Realeyes’ Max Kalehoff on the new metrics that matter in advertising

2025 has been the year of outcomes over impressions — and TV and streaming are leading the charge. Legacy broadcasters and newcomers alike are working to make buying and measuring TV ads as seamless as search and social, without compromising the insights and control that advertisers rely on.

One of the year’s standout partnerships paired Nielsen with creative intelligence and attention measurement firm Realeyes to launch an Outcomes Marketplace, giving advertisers deeper insight into how their ads drive business results. Realeyes powers the platform with a scoring system that predicts sales outcomes and measures creative efficiency across video and social platforms.

The Current spoke with Max Kalehoff, chief growth officer at Realeyes, about the nexus moment for outcome-based advertising, the company’s already on board, and what’s next.

It sounds like this partnership combines Nielsen’s strength in reach and frequency data with Realeyes’ attention metrics. Is that right to frame it that way?

Yes, because marketers are on the hook more than ever to drive accountability and results. Shareholders demand it.

The most important business outcomes are company valuation, customer lifetime value and sales. That’s the North Star.

[Marketers are now asking,] “How much can we tie our investment in advertising to sales impact, including growth and ROI?” And I think that really embodies the Nielsen outcomes suite of tools that complement the reach and frequency measurement that it’s so known for.

Have any companies signed up for this outcomes marketplace? When it was announced, Captiv8 was mentioned as a company that signed on.

Publicis’ Captiv8 is among our first customers.

We’re seeing excitement from some of the world’s largest retailers, the largest social-video and publisher platforms, major consumer advertisers, and the big agency holding companies. Demand is coming not just from the U.S., but globally including Latin America, Europe and Asia.

Momentum is building.

Why is the industry shifting toward outcomes now?

This is a philosophical question. It’s a combination of technology and competition.

If you had to point to a significant event that’s driving the conversation of outcomes, it started with Google’s debut of paid search advertising. They introduced very black-and-white sales and revenue outcomes to advertising that had never been seen before at that scale.

What did that lead to? Well, it happens to be that the architects of the Google advertising system then went to Facebook.

Those same folks created the modern system of online advertising that was objective based. Ten years ago, you could log into Facebook Business Manager and say, “My objective is sales or awareness.” That is objective-based advertising.

In order to be competitive, all the big platforms have started to frame the sale of their networks as objective-based. You invest in them to achieve certain business objectives. And that’s led to a label that we now call business outcomes. The scrutiny that companies have and CMOs have is not getting easier.

So competition and technology has led us up to this moment over the past 15 years.

Can you explain Realeyes’ ad scoring system? How do you predict sales outcomes?

The first solution out of the gate was human data collection using webcams to passively observe the visual attention and emotive response of viewers. That’s what we started on. That’s the gold standard of attention data. It’s all opted-in, first-party data.

We’ve built another product that complements that. It’s a scoring engine that can accurately predict human response and sales lift impact by analyzing ad content and placement alone, without needing to observe actual people.

So far, we’ve completed almost 19 million webcam observations of people viewing and responding to ads in 90 countries on every major platform. We use that as training data.

We can score a thousand ads in minutes, sometimes seconds, which means that there’s almost infinite scale. Every single ad content per unit can be scored and managed to achieve attention and drive sales, with consumers as an outcome. And because we know the link to sales potential, [this]can be used to predict sales.

We have multiple partners now that take attention data and plug it into their sales forecasting models to better predict sales performance and how advertising influences it. So that brings it back full loop into why attention has to do with outcomes.

Reach, frequency, confidence and trust behind that is the baseline. That’s what Nielsen’s celebrated for. And now we’re building on top of that to connect the exposure of advertising and media to business outcomes. The market wants a platform that brings it all together.

Attention equals sales. Attention equals business outcomes. And it’s predictive of what will happen in the future.

How long has Realeyes’ ad scoring system been around?

Human data collection, the way we do it today, has been around for about six or seven years. And what we call synthetic scoring, the ad assessment engine that requires no human participants, that’s been in-market for almost a year.

And do you have a ballpark percentage of how accurate these predictions are when you’re within the synthetic scoring system?

The predictions are 80% accurate to the ground truth of the training data. So, 8 times out of 10 times, you’ll get the same exact answer as if you were to actually sample a human participant panel.

For very high-stakes scenarios, like if you need to assess your Super Bowl ad, we recommend using the human sample because it’s the highest quality fidelity.

But for your halo ads, often 98% of an advertiser’s content volume, human data collection simply fails due to speed, scale and cost. So for the 98% of ad content that otherwise goes unmeasured, an 80% accuracy rate still achieves a transformational improvement by appending performance and learning to every single ad.

Our mid-tail and long-tail, especially with ad content volume, is exploding because of generative AI, that’s where the synthetic data methodology shines. Because of the economics, speed and scale, this enables an advertiser to append attentional performance and sales potential performance to every single ad unit that goes in front of a person for the first time ever.

And then what happens? Do advertisers get anything if the prediction is wrong?

No, it’s not a guarantee. Though if we’re right, most of the time when millions of decisions and billions of advertising dollars are at stake, you’d be wise to trust the probability of the data to achieve better outcomes.

The alternative is to fly blind or flip a coin.