One number tells the story. If your Meta CPAs rose 15–30% in March 2026, you might have cut budget, flagged a performance issue, or triggered a cross-channel reallocation. But the campaigns didn't perform worse. Meta changed what it was counting.
On March 3, Meta restructured → link to how conversions are attributed. Link clicks still earn 7-day click-through credit. But likes, saves, shares, and comments - previously bundled into click attribution - were reclassified → link to into a new engage-through bucket with a 1-day window only. The video engaged-view threshold dropped from 10 seconds to 5. The result: fewer actions qualify for conversion credit, so reported CPAs rise and conversion counts fall. Not because your campaigns degraded. Because the ruler got shorter.
The measurement definition shifted and most teams haven't caught it
This a baseline problem. Every efficiency target, ROAS benchmark, and blended CPA goal built on pre-March data is now measuring a different thing. For teams spending 00K+ per month, even a 20% CPA discrepancy can trigger budget cuts, pause decisions, or cross-channel reallocations that aren't warranted by actual performance.
The first step is straightforward: set March 3 as a breakpoint in your reporting. Do not compare April performance to February performance without adjusting for the attribution definition change. If you're using third-party measurement tools - Northbeam, Triple Whale, Rockerbox - check their changelog. Meta is now formally partnering with several MMPs to surface the gap, which is an admission that the discrepancy is real and significant.
The European cost increase hiding in your P&L
While attribution is the urgent story, European advertisers face a second, compounding issue. Starting July 1, Meta will bill location fees - Digital Services Tax surcharges applied based on where your ads are served. Austria and Turkey: 5%. France, Italy, Spain: 3%. UK: 2%. The fees are calculated before VAT, meaning VAT applies to the inflated total. They don't appear in Ads Manager reporting. They appear on your invoice.
For any team with meaningful EU/UK allocation, this is an unbudgeted cost increase that creates a gap between in-platform efficiency metrics and actual P&L. Google Ads has been billing equivalent surcharges for two years. TikTok does not yet. That asymmetry will affect budget allocation decisions before the year is out.
While Meta shifts the ground, challengers are building the case
LinkedIn and Reddit both published meaningful performance data this week and together they tell a contrarian story.
Dreamdata's 2026 LinkedIn benchmarks → link to found it's the only major paid social platform delivering positive ROAS (121%), outperforming → link to Google Search and Meta. More important than the headline number: the average B2B deal now involves 88 touchpoints across 272 days. Any team measuring LinkedIn on a 30 or 90-day attribution window is systematically under-reporting its contribution. The platform isn't underperforming — it's being measured wrong.
Reddit's MAX Campaigns → link to, now at full rollout, are delivering 17% lower → link to and 27% more conversions versus manually managed campaigns in early data. More interesting: Reddit's automation comes with genuine audience transparency - a persona breakdown showing which clusters are converting, what creative they respond to, and what communities they engage in. That's operationally rare in a world of black-box optimization.
The takeaway
The Meta attribution overhaul is the most structurally significant platform change in years - not because it's catastrophic, but because it's invisible if you're not looking. CPAs didn't rise because performance fell. Baselines rose because the definition of a conversion changed. Every budget decision made without acknowledging that breakpoint is a decision made on bad data.
Set the breakpoint. Align with your third-party measurement. And consider whether the uncertainty in your Meta data is the right moment to accelerate tests on platforms where the measurement story is cleaner.