September 20, 2019

Bringing Back Ad Networks

When I first started building websites over a decade ago, there were a couple of ways for a small site to make money: Adsense and random, crappy ad networks. As I moved through college and early after graduation, publisher-led networks started to pop up.

I remember one project I worked on, we worked first with Say Media and then right before I sold my equity in that project, we had explored jumping over to the Complex Media Network.

By the time I got to my current posting, the ad networks were mostly an after-thought and all the talk was header bidding programmatic. And as we continued to move into this more programmatic world, more companies got in the middle of the advertiser and the publisher, a consequence of which was the drastic reduction in ad rates that publishers earned.

But let’s not forget that CPMs were once pretty okay. According to a Digiday article published in 2012:

Across the Complex Media Network, CPMs run between $13 and $17 on standard display depending on the category. It can go up from there if there’s custom content, like video. Antoniello says it has high CPMs because of its unique content. He explained that the company vets all of its publisher partners upon recruitment, so no linking strategies.

It’s wild to think about, but only seven years ago, publishers were earning double figures on their standard display. Now maybe you earn a couple of bucks.

Axios put together a great graphic that shows how the ad buying experience has evolved over the past few years.


On the left is when I was getting my start. On the right is how things have been the past few years. Those trading desks, DSPs, SSPs, exchanges, and networks are all eating up parts of what the advertiser is paying for an impression.

Think about it this way… Let’s say that the advertiser is willing to pay a $10 CPM. The agency obviously gets a cut, but that’s always been the case. Then it hits the trading desk. They take a cut. And then it hits the DSP, so they’re taking a cut. Along the way, there are some DMPs and other tools to help more efficiently target the ads. So, they’re taking a cut. Then there are tools that are actually connected to the individual publisher, so they’re taking a cut.

By the time it actually gets to the publisher, that $10 has been shaved down. Every step along the way is taking their pound of flesh.

It was all necessary, though, because all of these different sources gave advertisers the ability to really target who they wanted their ads to hit. It didn’t matter where the ads were, just that they were being seen by the right person.

But as GDPR and CCPA continue to become a reality, and more importantly, as browsers continue down their path of increasingly blocking cookies (I wrote about ITP 2.2 a month ago), these tracking ads are going to become more difficult. Without the data, advertisers don’t really know who is seeing the ad.

I’ve been talking to colleagues about this for a while now, but as this trend continues, there is going to be the resurgence of contextual ad networks. If you’re looking to target individuals that are consuming financial news, for example, it’s possible that there will be a network that has relationships with publishers directly.

Another option is what The Washington Post announced on Tuesday, which is the launch of its Zeus Prime ad product. It’s a self-serve platform that will allow advertisers to purchase premium ads on The Washington Post and, in the future, a network of other websites.

AKA: An ad network.

Axios broke the story, so it’s worth including a few sections from the story to dive deeper into what this announcement entails.

Advertisers often complain that they would like a better alternative to buying ads on Google and Facebook — where the content isn’t always vetted — but there are no other places where they can buy ads as quickly and efficiently in real-time. The Post hopes this product will change that, and put more ad money in publishers’ pockets.

While a publisher can choose to license Zeus Prime as a standalone product, if it wishes to join the ad network that The Post is building, it needs to license all three of The Post’s commercial software products, including Zeus Insights, The Post’s first-party data tool that is used for ad-targeting, and Zeus Performance, its advertising performance tool.

At its core, this is a licensed tech stack, which The Post’s VP of Commercial Technology and Development, Jarrod Dicker, told Axios will cost anywhere from “the low volume range, half million annually and at the high range, in the millions.”

Let’s look at those tools in a little more detail:

  • Zeus Insights: This is the contextual targeting component of the platform. As an advertiser, what sort of readers do you want to target? Said another way, what kind of content do you want to target, which you believe your target audience will be consuming?
  • Zeus Performance: This is the ad analytics platform. Advertisers are going to want real-time, high quality data to determine engagement around the actual advertisements.
  • Zeus Prime: These are premium, social-type advertisements. You can drop in a tweet, Facebook post, or Instagram post and then the ad is automatically created. You then add a CTA underneath it and voila, without having to design or code it, you’ve got a great looking advertisement. To achieve these high quality ads, The Washington Post partnered with Polar.
Source: AdAge

It took me a little while to really understand why this product actually mattered. I had a discussion with Sean over at Industry Dive about it and we were both pretty confused about why real-time ad buying was so important. Neither of us had run into situations where advertisers needed to get an ad up and running immediately.

It was after spending time looking at the above graphic that I finally realized the real excitement around this product.

There are already companies that give publishers the ability to offer standard IAB self-serve ads on their sites. I’ve had good success with BuySellAds.

But, I think advertisers are growing tired of buying little boxes. They want something premium. However, they are also not looking to sign insertion orders with publishers because those are slow, annoying and remove a lot of control from the advertiser.

What made this announcement appealing is the type of advertisement. Brands are able to purchase native-like ads from publishers. Before this, the only way to do it was to work with a specific publisher that partners with Polar or to ignore publishers and work with Facebook direct, which offers these premium products.

“Advertising today for publishers is on opposite ends, it’s either very premium for custom experiences or very cheap for audience targeting,” says Dicker. “We want to bring demand back to the middle. If we do that, we’ll be bringing an entirely new revenue opportunity for publishers to band together and really take on Big Tech companies.” 

There are also secondary benefits. Now that so many steps in the programmatic process are being removed, there are fewer participants taking a cut. You don’t need a DSP, trading desk, SSP, etc. to get your ads onto a publisher’s website. Instead, the advertiser can go directly to the publisher and purchase an ad.

Dicker says that he believes publishers will be able to earn CPMs of more than $10. That’s a huge boost from what so many publishers with programmatic advertising are earning.

The real opportunity is when The Washington Post starts to scale up its ad network, giving the advertiser the ability to go directly to the ad network, select the type of audience it wants to target (based on what content they’re reading) and then the ads can be distributed across all the premium publishers. This will open up additional spend for publishers that might be regional brands and not on the advertiser’s radar.

The thing is, Facebook still has a ton of scale and the targeting abilities are better, not to mention giving advertisers the ability to bring in their own data. While this isn’t death to Facebook, it could at least be an attempt at publishers retaking some of the revenue. That’s what makes this project unique because if enough publishers license the tech from The Washington Post, scale naturally occurs.

So far, it doesn’t appear that The Washington Post is going to be taking a cut of the revenue that flows through the network—their revenue is from the licensing fees—so publishers will need to do some math to determine if the increase in revenue can offset the additional technology costs.

But let’s hypothesize for a minute… What if licensing fees go away?

That’s what Tony Haile, founding CEO of Chartbeat and now CEO of Scroll, is effectively saying here.

Rather than charging publishers for the various aspects of the publishing stack, give it all away for free. In exchange, you get a percentage of the revenue that the publisher then generates. It’s not too dissimilar to what Substack does here, but it would be at a far greater scale. Would The Washington Post make more money this way? Would it introduce more publishers to the stack than would otherwise have access? Could it help local news publishers that would greatly benefit from these systems, but don’t have six figures to license technology?

Nothing points to this happening, but the only two companies that are really building out these stacks are Vox and The Washington Post. There are only so many premium publishers that can afford licensing fees. What comes next?

All in all, I find this ad product to be very interesting. By partnering with Polar, The Washington Post gives ad buyers the chance to target audiences in a post-cookie world with premium ad-types. Whether publishers can make as much money as Dicker says remains to be seen, but I am more optimistic than I was on Tuesday.