Let me ask you something.

Name a vendor that is accredited for brand safety. Just one.

You said IAS. Or DoubleVerify. Or maybe Moat, if you have been out of the industry for two years and forgot that Oracle completely killed Moat on September 30, 2024 and stranded every advertiser still using it for measurement.

The point is the same regardless of which name you picked.

WRONG. You did not pick a vendor that is actually accredited for brand safety.

Because none of them are. NOT A SINGLE ONE.

Read that again. Question everything you know.

The MRC, the body that issues these accreditations, accredits very specific metrics.

Sophisticated Invalid Traffic filtration for connected television. CTV Video Viewable Impressions. TikTok Video Viewability. Those are real accreditations for real metrics.

They are also approximately none of what your CFO thinks she is buying when she signs the verification contract.

Brand safety classification. Contextual analysis. Made-for-advertising detection. Attention measurement. The category names that show up on every IAS and DoubleVerify deck on every Wednesday morning sales call in every Fortune 500 procurement department in America. None of those are MRC-accredited. The accreditation seal you have seen in fifteen vendor pitches over the last three years does not cover the product you are actually buying.

I will get to why that matters in approximately six paragraphs.

First, the bigger problem.

In January 2026, a peer-reviewed study examined 4,352 news articles across 51 domains and asked a simple question. Do the major verification vendors classify the same content the same way?

The answer was no. Not approximately no. Not directionally no. Just no. The researchers documented systematic disagreement between DoubleVerify, IAS, and Oracle Moat on the brand safety classification of identical content. An article one vendor called "unsafe" another vendor called "safe." A category one vendor flagged as "sensitive" another vendor ignored. The classifications were inconsistent at a level that the researchers concluded had "harmful consequences for both advertisers and publishers, leading to misplaced advertising spending and revenue losses."

This is not a controversial finding inside the industry. Anyone who has ever requested an appeal on a brand safety flag from one vendor while running a parallel campaign with another vendor knew this already. What is controversial is that the finding has now been published, peer-reviewed, and made available to procurement teams who can cite it in vendor reviews. The plausible deniability window has closed. The vendor's standard "our methodology is proprietary" response is now a response to a peer-reviewed study showing the proprietary methodologies produce different answers to the same question.

That is the procurement problem in one sentence. You are paying premium fees to vendors who cannot agree with each other on what they are measuring.

The industry response has been to act like this is fine. Like having three weather vanes pointing in three different directions during the same storm is somehow more accurate than having one. The expert framing in this category is that you should buy multiple verification vendors to "triangulate" the measurement. Triangulation requires three points that agree. Disagreement requires a referee. The verification vendors are the referee. Who is the referee's referee?

Nobody. That is the answer to that question. Nobody.

Here is the part that matters more than the accreditation question, and it is the part the trade press will not write because the trade press depends on the verification vendors for sponsorship dollars and panel fees.

IAS and DoubleVerify are not measurement vendors anymore. They are market governance infrastructure.

When DoubleVerify classifies a domain as "made for advertising," CPMs on that domain collapse. When IAS publishes a brand safety report on a publisher, agencies pause spend. When the verification vendors update their methodology, the entire programmatic supply chain rearranges itself to fit the new methodology, because the alternative is being classified as the kind of supply nobody wants to buy.

That is not what measurement vendors do. That is what regulators do.

Except verification vendors do not have regulators' accountability mechanisms. There is no public comment period before IAS changes its MFA definition. There is no due process when DoubleVerify flags a publisher's domain. There is no appeal beyond writing an email to a sales rep who has no authority to overturn the methodology. When a publisher loses revenue because a verification vendor's contextual analysis misread a news article about terrorism as terrorism-supporting content, the publisher's recourse is approximately zero.

The Senate noticed this. On March 27, 2025, Senator Mark Warner wrote to the FTC chairman and the Attorney General specifically naming IAS, DoubleVerify, and HUMAN Security. Warner asked the FTC to investigate whether these vendors had made false advertising claims about pre-bid bot avoidance. He asked the DOJ to investigate whether they had violated the False Claims Act by billing the federal government for services not delivered. The specific technical issue Warner raised was whether the pre-bid verification technology actually receives the User Agent field from DSPs like Google DV360 and The Trade Desk in real-time programmatic auctions. If they don't, they cannot identify declared bots pre-bid. Their marketing materials claim that capability anyway.

The trade press covered the Warner letter as a Senator-asks-FTC-to-investigate story. They did not cover it as a sitting U.S. Senator pointing out that the entire pre-bid bot avoidance category is potentially built on a technical impossibility. Those are different stories. Only one of them got covered.

