-
Nielsen will acquire DoubleVerify for $2.15 billion, combining audience measurement and media verification in a business serving more than $300 billion in advertising spend.
-
The deal raises concerns about independent media verification, as the company measuring inventory quality will be owned by a major audience ratings provider.
-
Independent studies suggest current ad fraud controls miss significant invalid traffic, including bot-delivered ads and made-for-advertising inventory that reduce media effectiveness.
-
Advertisers often receive verification scores without impression-level transparency, making it difficult to audit or challenge measurement results.
-
The article urges brands to strengthen accountability by using independent post-bid verification, updating vendor contracts, demanding ownership disclosures, and focusing on human or effective CPM metrics. *
On 6 August, Nielsen agreed to acquire DoubleVerify for $13.60 per share in cash, valuing the business at roughly $2.15 billion, a 30% premium to DoubleVerify’s 60-day average price.
The company expects more than $4 billion in annual revenue and says it will serve customers responsible for over $300 billion in advertising spend. Nielsen’s chief executive described the result as “a truly independent, end-to-end partner”, and DoubleVerify’s leadership promised “a single currency that scores media on both audience delivery and media environment quality”. (Nielsen, 6 August 2026)
So, the company that grades the quality of media inventory will now be owned by the company that sells audience ratings to the people who sell that inventory. They call this independence. Advertisers should at least be curious about the definition of “independent”.
Independent auditing in other areas of business is standard practice, or else we end up with situations like the 2001 Enron scandal. So why not here?
Who pays the inspector?
Verification is not a public utility. It is a product, and almost all of it is sold on a fee-per-thousand-impressions basis. That pricing is unremarkable until you notice what it rewards.
A vendor whose revenue rises with the number of impressions flowing through the pipes has no commercial incentive to argue that the pipes should carry fewer.
Imagine a food hygiene inspector paid by the number of meals served, rather than by the number of dirty kitchens closed. You would not accuse that inspector of dishonesty. You would simply stop treating a clean rating as proof of a clean kitchen.
This is a structural conflict, not deliberate vendor malfeasance. They are part of a supply chain that is opaque by design, so all it takes is an incentive that slightly tilts away from the advertiser’s interest, applied across trillions of impressions.
The evidence that something is being missed
There is ample proof that the current model is wanting: independent researchers have tested it. In 2024, Adalytics analysed more than a petabyte of web traffic and reported that ads for Fortune 500 brands, plus US government bodies including the Navy, the Army, the CDC, the Department of Veterans Affairs and the Postal Service, had been delivered to declared bots running in data centres, despite those advertisers paying for pre-bid bot filtration.
The same research reported that Integral Ad Science’s publisher services pixel classified declared bots as valid human traffic 17 per cent of the time, and non-declared bots 77% of the time (Adalytics).*
*[Note: The named vendors disputed the findings with legal proceedings underway.]
What is hard to dispute is the shape of the exchange: an outside party with no revenue at stake produced the uncomfortable numbers, and the parties paid to produce them did not.
That pattern is not new. The Programmatic Media Supply Chain Transparency Study by the ANA (Association of National Advertisers) put made-for-advertising inventory at around a fifth of programmatic impressions.
They found that non-viewable and invalid traffic impressions, as well as made-for-advertising ad spend, account for a 35% loss in media productivity, costing the industry some $120bn annually… and rising.
These findings came from independent research, not from the measurement layer that was supposed to be watching.
In our experience of dealing with multiple advertisers, we have found even these figures to be understated.
Who is watching the watchers?
The industry’s answer to that question is accreditation: the Media Rating Council audits and accredits measurement products, one product at a time, submitted by the vendor that built it.
Meanwhile, the advertiser, the party carrying the loss, usually receives a score rather than the evidence behind it. You get the grade without the marking – it’s what we call a ‘black box’.
When a discrepancy appears, the advertiser is in the awkward position of challenging a number they cannot inspect, produced by a vendor their agency selected, under a contract that rarely says what happens if it is wrong.
Consolidation sharpens that problem. A single currency helps with planning, but it is also a single point of failure, and it shortens the list of organisations that must be persuaded before an inconvenient methodology quietly changes.
Fewer scorers means fewer second opinions, and the timing is awkward: since September 2025, the UK’s Economic Crime and Corporate Transparency Act has exposed large organisations to criminal liability where an associated person (agencies and ad tech intermediaries included) commits fraud on their behalf without reasonable prevention procedures in place.
What accountability would look like
None of this requires ripping out your stack. It requires four things advertisers can ask for this quarter:
- Focus on post-bid, impression-level data, as this describes ‘what happened’ as opposed to ‘pre-bid,’ which described what was intended to happen: As log-level data is populated with, inter alia, pre-bid data which can be ‘gamed’, it is vital to have a 3rd party, independent post-bid data source.
- Include discrepancy and remediation terms in vendor contracts, so being wrong carries a price: Evolve contracts to paying on provable outcomes and not volumes purchased
- Ask every measurement partner to disclose, in writing, its ownership and any revenue relationship with the sellers of the inventory it grades.
- Train brand teams to think in terms of ‘eCPM/hCPM (effective or human CPMs) and not on least cost/max volume
That last one is the category my company works in, so take it as an interested opinion: a scorer with different incentives is the check most likely to surface what the first one passed over.
The same test applies to us. TruthsetsOnline sells forensic measurement, so we benefit if you conclude that your current verification is missing things. That is why we do not ask you to trust us; we provide a ‘glass box’ approach.
We ask you to run our tags next to your existing tools and compare the outputs, at impression level, on your own campaigns. If the numbers agree, you have lost nothing but a fortnight. If they do not, you have found the gap that nobody in the current arrangement is paid to find.
If you want a free forensic audit of your campaigns, InMail me, and we will compare the evidence side by side.
* Summary created by AI
Marc Dhalluin is the founder of TruthsetsOnline, a digital transparency and ad fraud analytics platform. Marc has built, transformed, and rescued brands generating over $500 million in combined annual revenue. A longtime collaborator of fraud researcher Dr Augustine Fou, he now focuses on one thing: proving that genuine advertising reaches real people. Based in Los Angeles, Marc writes regularly on digital transparency, programmatic fraud, and what brands should actually be measuring.
Connect with Marc on LinkedIn or visit TruthsetsOnline.com
