AT A GLANCE — MAY 2026 KEY FINDINGS:
• Meta rebuilt attribution — passive engagement (likes, saves, shares) no longer counts as conversions
• Google restructured GA4 attribution in April — reported ROAS may have shifted with no campaign changes
• AI Max expanded to Google Shopping — Google is now writing your ad copy from your product feed
• TikTok launched its Attribution Portfolio — top-funnel impact still undervalued by up to 10.7x in last-click models
• Wicked Reports benchmark: average 23% gap between platform-reported and verified new customer CPA in May 2026
• DTC customer acquisition costs up 40–60% since 2023 — every attribution error is now more expensive
May 2026 was one of the most consequential months for paid media measurement in recent memory.
Three major platforms moved simultaneously — Meta, Google, and TikTok — each making changes that independently would warrant attention. Together, they signal something bigger: the era of trusting platform-reported numbers as a primary decision-making tool is over.
I'm going to break down exactly what happened, what it means for your numbers right now, and what you need to be watching in June. No fluff. No broad advice about "embracing AI." Just the specific changes and what they cost you if you miss them.
Meta didn't send a press release. They rarely do. But two changes in May are already distorting reporting for most eCommerce brands running paid social.
Buy Now Pay Later products are now explicitly classified under Meta's Credit Special Ad Category. If your brand promotes BNPL payment options — Klarna, Afterpay, Affirm, or any similar checkout feature — you're now subject to credit advertising restrictions whether you selected that category or not.
What this means in practice: targeting restrictions have tightened, financial disclaimers are now required in ad copy, and brands that promoted BNPL as a standard checkout feature without HEC classification are now out of compliance. Meta has implemented 47 policy updates across Facebook and Instagram advertising in 2026 alone — the BNPL reclassification is the one most eCommerce brands aren't prepared for.
If you're running catalog ads or collection ads that mention payment options, audit those now. A compliance flag at the account level doesn't just pause individual ads — it affects account health scores and can throttle delivery across your entire account.
Meta has updated click-through attribution to count only real link clicks. Likes, saves, shares, and comments are now tracked separately under a new "Engage-through attribution" category and are no longer bundled into conversion data.
This means a brand that previously reported a 4.2x ROAS may now see 3.1x — not because sales dropped, but because passive engagement is no longer counted as a conversion event. If you're still running year-over-year comparisons or benchmarking against Q1 2026 numbers, stop. Those comparisons are meaningless until you've normalized for the attribution change.
The maximum retention window for a custom audience built from the "Purchase" event is reportedly jumping from 180 days to 730 days. This sounds like a win — and for retargeting, it is. But for new customer acquisition tracking, a wider purchase audience window means more existing customers are being swept into prospecting campaigns. Without a hard cap on existing customer spend in Advantage+, this makes the over-reporting problem worse, not better.
Starting June 22, 2026, Facebook will replace Nielsen's DMA with Comscore Markets for automotive ads. If you're not in automotive, this doesn't directly hit you — but it signals Meta's broader move toward rebuilding its audience measurement infrastructure. Every time Meta replaces a measurement partner, the transition period creates reporting inconsistencies. Watch your audience size estimates and reach metrics in June for any unexplained shifts.
Google's AI Max is expanding to include Shopping campaigns and travel-specific ad formats, with the new AI Brief tool to guide ad messaging and audience targeting. This is significant for eCommerce brands. Here's why.
AI Max for Shopping will let text ads show from Google-generated copy based on Merchant Center data. Searchers won't necessarily see image-based product listing ads, even though it's a Shopping campaign.
Translation - Google is now writing your ad copy from your product feed. You have limited control over what it says, and the creative it generates may not reflect your brand positioning, your compliance requirements, or your current promotions. For health and wellness brands especially, Google-generated copy that makes a product claim you didn't intend is now a real operational risk.
The larger issue for attribution : AI Max is becoming a more important part of Search campaign strategy, with ads eligible to appear above, below, and in some cases within AI Overviews. When your ad appears inside an AI Overview — a summary page, not a traditional search results page — the user journey is fundamentally different. The path to conversion is longer and less direct. If your attribution model is built around click-through windows, you're going to systematically undercount conversions from AI Overview placements.
A lot of marketers are searching for "GA4 attribution model restructure April 2026" because reporting behavior, conversion controls, and attribution analysis inside Google Analytics changed materially in 2026. The practical shift is not one single button rename — it is a broader structural change in how GA4 handles key event reporting, conversion management, cross-channel reporting, and alignment with Google Ads.
Here's what's actually broken for most accounts: The GA4 update in April 2026 changed the default attribution model and adjusted how credit is distributed across touchpoints in multi-channel conversion paths. Your reported conversion values and ROAS may shift even if nothing about your campaigns changed.
If your Google Ads performance looked materially different in April and May compared to Q1, this is the most likely explanation. The platform didn't change how well your ads work. It changed how it counts what they do.
The practical test: pull your blended MER for April and May from your actual revenue data. If MER held flat while Google-reported ROAS shifted, that's a measurement change, not a performance change. If MER also shifted, that's a real signal.
GA4 revenue typically runs 20–30% below Shopify revenue due to iOS privacy restrictions and ad blockers. Use your payment processor as the revenue source of truth. Never use GA4 as your finance number. This has always been true. In May 2026, after the April attribution restructure, it's more important to internalize than ever.
Performance Max is gaining better reporting visibility, with channel-level visibility letting advertisers now get a better view into where Performance Max is spending budget across channels. This is genuinely useful progress. For the first time, you'll be able to see where PMax is actually allocating spend — not just the aggregate result. Use this to check whether PMax is cannibalizing your branded search spend or over-indexing on bottom-funnel retargeting at the expense of new customer reach.