On May 22, 2025, Electrical Workers Pension Fund, Local 103, I.B.E.W. filed a securities class action against DoubleVerify in the Southern District of New York. Case number 1:25-cv-04332. The complaint runs through the alleged misconduct in significant detail, but two paragraphs in particular should be sitting on the desk of every CMO who has ever signed a DoubleVerify contract.

Paragraph one. The complaint alleges that DoubleVerify systematically overbilled its customers for ad impressions served to declared bots. Not undeclared bots. Declared bots. Operating out of known data center server farms. The Wall Street Journal corroborated the allegation independently. The Journal documented tens of millions of instances over seven years where ads from brands including Hershey's, Tyson Foods, and T-Mobile were served to bots across thousands of websites. The verification vendor whose entire product category is built on bot detection allegedly missed declared bots running out of named data centers for seven years.

Paragraph two. The complaint also alleges that DoubleVerify's risk disclosures to investors were materially false and misleading because they characterized adverse facts that had already materialized as mere possibilities. Specifically, the complaint alleges that DV failed to disclose to the market that its customers were shifting ad spending from open exchanges to closed platforms where DV's technological capabilities were limited and competed directly with native tools from Meta and Amazon. DV's Activation Services in connection with closed platforms would take several years to monetize. DV's competitors were better positioned to incorporate AI into closed-platform offerings.

Translate that into procurement language. Every DV contract that included Activation Services for Meta, Amazon, or other closed platforms was priced against a capability claim that DV's own investors now allege was never deliverable on the represented timeline. The advertiser was paying for a product the 10-K eventually acknowledged to investors was years away from functioning as marketed.

The DV stock fell from a peak above $39 to $13.90 across the alleged class period. A decline of more than 70 percent. The market knew. The market priced the disclosure failures into the equity. The advertisers who were paying the bills did not.

This is the structural fact that should determine how every advertiser thinks about verification spending in 2026, and it is the fact that has never been disclosed to advertisers in their contracts.

IAS and DoubleVerify collect revenue from both sides of the same transaction they purport to referee independently.

DoubleVerify segments its revenue into three streams. Activation. Measurement. And Supply-Side, which is revenue from the platforms and publishers being measured. In Q1 2025, DV's supply-side revenue grew 35 percent year-over-year to $16.5 million. In Q4 2024, it grew 34 percent. DoubleVerify extends its data solutions to supply-side platforms including Criteo's Commerce Grid and Index Exchange. The 10-K describes supply-side revenue as generated from platforms and publisher partners who use DoubleVerify's data analytics to evaluate, verify, and measure their inventory.

IAS does the same thing. The publisher revenue segment accounted for approximately $76.5 million of total revenue at last disclosure. IAS acquired Publica, a CTV ad platform, for $220 million in 2021, explicitly acquiring SSP-adjacent technology that "helps increase publisher yield." The vendor verifying whether your ads ran adjacent to safe content is also selling tools to the publishers that produce the content being verified.

The IAB Australia's own purchase guidelines ask buyers directly. "Is your verification technology independent from any potential conflicts of interest such as the selling of media?" Neither IAS nor DoubleVerify is independent under that definition. Neither surfaces this prominently in advertiser contracts. A CMO signing a verification contract has a reasonable expectation that they are buying an independent auditor. They are not.

Check My Ads Institute named the structural problem in their policy platform. They called for "codifying a best interest duty for all adtech intermediaries and verification vendors who work on behalf of advertiser and publisher clients." The Best Interest Duty language is the language financial services regulators use for registered investment advisers. The fact that a policy institute has had to propose extending fiduciary duty concepts to verification vendors tells you exactly where the structural reform conversation is heading.

Here is where I am supposed to do the polite editorial thing and remind you that none of these allegations have been adjudicated. The class action is pending. The Senate inquiry is ongoing. The Warner letters are not findings. The polite editorial thing is not really the ADOTAT thing. All of those statements are true. None of them changes what your contract says. None of them changes what your CFO has been signing for the last five years without reading the 10-K of the vendor cashing the checks.

A Prohaska Consulting and Advertiser Perceptions study with 110 US advertisers, publishers, and adtech professionals, quantified the publisher-side cost of the verification methodology nobody has been willing to audit.

US publishers estimated that 30 percent of their advertising inventory went underfunded or undersold in 2023 due to brand safety blocking. Ninety-seven percent of publishers said their business had been negatively impacted by category-level news blocking. One in three agencies or marketing professionals had been given a directive not to spend ad dollars in news for a period of time.

