Meta has changed the rules again. If a brand's reporting still looks the same as it did six months ago, it's almost certainly looking at the wrong numbers.
What Actually Changed
Earlier this year, Meta overhauled its attribution framework in ways that caused significant reporting shifts across eCommerce. Click-through attribution for website conversions now counts only genuine link clicks. Likes, shares, saves and other social interactions no longer count as click-through conversions, they've moved into a renamed category called engage-through attribution, previously known as engaged-view attribution. Meta also shortened the video engaged-view window from ten seconds down to five, reflecting how quickly consumers now convert through Reels content.
On the surface, that sounds like tidying up. In practice, it exposed something more fundamental: how dependent most brands still are on platform-reported numbers that Meta can, and does, change without much warning.
Why the Change Happened
For years, Meta counted all click types toward click-through conversions, while third-party tools like Google Analytics and Triple Whale counted only link clicks. The result was a persistent gap between what Meta reported and what every other tool showed, eroding confidence and making cross-channel budget decisions harder than they needed to be.
Meta's update was designed to close that gap, and its own recommendation is clear: use engage-through attribution to capture the full value of social interactions, particularly the high-value behaviours unique to social platforms. For sports and apparel brands where Reels is a significant acquisition driver, the five-second engaged-view threshold means fast-converting video content is measurable in ways it wasn't before.
Meta is also now partnering directly with third-party measurement platforms including Triple Whale and Northbeam, bringing click and view data into their models. A pixel that sits outside the Meta ecosystem, tracking customer journeys across channels, isn't optional for brands running across Meta, Google, TikTok and email at once. Without it, channels get credited or discredited based on whichever platform's model was most recently updated.
The question worth asking right now: where does Meta-reported revenue diverge from actual revenue? If that gap is significant and growing, there's a measurement problem, and budget decisions are likely being made on the back of it.
The Risk Hiding in a Healthy ROAS
For large sports and apparel brands, paid social can quietly become a retention channel dressed up as an acquisition one. Existing customers click the ads, convert, and Return on Ad Spend (ROAS) looks healthy, while the brand is really just paying to reach people who'd have purchased anyway.
Tightening audience definitions and building proper exclusion lists isn't a technical detail, it's a commercial decision. Excluding existing customers, recent purchasers and high-frequency site visitors from prospecting campaigns changes what Meta is actually doing with the budget: stop subsidising loyalty, start funding growth. This matters more in sports and apparel than most categories, since a customer buying a performance running kit isn't the same person buying lifestyle streetwear, even when they're the same individual.
Creative Is the Lever Still in Reach
Media buying at scale has become increasingly automated. Creative is the one lever brands still fully control, and most aren't pulling it hard enough.
Meta is now rolling out the ability to upload up to ten images and videos within a single ad unit, meaning five ads can each carry ten assets with five copy variations per persona, a level of creative density previously only available to brands with serious production budgets. The brands building clear personas, genuine creative diversity and enough asset volume to generate meaningful data now will have a structural advantage over those still running three to five creative variants per campaign.
Three Places to Start
- Measurement: review attribution settings in Ads Manager, make sure engage-through attribution is active, and audit the gap between Meta-reported and actual revenue using a platform-neutral tool.
- Audience: build a full exclusion architecture for prospecting campaigns, removing existing customers, recent purchasers and high-intent site visitors from new customer acquisition spend.
- Creative: map personas before briefing creative, identify gaps in the current asset library, and build toward the volume Meta's multi-asset ad units now make possible.
Want the Full Picture?
This is exactly the ground covered in Peak Performance, Velstar's guide to mastering eCommerce in the sports and apparel sector. The Paid Social chapter goes further into measurement, audience strategy and creative testing, alongside chapters on affiliate marketing, design, SEO, email and data.
Download the eBook to read the full breakdown, chapter by chapter.