. That means the first 5 seconds of any Reels creative isn't just a hook for retention, it's now a direct attribution window for purchase credit.
The play
-
Pull a pre/post attribution breakdown immediately. In Ads Manager, filter conversion reporting by attribution setting and compare the week before vs. the week after the rollout date. If total conversions (click-through plus engage-through combined) held flat but the mix shifted, this is reclassification. If total combined conversions dropped, then you have a real signal to investigate. Target: combined conversions within 5% of pre-change volume at flat spend.
-
Audit engage-through defaults across all active conversion campaigns. Go into each campaign's attribution settings and confirm whether engage-through is on or off. Decide deliberately. If you're running a prospecting campaign where clicks are the only meaningful intent signal, consider turning it off and measuring the reported CPA impact on a clean baseline.
-
Rebuild your Ads Manager vs. third-party reconciliation baseline. Now that Meta has aligned its click definition closer to how GA4 and MMPs count, run a 2-week comparison to see how much of the historical discrepancy has closed. Use the result as your new normalization factor for budget decisions.
-
Test 2-second-first Reels creative. Structure a new creative variant where the product, offer, or payoff lands before the 2-second mark. Measure 5-second hold rate (share of viewers who pass the 5-second mark) and engage-through conversion rate separately from your control. A 10-point lift in 5-second hold rate is a reasonable bar to call the test directionally positive.
Watch-outs
The default-on engage-through window is the main trap. If you're reporting blended CPA to a client or stakeholder and engage-through conversions are inflating the total, you may be hitting a CPA target that isn't achievable once you strip out the assist credit. Check the split before you call a campaign efficient.
Also watch for historical comparisons in any automated rules or budget scripts you're running. If you have a rule that increases budget when 7-day CPA is below a threshold, and that threshold was calibrated on the old (wider) click-through definition, the rule is now working off a different denominator and may trigger incorrectly.
The WhyItWon angle
This change adds a new layer of uncertainty to what "a conversion" even means across your creative mix. Engage-through conversions, click-through conversions, and 5-second-view-credited conversions can each behave differently depending on creative format, placement, and audience temperature. Before you spend to find out which creative earns which kind of conversion signal at what cost, WhyItWon reads your existing ads, your rivals' creative, and your customer signals to score the likely winners. With Meta's attribution buckets now more granular, knowing in advance which creative is likely to drive actual link clicks versus passive engagement influence is worth more than it was last month.
Meta
