When he founded Adelaide Metrics, Marc Guldimann was on a mission to clean up advertising. Too much of it was poor quality, intrusive to the viewer, and a waste of money for buyers.
His solution? A media quality metric that predicts outcomes more accurately than existing metrics. The theory being that such a solution would help advertisers invest in higher-quality placements and publishers more likely to create them. The big picture? A vibrant market built upon a more precise currency that can’t be gamed. Adlook caught up with him in Los Angeles to see how far he’s succeeded.
Key takeaways:
“As a publisher, you want certainty,” says Marc Guldimann. “With our system, publishers don’t have to gamble.” As founder and CEO of Adelaide, he heads a team dedicated to bringing that certainty to the media industry.
A machine-learning model
Adelaide introduced an omnichannel metric called AU, which measures any media placement’s probability of attention and outcomes. Designed to prevent gaming, AU is generated by a machine learning model that considers the characteristics of media proven to drive attention and is trained using full-funnel outcome data. While binary metrics like viewability treat reach as homogenous, AU provides a multidimensional view of media quality and a clear sense of what’s delivering results. Having analyzed 38 case studies in its latest “Outcomes Guide”, Adelaide claims that advertisers see an average of 31 percent upper-funnel and 56 percent lower-funnel lift using AU versus traditional metrics.
“With our system, advertisers can guarantee their media investment reaches a specific quality threshold,” says Guldimann. “For example, they can go to the market and say, ‘We need 33 AU and above’, knowing that’s the optimal level of quality for their KPIs. Then, the media seller simply says, ‘Okay, here you go.’ It creates a much more trusting environment.” they can get a 33-ranked placement.
Adelaide has spent the past four years developing its metric and has achieved some big wins recently. It concluded a deal with Adobe Advertising in June 2023, equipping advertisers with a push-button solution to apply attention measurement tags to campaigns and creatives, while delivering attention insights and trends within the platform. This comes on top of earlier deals with corporations including Microsoft, Mars, the NBA, and AB InBev.
Partnering with Adlook
In April this year, Adelaide partnered with Adlook to build a deep learning powered AU-CPM buying model to optimise attention goals and boost campaign performance, whatever the CPM. The partnership “advances our mission to establish a more fair and transparent market where publishers are rewarded for high-quality media, and advertisers can ensure they’re driving predictable business outcomes,” says Guldimann.
“We inject a quality score into the market so advertisers can understand what they’re paying too much for versus too little,” he continues. “We’re saying, ‘Don’t buy rip-offs, buy quality.’ Many issues in our industry can be traced back to incentives created by broken metrics—advertisers’ obsession with cheap reach, publishers answering to viewability and CPM targets with tons of low-quality inventory. We’re trying to solve these problems and realign buyer and seller interests around quality.”
In a further recent deal, the US National Basketball Association (NBA) partnered with Adelaide and found that, after implementing AU, it gained a 20 percent increase in attention score for a playoff campaign, while saving hours of time each week. “We now prioritize channels, partners and tactics that have strong attention scores,” said Larisa Johnson, NBA’s vice president of paid and CRM media strategy. “It’s a complex measurement, but it’s invaluable to reaching your target audience.”
An arc of attention
The complexity stems from understanding “the arc of attention over time,” explains Guldimann. “Media creates an opportunity for attention, then creative holds that attention for as long as the audience likes what they’re seeing. So, when an advertiser optimizes impressions to increase the duration of attention, media is not the only input affected.”
Without understanding these nuances, users can fall foul of Goodhart’s Law (‘when a measure becomes a target, it ceases to be a good measure.’) and introduce unintended biases. “For instance, we know older audiences spend more time viewing ads than younger people. Adults spend more time with things they’re familiar with. Optimizing creativity to max attention may favor more salacious content,” says Guldimann. “The most important lesson here is to steer clear of optimising the amount of attention and instead use attention data to make good investment decisions,” he stresses.
Adelaide’s foundation was in part a response to people trying to game the advertising system, offering placements with artificially inflated viewability metrics. With its own system relying upon automation, Guldimann recognizes that Adelaide must work hard not to fall into the same trap. “We’re always looking for people trying to game the system,” he says. “It’s cat and mouse, but we’re built in such a way that we can stay ahead of the game. We want to replace viewability and bring more trust to the market.”
Inertia, parsimony and greed
Ranged against Adelaide are the forces of inertia, parsimony and greed. “A lot of agencies have to deliver on CPM figures—everyone wants the lowest CPM, which equals the biggest revenue. So, everybody wants more small, sticky impressions. I hope we can dry up the well of refreshing ad slots, shrinking video players, and ineffective formats and ultimately right this wrong,” he says.
In terms of sustainability, Guldimann argues that attention metrics are the way forward. “This is the solution to reducing carbon emissions from media. You can fiddle with SPO [Supply Path Optimization] but the best solution is serving fewer, higher-quality ads.”
The trouble is, as Guldimann concedes, that AI is likely to spawn an almost unlimited amount of new content in the coming years, pushing supply and demand far out of balance. This will create yet more challenges for advertisers, as they struggle to identify which content is consumed by which people, and consumers themselves prioritize channels with zero (or fewer) ads.
With this blizzard of content on its way, Guldimann advises advertisers to look carefully at his company’s AU metric. “The probability of attention is VERY strongly correlated with outcomes,” he says, just to make that super clear.
Marc Guldimann studied Social Decision Sciences at Carnegie Mellon University in Pittsburgh, United States. He then founded a series of digital media companies, starting with Spongecell in 2005, which was sold to Flashtalking in 2018, then Parsec in 2015, sold to Kargo in 2020 and most recently Adelaide Metrics, which he founded in 2020. Guldimann also co-founded The Attention Council, a non-profit organization based in New York, which promotes the use of attention metrics in the advertising and marketing industries.