A quarter or more of US advertisers were blocking ad placements next to news stories containing the words "election," "Hamas," "sex," or "Israel." Read that sentence twice. Major Fortune 500 brands paying premium fees to IAS or DoubleVerify to filter out the word "election" during a presidential election year. The verification industry built the blunt instruments. The verification industry sold the blunt instruments. The verification industry continues to ship the blunt instruments and treat the resulting publisher revenue collapse as someone else's problem.

The most damaging number is the one buried at the bottom of the Prohaska report. The industry estimated 40 percent of programmatic ad dollars flowed to made-for-advertising sites in 2023. Forty percent. The verification vendors are paid specifically to classify and reduce MFA exposure. The classification is happening. The forty percent is still flowing. Either the classification is not working or the verification industry has a structural incentive to keep the MFA category large enough to justify the next product line. Either reading is procurement-relevant. Neither reading is in any IAS or DoubleVerify investor communication.

The verification vendors are not measurement infrastructure. They are an arbitrage layer that profits from the dysfunction they were hired to reduce. The Prohaska data is the documentary support. The advertisers paying the bills are the funders of the arbitrage.

Let me come back to where I started.

Name an accredited brand safety vendor.

If you said IAS, you named a vendor that has MRC accreditation for Sophisticated Invalid Traffic filtration for CTV. That is real. That is also not brand safety classification. The brand safety category your contract says you are buying is not what the vendor is accredited for.

If you said DoubleVerify, you named a vendor that has MRC accreditation for CTV Video Viewable Impressions, and in April 2026 was the first measurement provider to receive MRC accreditation for TikTok Video Viewability. Both are real. Both are also not brand safety classification. The accreditation seal on the slide deck does not cover the product line you are paying for.

If you said Moat, congratulations. Moat is dead. Oracle shut the entire business down on September 30, 2024. The historical data export window has closed. Any advertiser whose attribution model depended on Moat continuity has been operating without that continuity for over a year and most of them did not notice until their year-end review.

The accreditation seal you have been treating as a baseline credential for the last decade does not cover what you think it covers. The MRC audit reports that underlie the accreditation are confidential between the MRC, the auditor, and the vendor. You cannot read them. You cannot validate the methodology. You cannot appeal the classification when it goes against you.

You can pay the invoice though. The invoice you can pay.

The Media Rating Council wanted to clarify a few important points about how their accreditation process works.

MRC audit reports are kept confidential by design. This protection exists so that companies applying for accreditation can provide full transparency into their operations without worrying that proprietary information will be exposed publicly. Without that safeguard, it's unlikely any measurement provider would willingly open up their systems to the level of scrutiny the process requires. That said, MRC members do have access to the audit reports, provided they agree to confidentiality terms.

What MRC accreditation actually signifies is straightforward: the measurement provider is in compliance with the published, relevant MRC standards. In other words, accreditation confirms that the provider meets industry-developed standards that are themselves publicly available.

The MRC also wanted to address a common misconception about the financials. The fee for an audit does not go to the MRC. It goes to the auditing company that performs the work. The MRC makes no money off audits whatsoever.

Part Two is the procurement memo Publicis paid FirmDecisions to write and your company didn't. Fifteen questions reformatted as the questions your account team should be answering, with the red-flag response language that should escalate the conversation to your CFO or general counsel within twenty-four hours. Eight contract clauses drafted to drop into your 2026 IAS or DoubleVerify renewal, each paired with the resistance pattern to expect from the vendor's contract team, calibrated to the public record on supply-side revenue, pre-bid technical capability, MRC accreditation scope, and the Senator Warner inquiry. The single sentence to add to your MSA that closes the Big Four auditor defense before your verification vendor deploys the same position The Trade Desk took against FirmDecisions in March.

What you are missing without it. The Russian-nesting-doll section explaining why your verification spend has been growing faster than your media spend for three years without measurable media quality improvement. The four-stage deployment plan for using the memo in negotiation without triggering a defensive posture from the account team. The "what your account team will actually say" forecast that lets you walk into the next QBR knowing in advance which clauses the vendor will accept, which they will soft-pedal, and which they will refuse outright. And the one-page structural file with the side-by-side scorecard for DoubleVerify and IAS, the litigation status, the regulatory exposure, and the Moat orphan attribution problem your CFO is going to ask about the day after your board reviews 2026 KPIs. Subscribe to read it before Wednesday's renewal call.