Meta updates its click-through attribution: what you need to know

Meta is tightening attribution: click-through conversions for website and in-store will only be attributed when there was a link click, not just any kind of interaction (like, share, save, etc.).

Conversions that were previously counted under “click” without an actual link click will move into a new bucket: engage-through attribution, which replaces and renames the old engaged-view attribution.

Why Meta is doing this (and why it matters even if you “measure in GA”)

Meta’s positioning is about simplifying measurement in a more social-first world.

The practical impact is that reporting should align better with what third-party analytics tools typically consider a “click”, because many external tools won’t treat a like or a profile tap as a click that explains a conversion.

The key point: with the rollout starting March 2026 (gradual), you may see shifts in reported performance even if real-world performance hasn’t changed, because the attribution definitions are changing.

Before vs after: what used to count as click-through

This is the part that has caused confusion for years. Historically, Meta could attribute a conversion as click-through when someone clicked any part of the ad, not necessarily the outbound link.

It appears that this could include non-link interactions, as long as the conversion happened within the attribution window.

With the update: click-through = link click. End of story.

What engage-through is (and what ends up inside it)

Engage-through is designed to capture the value of social interactions that don’t immediately send traffic to your site but may influence later conversions.

Engage-through will include:

• conversions previously counted under click-through that were driven by non-link actions (likes, saves, comments, shares, etc.)
• the “video engagement” component that used to sit under engaged-view (now renamed into the new framework)

The change that impacts video and Reels

Meta is also shortening the engaged-view threshold for video from 10 seconds to 5 seconds, citing faster behaviour patterns, especially on Reels.

Search Engine Land also reports the stats shared: 46% of purchase conversions from Reels happen within the first 2 seconds of attention.

This isn’t a footnote. It’s Meta saying: “engagement now happens faster, so we’ll recognise shorter attention”.

What you’ll see in Ads Manager: different numbers, different performance?

In practice, you’ll see a reallocation:

• some conversions that previously appeared under click-through may shift into engage-through
• some conversions may no longer be counted the way you’re used to if they no longer qualify under the new windows and definitions

So the right interpretation isn’t “Meta is taking conversions away”.

It’s: Meta is separating intent (link clicks) from interaction (engagement) more cleanly, and that changes how reports read.

How to read reporting without getting tricked by “before vs after”

You need a bit of method here.

First: for the first 2–4 weeks, avoid snap judgements like “ROAS collapsed” if you’re comparing metrics whose attribution buckets have changed.

Second: split your analysis into:

• traffic-driven performance (link click → session → conversion)
• influence-driven performance (engagement → later conversion)

Third: if your objective is leads (not purchases), it often makes less sense to attribute strong performance to non-link engagement, because the desired behaviour is usually direct: see → click → submit. This “behavioural alignment” is central to Loomer’s analysis.

New rules, better performances?

If you run Meta for performance, this update is a useful reminder: you’re not buying “conversions”, you’re buying behaviours and the signals behind them.

Meta is now separating those behaviours more clearly.

The useful question is simple: are you optimising for actions that actually move people to your site (link clicks), or are you counting social interactions as “performance” when they may be closer to influence?