^{In 2014, the actor and his friend, the entrepreneur Daniel Nadler, founded EDO to compete for the last frontier of legacy media measurement. Now, investors are placing bets on startups that could win a slice of a $65 billion TV ad market.}

The actor Edward Norton has played a lot of characters. He’s been an insomniac underground fighter. A private detective with Tourette syndrome. A border patrol captain and friend to the concierge at The Grand Budapest Hotel. An illusionist. A scout master. A chauffeur. A dog. The Hulk. A neo-Nazi. A Catholic priest.

But here’s one role that would never make it into a Hollywood script: The cofounder of a startup challenging the status quo in the wonky world of TV measurement. In this case, however, the plot is real life.

Norton is the cofounder of EDO (Entertainment Data Oracle), a data and analytics company that applies machine learning to how advertising is measured across traditional TV. Having spent his career marketing films and understanding spending strategies, he recalls seeing firsthand how data science could help the entertainment industry. The idea: Wean the marketers off its reliance on old-school TV ratings to better connect the dots between the TV ads people see and what they search for online after.

“If Fox spends inefficiently on Grand Budapest Hotel, [director] Wes Anderson and all the rest of us actually make less money,” Norton tells Forbes. “So it affects the compensation of creative people and affects the risk profile of studios assigned to different types of content. If they can have more confidence in the efficacy of their spend, they can maybe make more interesting stuff.”

As traditional networks try to give advertisers the same accuracy expected from digital giants like Google and Facebook, measurement has been increasingly top of mind. That’s leading EDO and its peers to compete for what many say is the last frontier of the legacy media measurement. In the past year, investors have poured hundreds of millions of dollars into companies they hope will win a slice of a $65 billion TV ad market.

EDO, founded in 2014, is just one of a number of startups vying to compete with Nielsen, the longtime industry stalwart that experts say has had a near-monopoly hold on the TV ratings and measurement industry. Nielsen, founded a century ago, has faced a number of setbacks as advertisers and networks demand better ways to track audiences in the streaming era. When the Media Ratings Council—an influential nonprofit that audits how companies measure media–suspended Nielsen’s accreditation last summer, partners began looking for alternatives, startups entered newly open doors and investors started betting on potential challengers.

“It’s ridiculous that you’re proposing brain surgery on me and my business with a Stone Age axe in the era of the gamma knife. I want the gamma knife. I don’t want a Stone Age axe on my head, meaning my company, because this is too freaking important to me.”

On top of the massive pot of TV ad market dollars, there’s a lot of money to be made by chipping off Nielsen’s business. Nielsen—which is preparing for a private $16 billion acquisition led by Elliott Management and Brookfield Asset Management—reported $3.5 billion in revenue in 2021 with fourth-quarter revenue of $894 million and first-quarter 2022 revenue of $877 million. (For comparison, ComScore, a legacy rival of Nielsen, brought in $96.5 million in the fourth quarter of 2021.)

Vinny Rinaldi, head of media analytics, data and technology at The Hershey Company, says the advertising world prides itself on consistency. But that doesn’t work when things change.

“We’re going into a world where TV of the old is entering a new way,” Rinaldi says. “The gap is tightening and closing with what we define as TV…Who sits down at night and says, ‘Do you want to watch linear tonight or streaming?’ Nobody does that.”

It might seem incongruous to have a prolific creative like Norton as the mind behind a company using machine learning, but he’s not doing it solo. He cofounded EDO with his friend, the poet and entrepreneur Daniel Nadler, who sold his AI startup Kensho Technologies to S&P in 2018 for $550 million, which at the time was the largest AI deal in history. They also have the backing of billionaire Jim Breyer, a celebrated investor who made early bets on other startups-turned giants like Facebook, Spotify, Etsy and Marvel Entertainment. Breyer’s firm, Breyer Capital, led EDO’s $12 million Series A round in 2018.

“Edward is extraordinarily insightful and thoughtful whether it’s media, whether it’s the worlds of colliding industries of entertainment and technology,” says Breyer, who knew Norton while on the board of Marvel. “He has great passion, great insight, and puts significant work in in terms of understanding next-generation opportunities, which is not always the case in Hollywood or the world of entertainment.”

Using algorithms that “watch” DVR programming across 120 cable and broadcast networks, EDO’s tech matches around 100,000 commercials a day with data about what people are searching for on Google in the minutes after ads air. Over the past seven years, EDO’s tech has watched more than 200 million ads, which lets the company run data models to see how a 30-second ad works on live sports versus cable entertainment, how certain demographics react differently than others, or how ad placements early in a show perform compared to those that are later. EDO can also help TV and film studios measure exactly when viewers are more engaged with a movie or show.

“Search, when used properly, is highly predictive of market share,” says EDO CEO Kevin Krim. “If you find someone who is growing their share of search before share of market, then their share of market eventually catches up.”

Krim—who was a media executive at Bloomberg and CNBC before joining EDO in 2015—says EDO’s analyses of search data show a 90% correlation between organic search and future market share for auto and restaurant brands. For CPG brands, there is a 80% correlation and a 70% correlation for insurance brands.

“When you go to Google as an advertiser or go to Facebook as an advertiser, you don’t say ‘I would like to reach 67% of males 18 to 49,’” he says. “And yet that’s the way TV’s been sold since the 1950s or so.”

While talking with Norton about the idea for EDO nearly a decade ago, Nadler—who wrote his Harvard doctoral thesis about economic and statistical modeling—was surprised that Hollywood still heavily relied on emotion and intuition when deciding which movies to green-light and how to market them. Although alternative data was a radically new idea in 2014, it’s now the norm to measure search traffic, social media metrics and other proxies for measuring awareness and intent.

