Lunio, a company that sells invalid-traffic detection and filtering tools to advertisers, reported that retail search campaigns running Google’s AI Max were exposed to 72 percent more invalid traffic than standard search campaigns over the same window. The finding lands 19 days before Google begins converting two legacy campaign settings to AI Max with no opt-in required. Lunio has a commercial interest in that outcome: a bigger invalid-traffic problem is a bigger market for the tool it sells.
PPC Land, which first reported the study on August 12, 2026, described a sample exceeding 414 million clicks. Those were logged between October 2025 and June 2026 on Google Ads, Bing, LinkedIn and Meta plus assorted native and social platforms, then narrowed to retail brand accounts. Within that window, the invalid-traffic rate on AI Max retail search campaigns climbed from 2.46 percent in Q4 2025 to 5.28 percent by Q2 2026. Standard search campaigns in the same accounts moved the other way, easing from 3.72 percent to 3.07 percent. One campaign type worsened. The other, running on the same platform against the same advertisers, did not.
That 72 percent figure is a rate comparison confined to Google search: AI Max clicks measured against standard search clicks. A separate figure in the report states that AI Max accounted for 68 percent of all invalid clicks Lunio detected across the full Google search dataset. Those two things measure different quantities, one a proportion of a total and the other a frequency, and the report keeps them apart. Any campaign type big enough will naturally account for much of whatever it is measured against, so volume alone could explain a large share. What makes that explanation harder to sustain here is that the rate diverges too, not just the total.
Every figure in this report comes from Lunio’s own classification of what counts as invalid: coordinated bot activity, scraping, competitor click manufacturing, and accidental clicks, sorted by Lunio’s own detection system. None of it has been independently replicated, and the company’s revenue depends on convincing advertisers that the problem it measures is large. That does not make the numbers wrong. It does mean the rate should be read as one vendor’s methodology, not an audited industry benchmark.
Retail invalid traffic averaged 5 percent across the full dataset, rising from 4.18 percent to 5.54 percent over the three quarters measured. Google Shopping carried the highest rate of any campaign type, climbing from 4.16 percent to 7.51 percent, an 80 percent increase. Off Google, Meta averaged 5.99 percent and TikTok 5.56 percent, with TikTok’s rate up 68 percent across the period.
Lunio does not publish a market-wide loss estimate. It instead works a single hypothetical case: a retailer spending $10 million a year at a $3.70 average cost per click. At the 5 percent overall rate, that arithmetic produces roughly $500,000 in wasted spend, and at a conservative three-to-one return-on-ad-spend ratio, about $1.25 million in lost revenue opportunity. The example is a worked illustration, not a measured result. It scales linearly with an advertiser’s own budget and assumes blocked clicks would have converted at the account average, a modeling convention rather than an observed outcome.
Rather than pin it on one platform decision, Lunio chief executive Nick Morley tied the sector’s exposure to how retail trading is structured across the year. Automated media buying and short, compressed sales-event budgets, he argued, leave retail advertisers more exposed to invalid traffic than accounts in other verticals.
This is the third data point this publication has tracked on the same underlying mechanism. Reporting here on August 7 covered Google’s plan to auto-upgrade eligible Search campaigns to AI Max from September 1 with no opt-in, and reporting on August 10 covered a Smec study finding that algorithmic broadening already absorbs a large share of exact match impressions. Lunio’s contribution is a claim about traffic quality rather than traffic share: broader matching reaches more inventory, and Lunio argues more of that inventory fails its own validity check. The September 1 default means the question stops being one that only opted-in advertisers have to answer.
Retailers whose accounts convert automatically on September 1 will enter fourth-quarter budget season on a setting Lunio’s data ties to nearly double the invalid-traffic rate of the one it replaces, measured on data that ends in June. Search and paid-media teams managing retail accounts should pull their own invalid-click and conversion figures by campaign type before that date, rather than treat either Google’s performance claims or Lunio’s vendor figures as a substitute for account-level evidence.
Per a PPC Land report by Luis Rijo, published August 12, 2026, citing Lunio’s retail invalid traffic study.
