Part 3 of an ongoing series. If you haven't read Parts I and II, go do that. I'll wait. I'm a rabbi. Patience is in the job description.

A few weeks ago I started writing about the TV measurement industry because I thought it would be a nice little series about some companies that measure television.

Three parts and several hundred pages of internal documents later, I've discovered that a $90 billion industry is being run like a homeowners association where nobody reads the bylaws, everyone hates the board president, and the plumbing hasn't been updated since 1987.

It gets stranger every week. I am not making any of this up. I couldn't. I'm a good writer but I'm not that creative.

Since Parts I and II, I've spoken to board members, C-suite executives, institutional investors, agency heads, network leaders, securities lawyers, communications professionals who are bad at communicating, and one person who I'm fairly sure was calling me from a closet so their colleagues wouldn't overhear. The picture isn't a scandal. It's worse. It's a system so deeply, boringly broken that fixing it would require effort, and effort is the one thing this industry has collectively decided it will never expend. Not even a little. Not even if you show them a chart.

I don't mean they're badly run. I mean they don't behave like organizations that exist in a competitive market where reputation matters. They behave like medieval fiefdoms. Suspicious of outsiders, terrified of transparency, governed by the emotional weather patterns of whoever's in charge that quarter. Which changes more often than my kids change their minds about dinner.

VideoAmp had a CEO for six months. Then he was gone. Then the CMO was gone the next day. Their own board members talk to me with documentation, but the people running the company are hiding behind an external PR firm that has perfected the art of not responding to emails. How did I find out about the leadership change? My interview got canceled. You don't cancel press interviews during a "smooth, planned transition." You cancel them because nobody knows who's authorized to talk. Purpose Worldwide received our questions weeks ago. Revenue. Vista debt. Leadership. AI pivot. Their response remains a masterwork of negative space. Really impressive, actually, in a performance art kind of way.

iSpot might be the strangest company I've ever covered. And I've covered ad tech, so the bar for strange is somewhere in the Earth's mantle. They won a lawsuit, technically, and came out looking worse. Asked for $47 million, got $18.3 million, lost on trade secrets and DMCA entirely, and revealed their platform could be accessed for 21 months on stale credentials. I change my Netflix password more often than that. Their president Julie Van Ullen publicly agreed to talk to me on LinkedIn, where people could see it, and then got pulled back by what I assume was an urgent internal memo along the lines of "absolutely not, stop talking to the rabbi."

The agencies don't just dislike iSpot's leadership. They refuse to discuss it on the record after saying mean things. Several mentioned point blank they think they are “asshats” and absolutely them. In ad tech! Where people trash-talk competitors over warm rosé at Cannes before the appetizers arrive. When professional gossips go quiet, that's not discretion. That's fear.

And yet. iSpot's clients defend them. Brands call me off the record to say the product works. Agencies won't say a kind word. Brands won't say a critical one. Nobody will say anything on the record. It's like reporting on a restaurant where every line cook says the chef is a psychopath but the Yelp reviews are four and a half stars.

Here's something genuinely weird: the most substantive organization I've encountered in this entire investigation is the VAB. The Video Advertising Bureau. I had no idea who they were a few weeks ago. They sound like a Washington acronym that exists to justify a conference budget. But they're producing actual research with actual data and forcing actual conversations about whether the system underpinning $90 billion actually works.

Their December 2025 report on Nielsen's Big Data + Panel found 45 to 58 percent of total hours showing over 20 percent audience variance in key buying demos. NFL hours with double-digit discrepancies. Average network drops of 18 to 30 percent across 158 networks. Sean Cunningham called it "voodoo math." Nielsen called the report "seriously flawed and manipulated," arguing the VAB compared Live-only data to Live+SD, excluded OOH and DTVR, and essentially compared apples to oranges. Nielsen ran its own analysis over the same dates using L+SD program data and says more than 70 networks showed gains.

Both sides have a point. The methodological debate is real. But the MRC appears to be taking the underlying concerns seriously regardless, which tells you something about where the weight of evidence is landing. When the accreditor starts asking questions, "the other side's math is wrong" is an explanation, not a resolution.

I said in Parts I and II that Nielsen had a communications lockdown. I was wrong. Not entirely, but enough that I need to correct it publicly, which is something I do because I'm not a coward.

Nielsen is doing interviews. Karthik Rao went on the Next in Media podcast. Brian Fuhrer did The Town, did rounds with AdExchanger, Sports Business Journal. He speaks regularly with the New York Times, the Wall Street Journal, New York Magazine. These are real interviews with real outlets asking real questions about Big Data + Panel and methodology. That's not a lockdown. That's a functioning communications operation.

