Engage-through is a creative signal, not a measurement column.
Meta introduced the engage-through attribution bucket in March 2026 and rolled it out as default in May. It captures conversions attributable to non-click engagements: likes, reactions, comments, shares, saves, and engaged video views (now defined at 5 seconds, down from 10) that precede a later conversion 1. Previously these signals folded into click-through attribution and inflated reported click-driven results.
The mistake operators are making this month is reading engage-through as a recovered conversion number, summing it with click-through, and reporting a blended figure that looks roughly like the old number. That math is wrong. Engage-through tells you something different from click-through, and the two should not be added together for decision-making.
This is how to read it without lying to yourself.
What engage-through credits and what it does not
Engage-through credits a conversion when, within the attribution window, a user took one of the qualifying non-click actions on an ad and then converted later 2. Qualifying actions include reactions, comments, shares, saves, and engaged video views at the new 5-second threshold.
What engage-through does not credit: pure impressions (those go to view-through, which Meta deprecated as default earlier in 2026), drive-by scrolls without engagement, or any action that happens after the click attribution window closes. Engage-through has its own attribution window, typically 7-day post-engagement, configurable in the campaign setup.
The semantic difference between engage-through and click-through is meaningful. A click is a strong intent signal: the user saw the ad, decided to investigate, and arrived on a landing page. An engaged view or a save is a weaker but still informative signal: the user reacted to the ad emotionally, but did not yet move toward purchase. The two are not interchangeable, and treating them as additive overstates the strength of the demand signal.
Why engage-through inflated under the old attribution model
Under Meta’s previous attribution structure, the upper-funnel engagement signals (likes, shares, saves, engaged views) folded into the click-through column. A user who watched a video for 12 seconds, did not click, and converted 36 hours later via direct traffic was being credited as a click-through conversion. The credit was directionally reasonable (the ad did contribute) but the column was misleading.
The new structure separates these. Click-through now reflects users who clicked and converted; engage-through reflects users who engaged but did not click; view-through reflects users who saw the ad but did not engage or click. Each column carries information; none should be confused for another.
The practical implication is that your historical click-through CPA was overstated, possibly meaningfully. The new click-through CPA is closer to the truth. The new engage-through column captures part of the difference. Some part of the old click-through credit is also not migrating anywhere; it was incidental and is now uncredited 3.
Where engage-through is useful
Three diagnostic uses.
The first is creative testing. A video ad with high engage-through volume but low click-through is doing the emotional work without closing. That is sometimes a sign of a strong hook with a weak call to action. Sometimes it is a sign of the wrong audience: the creative is reaching people who are interested but not in-market. Either way, the signal helps you decide whether to iterate the CTA, rebuild the offer, or reassign the budget.
The second is awareness campaign evaluation. For brands running upper-funnel campaigns, engage-through is a more honest measure of impact than view-through ever was. It tells you the campaign produced a memorable enough impression that the user engaged with it and later took action. Aggregate engage-through volume across an awareness flight is a reasonable proxy for brand interest, and it can be compared period-over-period for the same campaign type.
The third is creative concept identification. Engage-through engagement (likes, comments, shares) is a leading indicator of which creative concepts are landing emotionally. The concept that produces 3x the engage-through of its siblings, even if it has not closed conversions yet, is the concept worth producing more variants of. This is most useful in the first two weeks of a creative test, before fatigue and conversion data overtake the signal.
Where engage-through misleads
Three traps to avoid.
The first trap is adding engage-through to click-through and reporting the sum as total ad-driven conversions. The two attribution paths are not additive in any meaningful sense. The user who clicked and converted is a different user from the user who engaged and converted. There is also overlap: a user could engage, then click, then convert, and both columns may credit the same conversion under different attribution rules. Sum them and you double-count an unknowable share.
The second trap is using engage-through to defend an underperforming campaign. “Click-through is down but engage-through is up” is, in most cases, a story an operator tells when they do not want to turn off a creative. The right test is whether downstream revenue, measured in your e-commerce platform or CRM, is holding or growing. If downstream revenue is down, the engage-through number is irrelevant and the campaign is not working.
The third trap is treating engage-through as a primary KPI for optimization. Meta’s Smart Bidding still optimizes against click-through and the configured conversion event; the engage-through column is reporting, not bidding signal. Setting team targets against engage-through volume will produce creative that wins likes and loses revenue.
How to set up your reporting view
Inside Ads Manager, open the Columns dropdown for a campaign view and select Customize Columns. Add: Link Clicks (the new strict definition), Click-Through Conversions, Engage-Through Conversions, Engage-Through Revenue (if you have value tracking), and the Conversion Count Breakdown view (which shows first-click, last-click, and data-driven attribution side by side) 4.
That column set lets you read each campaign for three things: click efficiency, engagement quality, and attribution-model agreement. The first tells you whether the click experience is working. The second tells you whether the creative is resonating. The third tells you whether the platform’s attribution model is internally consistent for the campaign.
When all three look healthy and downstream revenue is holding, the campaign is working. When click efficiency drops, engagement quality drops, and attribution disagrees with itself, the campaign is not working and the fastest fix is usually new creative.
A useful weekly read
Once a week, for each active campaign, ask three questions:
What is the trend in click-through CPA over the last 28 days, compared with the prior 28? Click-through is the most honest single number Meta now reports.
What is the trend in engage-through volume? Rising engage-through with stable click-through suggests the creative is doing more upper-funnel work. Falling engage-through with stable click-through suggests creative fatigue at the top of the funnel; new creative tests should already be queued.
What does the downstream revenue platform say? Shopify, your CRM, or your post-purchase survey is the source of truth. If Meta reports both columns down but Shopify is flat, the campaign is still doing its job and the dashboard is mostly attribution noise.
The honest read of an account uses all three. The dishonest read picks the column that looks best and reports that one.
What this changes about how the firm reports
The firm’s reporting template for Meta accounts now leads with three numbers: in-platform click-through CPA, downstream revenue per dollar spent (pulled from the client’s commerce platform), and creative freshness (days since the top-performing creative was tested). Engage-through sits in the diagnostic appendix, not on the executive summary line.
This is the read that survives any attribution change Meta ships next. The platform’s reporting will keep evolving. The underlying question, did the spend produce more revenue than it cost, depends on data the platform does not own. Reading engage-through honestly means using it for what it tells you and ignoring it where it does not.
Anchor the report on downstream revenue. Treat in-platform attribution as a diagnostic. Treat engage-through as a creative signal. That is the order of operations that survives the next ten attribution changes, too.