If your Meta numbers felt off in July, you're not overthinking it.
Three separate changes hit Meta advertising simultaneously this month — each one quietly distorting a different part of your dashboard. Combined, they mean that the spend numbers, the audience sizes, and the reach metrics you're looking at right now in Ads Manager are all measuring something slightly different than they were in June. And in two of the three cases, Ads Manager gives you zero indication that anything changed.
This is not a platform outage. It's not a tracking problem you caused. It's structural — three definitional and policy shifts that landed at the same time, with no banner notification inside your account.
For ecommerce brands in fashion, swimwear, and health products who make budget decisions based on what Ads Manager shows, this matters. Here's exactly what changed, how each one is distorting your numbers, and what to do about it.
Three Changes. One Month. Every Metric Affected.
Let's name them before we go deep:
-
EU Location Fees (July 1) — Meta started charging Digital Services Tax pass-throughs on ads delivered to UK and EU audiences. These fees don't appear in Ads Manager. They appear on your invoice. So your dashboard ROAS is inflated relative to your actual cost.
-
Off-Platform Data Opt-Out Removal (July 2026) — Meta removed the setting that allowed users to disconnect their off-platform browsing and purchase activity from their Meta account. People who had previously opted out are now back inside your retargeting audiences — silently, without any action on your part.
-
Legacy Reach and Impression Metrics Deprecated (June 15) — Meta retired approximately 85 legacy metrics across its reporting API and replaced them with a Media Views and Media Viewers model. The new metrics count a different underlying event, so your reach and impression numbers are no longer comparable to anything before mid-June.
If you ran any of these through individually, each one is manageable. The problem is they all landed within weeks of each other, and they hit spend data, audience data, and reach data simultaneously. The result: almost every top-level number in your dashboard is measuring something different than it was in Q2.
Change #1: Your EU Spend Is Wrong in Ads Manager
Starting July 1, Meta is charging location-based fees on ads delivered to audiences in six countries: UK (2%), France (3%), Italy (3%), Spain (3%), Austria (5%), and Turkey (5%). These are pass-through charges for the Digital Services Taxes (DSTs) those governments impose on large tech platforms.
The fee structure is straightforward. The reporting gap is not.
Ads Manager does not include these fees in your campaign spend data. Your campaign-level cost, CPA, and ROAS figures inside Ads Manager still reflect only your ad budget spend. The location fees are added on top — but they only appear on your monthly invoice in the billing hub.
What this means in practice: if you're running any scale into the UK or EU, your Ads Manager ROAS is overstating your actual efficiency right now.
The math is simple. If you spent $10,000 in July with 30% of delivery going to UK and EU audiences, you could have $150–$250 in location fees sitting on your invoice that doesn't appear anywhere in your campaign dashboard. Pull your Ads Manager ROAS against that $10,000 and you'll get one number. Calculate against your true total cost and you'll get a lower one.
For fashion and swimwear brands who sell internationally, or health brands running UK and European prospecting, this gap compounds quickly. The $5,000 you think you're spending in France is actually costing $5,150. Your CPA looks better than it is.
The fix: Stop using Ads Manager as your source of truth for spend if you're running UK or EU delivery. Pull actual invoiced spend from the Meta billing hub for any total cost calculation — MER, blended CPA, true ROAS. The breakdown we covered when these fees were announced walks through how to factor them into your budget planning.
Change #2: Your Retargeting Audiences Grew — Without You Doing Anything
In July 2026, Meta removed the "Your activity off Meta technologies" control that previously allowed users to disconnect their off-platform browsing and purchase data from their Meta accounts. Users who had exercised that opt-out — whose Pixel visits, app activity, and purchase signals were effectively invisible to your Custom Audiences — are now visible again.
Meta was explicit: no new data is being collected. What changed is whether previously opted-out users appear in your Website Custom Audiences and lookalike seeds.
The immediate effect: retargeting pools grew. If you have a Website Custom Audience of recent site visitors, it is likely larger now than it was in June, without any change in your actual traffic. Lookalike seeds built from customer lists or purchase events became more representative of the full population of people who took those actions.
This sounds like good news, and in some ways it is. But here's where it gets complicated for the brands we work with.
For fashion and swimwear brands running seasonal campaigns: Your retargeting audience now includes people who visited your site months ago but whose signal was blocked. They may be lower-intent than your recent visitors. If your retargeting ROAS drops in July despite your creative staying the same, this is one explanation — you're now reaching a broader, less warm audience within the same audience definition.
For health and supplement brands: More concerning is the frequency and overlap dynamic. If your retargeting pool grew silently, your frequency reporting and CPM may shift as Meta redistributes impression delivery across a larger audience. Watch for frequency drops paired with CPM changes — that's the pool expansion showing up in your numbers.
The broader issue: most brands assume audience size changes only when they change something. That assumption is wrong as of July 2026. Meta can — and did — expand your retargeting eligibility pool without any action on your end.
The fix: Pull your Website Custom Audience sizes now and compare them to your most recent saved benchmark from Q2. If they grew materially without a corresponding traffic increase, you're seeing the opt-out removal in action. Factor a broader, potentially less-warm audience into how you interpret July retargeting performance.
