Pixalate today added a Clawback Warning label to its Media Ratings Terminal, marking connected TV apps, mobile apps and websites against which buyers have reported invalid traffic disputes in a rolling 12-month window, according to the company.

In Short

Pixalate, a company that measures ad fraud, today began showing a warning on apps and websites against which buyers have filed disputes over the quality of ad traffic, much like a "frequently returned" tag in an online shop. It matters most to the exchanges that sell this inventory, because each one normally sees only the disputes filed with itself, so a seller can look clean in one place while drawing complaints in another. The label shows how many of the last 12 months included a reported dispute and disappears after 12 months without one, so a marked seller can later look clean again.

What the label does

Pixalate placed the warning on the individual pages that its Media Ratings Terminal (MRT) keeps for CTV apps, mobile apps and websites. According to Pixalate, the terminal evaluates more than 17.5 million such properties, and the label is refreshed monthly over a rolling 12-month window, so that it "reflects ongoing activity rather than a one-off report." It sits beside the risk ratings and labels the company already publishes for each app and site.

The intended readers work on the sell side. According to Pixalate, the label gives supply-side platforms (SSPs) and exchanges "a new data point to research before onboarding a publisher." The company's own comparison is to Amazon's "Frequently Returned Item" label, which asks shoppers to check product details and reviews before buying; the Clawback Warning, it says, points to additional risk factors in the same spirit.

What earns the label is a record of disputes over invalid traffic (IVT) reported against an app or site. Pixalate describes the population as properties, ads.txt-verified publishers among them, that "exceed ad fraud clawback thresholds observed across multiple seller paths." The thresholds themselves are not given. Nor does the release define a clawback. In programmatic trading the word generally describes a buyer's attempt to recover spending on impressions later judged invalid, and the release does not say whether a report must end in a granted refund before it counts.

How the record is built

The release lays out the mechanism in three parts. The input is the set of apps and websites named in clawbacks filed through Pixalate's Clawback Tool, plus third-party submissions that are, according to Pixalate, "reviewed against supporting evidence before inclusion." The lookback is a rolling 12-month window that matches the one used by the Clawback Tool, refreshed monthly. The output is the label on the app's or website's MRT page, with a month-by-month record of clawbacks and a count of months affected, for which the release gives 4 of 12 as an example.

Two rules govern the life of a label. Pixalate applies it automatically and removes it once an app or website goes 12 months without a reported clawback, and according to the company no party can add or remove it manually. The terminal displays how many months were affected, but not who filed each clawback, and the warning is reported separately from impression-level IVT measurement.

System metrics

ParameterData, according to Pixalate
CTV apps, mobile apps and websites evaluated17.5 million+
Global ad platforms covered745
App stores coveredApple App Store, Google Play Store, Roku, Amazon Fire TV, Samsung TV, Apple TV, LG TV
Lookback windowRolling 12 months
Update frequencyMonthly
Data sourcesPixalate Clawback Tool; third-party clawback submissions

The label is live today in the terminal at ratings.pixalate.com. The Clawback Tool, which feeds it, is available to Pixalate Analytics clients, according to the company.

Why one exchange's view falls short

The premise rests on partial visibility. According to Pixalate, low IVT on a single exchange does not automatically make a publisher safe to work with. A publisher can "waterfall" its inventory, selling higher-quality traffic through some SSPs and lower-quality traffic through others, especially in international markets, while each exchange sees only its own disputes. The usage differs from the sequential ordering of demand sources that the word describes in mobile ad mediation; here it means traffic sorted by quality across sellers.

Pixalate lists four ways SSPs and exchanges can use the label:

  • Waterfall detection - spotting publishers that may sell better traffic through some SSPs and worse traffic through others.
  • Cross-ad exchange screening - viewing clawbacks reported across other supply paths.
  • Onboarding due diligence - researching a new publisher's apps and websites before the inventory enters the marketplace.
  • Continuous monitoring - checking supply already trading each month for new clawbacks reported by buyers.

Is a monthly, 12-month record enough to catch a seller who moves its weakest traffic between exchanges? The release does not test that question, and it publishes no count of labelled properties that would allow an outsider to judge how often the pattern appears.

A worked example from the release

The release carries a screenshot of an MRT page for a mobile game whose name is blurred. According to the banner on that page, a clawback had been reported for the app 9 times in the last 12 months. A separate panel records the months in which the app appeared in a Clawback Report across a window running from October 2025 to September 2026: one report each in February, April, May, June and September 2026, and four in August 2026. That makes 6 of 12 months affected, with September 2026 shown as the most recent.