“A lot of what I’m saying now is super obvious, but at the time it was still the old studio era,” Nadler says. “You still had a Bob Evans-type model where you have a guy who’s really good at feeling out whether a script will work, whether it should become a movie, how to finance the movie and what the budget should be.”

There are some differences when it comes to using machine learning in finance versus entertainment, Nadler notes. There’s a century of economic data at a granular level for public markets, but the measurable history of entertainment doesn’t go as far back. On the other hand, a chaotic financial system with many participants makes it impossible to predict. But with advertising, far fewer participants makes it a “much, much simpler experimental design.”

“It’s the classic experimental designs of psychology in the 1950s and 60s,” he says. “Which is where we have a participant subject and we’re going to give them stimuli A and a stimuli B and we’re going to see their reaction and measure their reaction.”

There are other factors that also could play into EDO’s appeal. Showrunners want more data from streaming services about how content performs. Marketers say platforms like Hulu and Roku have failed to provide enough data on ad performance. Meanwhile, Netflix’s recent subscription slowdown has the company planning to peel back investing in its own content as the possibility of an ad-funded Netflix has marketers eyeing new territory. (Disney also plans to offer a cheaper ad-supported subscription tier later this year that will reportedly include 4 minutes of ads per hour.)

Networks and advertisers have been focused on testing to see what works best in an ever-evolving landscape. NBCUniversal, after running pilot programs during the Super Bowl and Winter Olympics, named iSpotTV as its official currency—industry lingo for a measurement metric. Disney tested around 100 vendors before naming Samba TV as its first third-party vendor in March. Horizon Media, one of the largest U.S. media-buying agencies, is planning to commit as much as 15% of its upfront ad-buys to different currencies. And just last month, EDO inked a new deal with Discovery as its official partner for measuring how ads affect consumer behaviors.

“I think marketers universally regret that they let the Googles and Facebooks of the world grade their own homework,” says Chris Kelly, CEO and cofounder of Upwave, a marketing analytics startup. “Marketers have bellyached for years about how vulcanized measurement is in digital and I don’t think they’re going to allow similar vulcanization on TV.”

Despite the popularity of streaming, Norton notes that increased scrutiny of digital has shown weakness with mobile advertising. (For example, muted videos and multitasking users.) However, he thinks TV is still a highly effective medium and says companies like EDO now have ways to prove the medium’s impact at a more granular level.

“It’s ridiculous that you’re proposing brain surgery on me and my business with a Stone Age axe in the era of the gamma knife,” Norton says, referencing the advanced radiation treatment used on cancer patients. “I want the gamma knife. I don’t want a Stone Age axe on my head, meaning my company, because this is too freaking important to me.”

EDO’s pitch is gaining traction. The company says its 2021 annual recurring revenue was $16 million, and top brands like Toyota, New Balance, IBM, and Royal Caribbean to measure how ad creative and length impact effectiveness. Other clients include entertainment powerhouses ViacomCBS, Warner Bros, 21st Century Fox and NBCUniversal. EDO is also attracting new investors: Last month it announced $80 million in new funding from Shamrock Capital. (The startup wouldn’t disclose its current valuation, but it was valued at $89 million after its Series A round in 2018.)

“It’s indicative of the change of the industry more broadly,” says Laura Held, a partner at Shamrock Capital, which in April invested $80 million in EDO. “The industry is ready to embrace alternative solutions in a way they would not have a few years ago…For so long everyone was reliant on Nielsen as sort of the industry standard and it’s hard to break that even if it’s not the best way to do things over time.”

Michael Piner, EVP of advanced media at the IPG-owned agency MediaHub, who has been using EDO for two years, says the data gives clients a TV equivalent of online click-through rates. Although data for online searches and TV-related search have been around for a while, he says EDO “really operationalized it and brought scale to it.”

“It’s a leading indicator of outcomes because it’s an immediate measure of an increase in consideration and intent because people not only saw your spot,” Piner says. “But also started to engage and search for it…People Google something when they’re interested in it.”

Better tech doesn’t necessarily mean a better understanding of audiences overall, notes Mitch Metcalf, a former broadcast executive at NBC and ABC. He notes that although startups might be better at measuring certain subsets of audiences, it doesn’t mean they see the whole thing at once. And although there are still issues with Nielsen’s sample sizes, he says they’re at least they’re measuring “real people and real homes.”

“The problem is that it’s an excellent, excellent view of a subset of the audience,” Metcalf says of measurement startups. “It doesn’t really give you a view of all smart TVs or, God forbid, TVs that aren’t smart.”

Whether EDO can hit Nielsen with its data-powered slingshot is a story that doesn’t have an ending—yet. Later this year, the goliath plans to debut its Nielsen One offering, which aims to have a better overall view across streaming, linear and other online platforms. There are also still questions about whether marketers and networks will totally make the switch or simply wait for Nielsen to catch up.

“Nielsen is so entrenched in advertising measurement and insights,” says Tina Moffet, an analyst with the research firm Forrester. “They understand advertising as a whole and the practice of advertising. I don’t think they’re going away anytime soon.

When asked how the potential of EDO compares to past investments across other startups that then became giants, Breyer says one of the most significant criticisms he’s received in the past with investments like Facebook, Spotify and Etsy is that the people have thought the market opportunity might be limited. However, that’s never worried him.

“Once a company has embedded itself into a circulatory ecosystem,” Norton says of legacy companies. “The length of time that it can sort of mush along on cruise control without significant adapting…It’s sort of like Stockholm syndrome.”