Change #3: Your Reach and Impression Numbers Changed What They Mean
Effective June 15, 2026, Meta deprecated roughly 85 legacy metrics across its reporting API — including Post Reach, Page Reach, Story Impressions, Video Impressions, and 3-second Video Views. These were replaced with a Media Views and Media Viewers framework.
The new metrics count a different underlying event.
The old model: an impression was counted when content was delivered to a feed, whether or not it was visually rendered on the user's screen. The new model: a media view is counted when content is actually rendered — visually displayed — in the viewport.
This is a more accurate measurement. It's also not comparable to your historical numbers.
Depending on how your content performs, you may see reach numbers appear lower (unique viewers in the new model) or impression counts that look higher (repeat views are now counted). Neither means your actual reach changed. It means the metric definition changed.
For anyone running brand awareness or upper-funnel objectives where reach and impressions are the primary KPIs, this creates a reporting gap that is easy to misread as a performance drop.
For fashion brands running creative testing: Thumb-stop rate, hook rate, and video view metrics are all calculated off impression-based denominators. If the denominator definition changed, your derived ratios changed with it. A thumb-stop rate you were benchmarking against June data is no longer a clean comparison.
The fix: Establish a new baseline for all reach and impression-derived metrics from July 1 onward. Treat any comparison to pre-June 15 numbers as apples-to-oranges until you can re-anchor. If you use third-party reporting tools, check whether those tools have updated their Meta API integration to use the new metric definitions — many hadn't as of mid-July.
Why All Three Together Are Dangerous
Individually, each of these changes is annoying but manageable. Together, they create a specific risk: you make a wrong budget or creative decision because you're interpreting a dashboard artifact as a real performance signal.
The scenario that concerns us most: an ecommerce operator sees ROAS declining in July. They look at their dashboard. Reach is down (metric redefinition). Retargeting is underperforming (pool expanded with lower-intent users). EU CPA looks worse than expected (because actual costs are higher than dashboard shows). They conclude creative fatigue or a campaign structure problem and start pulling levers — pausing campaigns, refreshing creative, restructuring ad sets.
None of that addresses the real issue. The issue is that Ads Manager is showing them a version of their numbers that doesn't reflect actual performance.
The correct read: some of your July "decline" is real, and some of it is definitional. The job is to separate them before you touch anything.
What to Pull, What to Ignore, and How to Rebuild Your Real Numbers
Here's the practical layer — what to actually do with each of these.
For spend and ROAS:
- Pull invoiced spend from the Meta billing hub (not campaign-level Ads Manager spend) for any UK or EU delivery
- Recalculate your true blended ROAS and MER using total invoiced cost, not Ads Manager cost
- If you don't track MER separately from ROAS, July is a good month to start — the gap between the two just got larger
For audience performance:
- Pull audience size history on your key Website Custom Audiences and compare to your last Q2 benchmark
- If retargeting ROAS dropped in July, segment by audience recency before diagnosing creative or offer — you may simply be reaching a broader, colder slice of the audience
- Don't build new lookalike seeds from purchase events in the first two weeks after the opt-out removal; let the seed stabilize before extracting new lookalikes
For reach and impression metrics:
- Establish July 1 as your new benchmark baseline for all reach, impression, and video-view metrics
- Flag any pre-June 15 reach or impression data in your reporting as under a different definition — don't blend it with current data
- For creative performance comparisons, anchor to hook rate, video-through rate, and downstream conversion events (which are not affected by this deprecation) rather than raw view counts
The principle underneath all of this: Right now, in July 2026, Ads Manager is not a single source of truth. It's three different measurement frameworks running simultaneously, and only one of them (invoice-based spend) is outside Ads Manager entirely. The brands that navigate this well aren't the ones who trust the dashboard. They're the ones who know which number to pull from where.
What This Means for the Rest of H2
July's reporting distortions are temporary in one sense — the definitions will stabilize, and you'll have new baselines to work from. But the underlying direction Meta is moving in is permanent.
Ads Manager has been drifting away from transparent, advertiser-controlled reporting for two years. The H2 2026 automation push is part of this — more AI control, less signal handed back to advertisers. July's simultaneous changes are three more steps in that direction.
The brands that hold up over the next two quarters are the ones building reporting infrastructure that doesn't depend on Ads Manager as the sole source of truth. That means tracking MER from invoiced spend, maintaining audience-size benchmarks independently, and keeping creative performance anchored to downstream conversion data rather than top-funnel impression metrics.
That's not a complex system to build. But you have to decide to build it — and July 2026 is a reasonable forcing function.
What to Do Right Now
If you're an ecommerce brand running Meta ads with any UK or EU delivery, pull your July invoice today and compare actual total spend to what Ads Manager shows. That gap is your starting point.
Then check your key retargeting audience sizes. If they grew without a traffic explanation, you're already inside the second change.
If you have questions about how these three shifts are affecting your specific account — where the real numbers are versus what the dashboard is showing — this is exactly the kind of clarity we provide in a strategy session.
Book a strategy call here. No pitch. Just a clear picture of where your numbers actually stand.