Two details stand out. The count of reports (9) and the count of months affected (6) measure different things, so a single bad month with several filings, as in August, weighs differently in each. And the release text cites 4 of 12 as its example of the months-affected figure while the screenshot shows 6 of 12; the two are separate illustrations. A footnote beneath the panel states that the numbers above each month count Clawback Reports in that month, that the reporting party is not shown, and that the counts reflect reports generated by Pixalate clients and externally submitted reports. The app's page also lists a developer country of Cyprus and downloads of 100M to 500M.

What the release leaves open

The footnote matters for interpretation. If the counts draw on reports from Pixalate clients and external submitters, coverage depends on who files, and a property trading mainly with buyers outside that group could carry fewer reports than its traffic would warrant. The release gives no figure for how many clawback reports feed the terminal each month, how many properties are currently labelled, or how submissions from third parties are verified beyond the phrase about supporting evidence. It also offers no route for a publisher to contest a label, other than the passage of 12 clean months.

There is a measurement-governance question as well. Pixalate states that it is accredited by the Media Rating Council(MRC) for the detection and filtration of sophisticated invalid traffic, and the release reports the new warning separately from impression-level IVT measurement. It does not say whether the label falls within the scope of that accreditation.

The company's disclaimer is explicit about status. According to Pixalate, its opinions are "neither facts nor guarantees," the data is grounded in its proprietary technology and analytics, and the information is shared "not to impugn" any entity, person or app. It also cites the MRC in noting that "fraud" in this context is a custom definition for advertising measurement, not a legal one. In short, the label is a vendor's signal, offered as a prompt to "look closer rather than a verdict," in the company's words.

Context for the marketing community

Recovering money spent on invalid traffic is not a new idea. In September 2017, Google said that AppNexus, Index Exchange, OpenX, PubMatic, SpotX, Teads, Telaria and DoubleClick Ad Exchange had committed to refund invalid traffic detected within 30 days after monthly billing, commitments it put at 90% of programmatic buying. The 30-day figure belonged to that commitment, and the release says nothing about current refund terms. What the new label adds is memory: a 12-month record attached to the inventory rather than to a single invoice.

Supply-chain data has its own record of unreliability. A February 2024 Pixalate analysis found that 25% of programmatic traffic carrying SupplyChain object data failed validation, and failing traffic showed invalid traffic rates 64% higher. That figure comes from the same vendor now publishing the label, a point that applies to both datasets.

Buyers already have a blunt instrument on their side of the market. In May 2025, DV360 added a seller ID blocklist that lets advertisers block specific exchange and seller IDs, at an additional 1.5% charge on media costs. A sell-side label and a buy-side block list address related problems from opposite ends of the transaction; the release does not say whether Pixalate's label data can feed such controls.

The note in the release that ads.txt-verified publishers can be labelled also fits recent data on how authorisation files are used. DataBeat's September 2026 Sellers Report, published on September 18, found that publishers among the top 50,000 US websites added a net 54,942 ads.txt lines, 44,608 of them through reseller relationships against 10,334 direct. An ads.txt entry records who may sell a publisher's inventory; it says nothing about the quality of the traffic that reaches that inventory.

Pixalate itself has been widening its supply-side data offer. On June 17, 2026, it published the OpenEPG Index 1.0, and on September 21, 2026, it released OpenEPG DB 1.0 and a Pre-Bid API in private beta for DSPs, SSPs and exchanges in the United States. Its accreditation history is longer: the company has held MRC accreditation for sophisticated invalid traffic detection and filtration in OTT since November 2018. The Clawback Warning is a different kind of product from all of these, built from buyer reports rather than from impression-level detection, and whether SSPs adopt it in onboarding is not addressed in the release.

Timeline

Summary

  • Who: Pixalate, a global ad fraud and privacy compliance platform founded in 2012, with SSPs and exchanges as the intended users of the label.
  • What: A Clawback Warning label in the Media Ratings Terminal that marks CTV apps, mobile apps and websites with invalid traffic disputes reported against them, drawn from Pixalate's Clawback Tool and reviewed third-party submissions, refreshed monthly over a rolling 12-month window.
  • When: Today, with the label live on release; the window covers the preceding 12 months and a label lapses after 12 months without a reported clawback.
  • Where: On CTV app, mobile app and website pages in the Media Ratings Terminal at ratings.pixalate.com, covering 17.5 million+ properties across seven app stores and 745 global ad platforms; the announcement is datelined London.
  • Why: According to Pixalate, low IVT on one exchange does not make a publisher safe, because publishers can route better traffic to some SSPs and weaker traffic to others while each exchange sees only its own disputes; the label gives sell-side teams a cross-exchange data point before onboarding.