Blog · Supply chain
A publisher clean on your exchange IVT log can still carry a clawback warning on another seller path.
On October 6, 2026, Pixalate introduced a Clawback Warning label in its Media Ratings Terminal. The label marks CTV apps, mobile apps, and websites with invalid traffic disputes reported against them, drawing on transactions monitored across 745 global ad exchanges. Pixalate evaluates more than 17.5 million apps and sites and refreshes the warning monthly over a rolling 12-month window. For SSP inventory quality teams, that is a new onboarding signal. It is not a substitute for your own impression IVT rate on the same publisher this week.
What clawbacks measure versus what IVT measures
Impression-level invalid traffic classification asks whether a specific served event looked invalid under a vendor model at bid or render time. A clawback, in this product framing, is a buyer dispute record: inventory that led to credits or makegoods after delivery. Pixalate states the warning is meant to prompt research before onboarding, not to act as a verdict, and that MRT shows how many months were affected, not who filed each clawback.
Those are related problems with different denominators. A publisher can serve impressions on your exchange that pass your filters while the same bundle generates disputes on another path you never see in your logs. Your dashboard shows a low post-bid IVT share. MRT can still show a Clawback Warning with a months-affected count such as four of twelve, refreshed monthly.
Pixalate also separates the warning from impression-level IVT measurement in the announcement copy. Treating the label as a restatement of your quarterly IVT benchmark misreads the object. It is dispute history aggregated across exchanges, not a replay of your auction telemetry.
Waterfall inventory splits quality by seller path
The October 6 post names waterfall detection explicitly. Publishers can sell higher-quality traffic through some SSPs and lower-quality traffic through others, especially in international markets. Each exchange sees only its own disputes and only its own impression stream. Low IVT on your path therefore does not automatically make a publisher safe to onboard globally.
OpenRTB still identifies inventory with app.bundle, site domain, and publisher identifiers. It does not carry which sibling SSP received the bad slice last month. Screening tools that only query your internal IVT table will miss cross-path clawbacks until a buyer escalates outside your ticket system.
The same release compares the label to an ecommerce high-return badge: a reason to read reviews before purchase, not a ban. For supply teams, the operational read is due diligence on apps and sites already in market, not only at signup. Continuous monitoring each month for new clawbacks reported by buyers elsewhere is part of the stated use case.
How the label is built
Inputs include apps and websites named in clawbacks filed through Pixalate Clawback Tool plus third-party clawback submissions reviewed against evidence. The lookback is a rolling 12 months, matching the Clawback Tool window, with monthly refresh. Output is the Clawback Warning on the app or site MRT page with month-by-month records and a months-affected count.
Pixalate applies the label automatically and removes it after 12 months without a reported clawback. No party can add or remove it manually. Coverage spans Apple App Store, Google Play, Roku, Amazon Fire TV, Samsung TV, Apple TV, and LG TV alongside web properties, including ads.txt-verified publishers named in the release.
System metrics published with the launch include 17.5 million plus evaluated properties, 745 global ad platforms in the monitoring set, monthly updates, and the 12-month lookback. Those figures describe the rating universe, not your exchange share of it.
What SSP teams can do on the bidstream side
Clawback warnings live in a ratings terminal, not in OpenRTB. Your ingress pipeline still controls whether inventory enters the market at all. Pair external labels with bundle and domain checks you can enforce: app-ads.txt authorization, schain completeness, and consistent app.bundle typing before the request reaches buyers.
Structural risks such as delisted apps that keep bidding are a adjacent failure mode we covered separately on RTBlint. Impression IVT does not catch delisting; app-level exclusion does. Clawback warnings are another app-level or site-level signal that impression filters can miss when the events on your path look normal.
- Onboarding: research MRT clawback months before approving new publishers, especially when they already sell through multiple SSPs.
- Waterfall review: compare dispute labels with sell-path maps when international traffic spikes without matching quality on your exchange.
- Do not merge dashboards: keep clawback research separate from impression IVT rates so a falling IVT chart does not hide rising dispute months.
- Validate requests: paste production CTV and mobile requests into the bid request tester so policy runs on the same shape buyers receive.
The honest limit
A vendor-maintained clawback label reflects disputes filed through specific tools and reviewed third-party submissions. It will not list every buyer credit in the market, and it is not a pre-bid field you can require in OpenRTB today. It closes a due diligence gap: publishers that look fine on your path while accumulating disputes elsewhere.
RTBlint checks bid request shape and consistency. It does not call MRT or classify invalid traffic. A clean RTBlint result is not a clean clawback history. It is proof the request you are about to sell is structurally readable.
Sources
Field names refer to OpenRTB as published by IAB Tech Lab. Clawback and system metrics quotes are from Pixalate October 6, 2026 blog post. RTBlint is independent and not affiliated with IAB, IAB Tech Lab, or Pixalate.