TikTok's evolving measurement suite gives advertisers a more complete view of ROI, capturing impact across every touchpoint. On TikTok, discovery leads directly to results, influencing the entire path to purchase to generate demand and drive sales.
The honest read on TikTok attribution in 2026: TikTok's impact is often undervalued in last-click attribution. Marketing Mix Modeling reveals TikTok can be undervalued by up to 10.7x because it excels at top-funnel awareness that converts later through other channels.
That's not a reason to trust TikTok's self-reported numbers any more than Meta's. It's a reason to have independent attribution that captures the full path — because if TikTok is influencing purchases that get credited to Google or direct, you're making channel allocation decisions based on fundamentally wrong data.
GMV Max represents TikTok's effort to merge media, creator content, affiliate activity, and shop mechanics into one performance engine. For sellers active in TikTok Shop markets, it is becoming harder to ignore. The attribution complexity inside TikTok Shop — where a purchase can be influenced by a creator video, a paid ad, an affiliate link, and a live shopping event all in the same session — is unlike anything existing attribution models were designed to handle.
TikTok Market Scope is expanding with Industry Analysis, Ecommerce Insights, and Creative Insights modules. More importantly, TikTok is adding pro features that account for a wider set of seller costs, such as affiliate costs, coupons, and platform fees, to produce a more realistic profitability view. This matters because GMV as reported by TikTok has historically not accounted for affiliate commission costs — making profitability look significantly better than it is. The updated reporting should close that gap. Watch whether your TikTok-reported margins shift materially in June when these features roll out.
Three platforms moving simultaneously on attribution and measurement in the same month isn't a coincidence. It reflects a structural shift that's been building for years.
The era of passive attribution — set up a pixel, run ads, trust the dashboard — is over. What's replacing it is an environment where:
Meta overhauled attribution in Q1, Google restructured GA4 in April, and TikTok launched its Attribution Portfolio in May. Each change introduces a transition period where historical comparisons break. If you're using any prior period as a benchmark right now, you need to verify that the measurement methodology hasn't changed between the two periods.
Performance Max campaigns now account for 67% of all Shopping spend. AI Max is expanding to Shopping. TikTok's Smart+ and GMV Max are automating spend allocation across creators, affiliates, and paid ads simultaneously. These systems are optimizing for outcomes — but they're reporting on their own performance. That's the fundamental conflict of interest that hasn't changed regardless of how sophisticated the AI gets.
CAC increased 40–60% since 2023 for DTC brands. An overwhelming 88% of subscription brands are seeing higher acquisition costs. In that environment, a 15% attribution error isn't a rounding issue. It's a budget decision that costs you real growth.
Based on data from Wicked Reports customers running between $50K–$500K/month in paid spend, here's what we're seeing across accounts in May 2026:
Average gap of 23% — meaning brands are being told their new customer acquisition costs 23% less than it actually does
The divergence between platform-reported performance and actual business performance is the widest we've measured since the original iOS 14 rollout in 2021.
Given everything that moved in May, here's the specific audit every eCommerce operator should run before making a single budget decision for June.
Pull your actual new customer count from your CRM or Shopify for March, April, and May. Compare against what each platform reported as new customer conversions for the same periods. Document the gap. This is your working attribution error rate.
If you're running ASC without a hard cap on existing customer spend, you are almost certainly over-indexing on retention and under-investing in acquisition — and reporting it as acquisition performance. Set the cap at 10–20% or lower depending on your growth stage.
If your property has fewer than 400 conversions for a key event, GA4 automatically falls back to last-click attribution without notification. Many brands are running data-driven attribution in settings but getting last-click results in practice, with no indication anything changed. Verify your model is actually activated.
Take one geography or cohort segment. Pause paid ads for two weeks. Measure the actual revenue difference against your control group. That delta — not your ROAS — tells you the true incremental value of your spend. In the current measurement environment, holdout testing is the only way to validate what platform numbers are telling you.
Total revenue from your payment processor divided by total paid media spend for April and May. If this number is deteriorating while platform ROAS is holding flat or improving, you have a measurement problem, not a performance problem. If both are deteriorating, you have a performance problem. These are different diagnoses with different solutions.
One trend that didn't make headlines in May but is quietly creating an attribution nightmare: brands in regulated categories like health and finance should manually review any AI-generated ad copy before publishing, as automated systems may produce language that conflicts with compliance requirements.
But the issue goes beyond compliance. When Google's AI Max writes your Shopping ad copy from your Merchant Center feed, and Meta's AI generates your Advantage+ creative variants, and TikTok's Symphony tool creates your video assets — the creative that's actually running may be materially different from what you approved. Attribution tells you which ads are driving conversions. But if you don't know exactly what creative ran, you can't replicate what worked or diagnose what didn't.
This is a problem that compounds over time and gets more expensive the longer it goes unaddressed.
Every major platform changed how it measures performance in May. All of those changes create transition periods where historical comparisons break. DTC customer acquisition costs are at a multi-year high. AI-powered campaign types are automating spend decisions faster than most brands can audit them. The brands that will come out of Q2 in a strong position are the ones with a verified new customer acquisition cost that doesn't come from any platform's dashboard — one that's anchored in their own CRM data, validated against their blended MER, and stress-tested through holdout experiments. That number doesn't exist inside Meta Ads Manager or Google Ads. It has to be built independently.
That's what Wicked Reports is built for.