How an AI-powered ad inventory increased Hearst Newspapers’ mobile CPMs by 30%-50% 

Andre Richards
Associate Director of Programmatic Yield

“We’ve been in partnership with Browsi since 2017, I’m pleased with the attention provided and willingness to work within our business needs. Introducing AI to the process of optimizing our ad inventory has helped us take our advertising game to the next level. We didn’t believe it’s possible to increase viewability and CPMs so dramatically. Browsi is revolutionizing the industry.”

We’re certain most of you are familiar with Hearst, the American multinational mass media and business information conglomerate. They’re essentially a media beast with more than 146 million people for an audience. What you might not be familiar with, though, is that Hearst – just like pretty much any publisher out there – sought to maximize the revenue coming in from their ad inventory.

Here’s a short story on how Browsi managed to help Hearst boost CPMs and viewability across platforms, ultimately bringing in more revenue – without damaging the user experience.

The Challenge

What Hearst understood is that in order to increase their revenue, they needed better ad inventory. With many direct and programmatic advertisers looking to buy its media, the opportunity to sell premium ad inventory couldn’t be ignored. Their ad inventory quality had to improve.

As the main challenge, standing in the way of that goal, they identified optimizing their ad layout for viewability and scale. Like many other publishers, Hearst was using wireframes to design their ad layout and offer fixed ad locations for their mobile sites, recording low ad viewability and as a direct result, lower than expected CPMs.

Delivering viewability for the direct campaigns and PMPs was also a challenge. Hearst was relying on historic viewability scores which required ad ops footwork and resulted in massive ad waste. They were looking at a glass ceiling.

The Trial (and Error)

Trying their best to optimize their ad inventory the old-fashioned, manual way, Hearst resorted to the usual:

  • A/B testing ad layouts and sizes
  • lazy loading some of the ads
  • resorting to sticky placements
  • Manually moving direct campaigns between placements in order to hit campaign goals and avoid ad waste

And yes, all these things showed improvements, albeit limited. Why?

Because at the end of the day, publishers simply can’t adapt manually to real-time variables around content, behavior, devices and ad stack. It’s humanly impossible to respond and make an impact in the milliseconds between a page loading and the visitor consuming it.

As a result, no matter how many ad layouts and sizes they’d try – they would always end up going with what the majority enjoys – not what *everyone* enjoys.

They could only go so far.

Browsi and AI to the rescue

As we said, Hearst was right to think that it was humanly impossible to make significant changes in that short window of time between a page loading and the visitor consuming it. But with the help of more than “mere” humans – Artificial Intelligence (AI) – it suddenly becomes, well – easy.

Using AI, we helped Hearst handle all the decisions around viewability, scale, and pricing in real-time – from page load until each ad is shown.

We deployed:

  • Personalized ad layouts
  • Real-time Viewability prediction
  • In-view ad refresh
  • Predictive ad loading

The Results

Hearst’s ad inventory now automatically adapts, in real-time, to user behavior, page structure, devices, and internet connection. The resulting ad inventory is personalized, bringing higher viewability, scale, and CPM without compromising the user experience.

And we have solid numbers to back it up. With Browsi, Hearst achieved:

  • A 30%-50% increase in CPMs for mobile
  • A 25% increase in viewability on its mobile platform

The best part is – Hearst Newspapers is delighted with the results.