Here's what it is though.

They're doing all of those interviews. They're not doing one with ADOTAT.

When I ask for Peter Naylor, it's "we need the right time and news hook." When I ask for leadership on the specific questions I'm raising, about BD+P accreditation, about the VAB findings, about the contract structures, about the patent litigation strategy, I get background messaging and a very polite request to chat through my angle first. Elaine is wonderful. Genuinely. Best PR person in measurement. Sends me things proactively. Always responsive. Always professional. This is not about her.

This is about the fact that Nielsen will put its executives in front of the New York Times, the Wall Street Journal, AdExchanger, Sports Business Journal, podcasts, trade publications, and when the visibly Orthodox Jewish rabbi running an ad tech publication asks the same questions, the answer is "let's find the right time."

I want to be very careful here because there are two different things happening in this industry and they need to be separated clearly.

I am not accusing Nielsen's communications team of anything nefarious I genuinely don't believe that's what's happening with them specifically. They have been nothing but professional and proactive. The team has engaged with me more substantively than almost any other company in this series.

What I think is happening with Nielsen is simpler and in some ways more frustrating: I'm not the New York Times.

I'm a trade publication they didn't take seriously until my series started getting forwarded around their own building. And by the time they realized the coverage mattered, the questions I was asking had gotten uncomfortable enough that "let's find the right time" became the permanent answer.

I should note that both the New York Times writers, and the WSJ read this, prolifically. I can see their opens, clicks.

That's a business calculation. I don't love it. I think it's shortsighted. But it's not bigotry. It's a company deciding which journalists get access based on outlet size and question difficulty. Annoying. Not hateful. There's a difference and the difference matters.

Parts I and II covered the supply side. The measurement companies. Their tech, their finances, their lawsuits, their CEO carousels.

Part III is the demand side. The agencies and brands spending roughly $90 billion a year on television, $55 billion linear, $33 billion CTV and streaming, and collectively deciding that whether any of it works is someone else's problem.

Twenty-six senior executives sat across from Nielsen last June and told them the product was broken. Then they renewed. Every single one.

Nielsen, to their credit, didn't deny this meeting happened. They asked me what my "why" is. The "why" is the next three parts.

The measurement companies are a mess. But they're a mess the buy side chose. Every renewed contract. Every JIC certification nobody transacts on in September. Every "we're committed to multicurrency" press release that quietly dies when actual buying starts. Those aren't accidents. Those are choices. And Nielsen is making them very easy to make. They just renewed Paramount multi-year. Warner Bros. Discovery multi-year. Seven of the largest agencies including six holding companies re-signed long-term Big Data + Panel deals in 2025. Gray Media renewed for local across 113 DMAs. Hubbard Broadcasting renewed in Albany and Albuquerque.

Everyone keeps signing. That's not because everyone loves the product. The VAB report tells you what the networks actually think. It's because the system makes staying easier than leaving. That's the story.

Agencies aren't incentivized to fix this. Switching doesn't save them a dollar. It creates operational complexity for the same fee. Holding company CFOs would rather set money on fire. At least fire is warm.

The infrastructure, Media Ocean, Prisma, every agency tool, was built for Nielsen the way America was built for gas. The alternative currencies exist. The pipes don't fit them. One major broadcaster had to build an entire Snowflake workaround just to ingest alternative data because their tools physically couldn't process it. That's not resistance. That's plumbing.

And underneath all of it there's a number that should keep every CMO awake:

98 percent.

98 percent of ad impressions in one major campaign analysis went to people who were going to buy the product anyway. Not correlated. Causal. The methodology isolated incremental effect from baseline behavior. 98 percent didn't move the needle.

Wanamaker said half his advertising was wasted. He was an optimist by 46 percentage points. We have perfected the art of measuring nothing accurately.

Parts 2 through 4 are for paid subscribers. That's where we open the contracts, name the tools, walk through the pricing trap, explain why 97 percent of linear impressions go to 55 percent of viewers while the other 45 percent starves, and talk to the people your favorite trade publications can't get on the phone.

Because apparently the people your favorite trade publications can get on the phone won't take my call. Which is fine. I have other calls. They've been very informative.

Subscribe. You don't want to miss what comes next.

Every company received formal requests for comment with specific questions backed by documentation. The ones that responded told better stories. The ones that didn't told the best story of all.

The Rabbi of ROAS will return.