AOL’s display performance is in comeback mode, but growth is lopsided. The company’s Q3 earnings presented a clear demonstration of the strength of third party network, which includes the flagship ad unit Advertising.com, as revenues jumped 18 percent over the same period in 2011. At the same time, revenue from AOL’s owned & operated sites comparatively lagged, gaining just 7 percent in Q3, a possible indication that the focus on premium direct sales and the interactive, branding-aimed Project Devil ad units will take time to attain meaningful traction.
Ad.com CEO Ned Brody spoke with AdExchanger about trying to connect the seemingly opposing forces of premium and programmatic ad sales, and how AOL is competing more directly with the likes of Google, Rubicon Project and PubMatic as it prepares to release its own formal supply side platform to publishers.
NED BRODY: To be honest, I don’t think there’s that much that’s new per se other than we’re executing on our core strategy, as shown by the Q3 results, which was in part driven by third party network revenues and the shift toward programmatic buying. That strategy is to be a significant player in the ad technology space with a full stack across display, mobile and video. We launched our second platform product in addition to Adtech, which was the [demand side platform] AdLearn platform this summer and we now have four of the five agency trading desks signed up.
We’re excited by what’s been going on in the programmatic space. About 18 months ago, we said, “Look, we’re going to double down on the programmatic space, grow out our video and our Ad.com display business and the Adtech business. We promised a number of announcements this year. And there’s a couple more to go.
What is AOL’s approach to programmatic?
There is a subdivision inside programmatic, where you have the managed services and then there is the platform business. It all coalesces around the use of data in an automated trading environment in the purchase of advertising. There are two ways to do that. You can either in a fully transparent SaaS model, which is platform based, or you can do it using any number of companies that do the optimization for you.
If you look at the traditional Ad.com business, people call it a network. It really is an SSP – plus an exchange, plus a DSP – all for internal use. AdLearn, which is the optimization piece, does about five to seven billion auctions a day. They’ve just always been internal. We’ve been in programmatic. You could say that [current Videology CEO and Advertising.com co-founder] Scott Ferber and Ad.com invented programmatic 12 years ago. We’ve been in that space ever since. And that business continues to boom.
How are you going about creating the more formal SSP offering, which has been mentioned over the past few months?
The side of programmatic that people highlight is the platform side. We made the decision 18 months ago that we would enter the platform space in a big way. We opened up AdLearn, which is now a DSP product. As you noted, we publicly stated that we’re working on an SSP product, which is really just an exchange product. That’s really what they are. So the SSP/exchange will be built off the Adtech platform. You can see us over time evolve that into our mobile and video businesses.
We believe the market’s really going cross format. You hear that a little bit today, but you going to hear a lot more about cross format tomorrow. Advertisers are going to become more “format-agnostic.” They just want to use the best format to match their either brand or direct response need.
So with the forthcoming SSP offering, do you see AOL going head-to-head with the likes of Rubicon and PubMatic?
Absolutely. They’re absolutely to compete with those guys.
I know people think of AOL and Ad.com in the same sentence. The reality is that Ad.com as a group serves 25,000 publishers – one of which is AOL. And AOL is in no way the majority of inventory — it’s in the 25-30 percent range. All of our products are built to serve the markets at large, so our SSP products will be set to compete with Rubicon and PubMatic and AppNexus and Admeld, just as our DSP product is built to compete with Invite and Turn and everybody else.
As you sharpen the focus on the sell side, the big question is always about supplementing and not cannibalizing their direct sales. Is that a fear that you’re able to address directly?
That’s obviously the age-old issue in any type of non-reserved environment. That was the argument in the very early days of network. “If I allow someone else to sell my inventory, how am I ever going to protect my direct sales? Isn’t that going to have a cannibalistic effect on my reserved?” And the answer is it doesn’t because among other things, publishers do have control over their inventory. And secondly, advertisers approach direct sales and programmatic differently and often, in complementary ways.
How does the dual emphasis on “premium,” through the Project Devil ad units, and programmatic, through the range of DSP and SSP tools you’re building, work to complement each other? What kind of traction are the Devil units getting?
Devil is doing extremely well. Number one, on the reserve side, it continues to grow well into double-digit rates. We launched the Devil Network a year ago. At the time, the idea was that at the end of the day, networks need to be more premium. They need to have better formats, they need to be on better inventory sources. We believe that a 300×1050 unit on a network of top quality publishers will resonate well, even on a non-reserve basis. There were some skeptics. But we now we have over 100 publishers running Devil Network campaigns or having agreed to and in the process of doing the work to do so.
Some major publishers still consider RTB to stand for a “race to the bottom.” Is RTB compatible with premium inventory?
There’s no doubt that RTB can encompass premium. There are a couple reasons why. On one hand, everyone is, “Wow, look at the rise of programmatic” and then everyone is saying, “Programmatic is destroying yields for publishers.” It’s not necessarily true, but there are clear reasons why many people thought this way and some still do.
Early adopters of programmatic were clearly performance advertisers. I think that is the stigma that is associated with programmatic buying. More recently, you’ve seen people like Kellogg’s and Ford and Proctor & Gamble enter programmatic buying to seek out the very best consumers for their offerings. That still, however, tends to be at a price point below reserve.
What happens for publishers, though, is you really have to run an optimized mix of reserved and non-reserved. We have experimented with different parts of AOL where we have moved from reserved to non-reserved on those sites and have seen increases in yield.
It all depends upon the demographics and the elements of your audience and whether you can do it but I think we have a long way to go. The reason we have a long way to go as an industry in terms of helping publisher yield through programmatic is that we’ve really only adopted the very most common and basic formats. The majority of the programmatic trading occurs around 300×250’s and your traditional banners, 780’s.
The move to a premium format and other models in programmatic has just begun. One of the reasons it’s taken so long is that, because of the structure of the industry, innovation has happened at a micro- rather than a macro level.
How so?
SSPs have been innovating and they’ve moved from very static models to bid resolution to private exchanges. That’s been a nice evolution. DSPs have innovated on that side as well. But no one has innovated the entire stack. One of the reasons is that when you have lots and lots of fragmented competitors grouped in these niches, the only links, or least the lowest common denominator between these entities are essentially around APIs. We believe you really have to participate in all elements of the stack — SSP, DSP, exchanges in general, in order to be able to innovate.
Let’s go back to Devil’s example. Let’s say that when AOL was thinking about Devil as a product, for a second pretend we didn’t own an ad server with Adtech. You can pick whatever third party you pretend that we might have used at that point. We go to that third party and we say, “Hey, guys. We want to really increase our yield by creating a 1050-sized unit. And we want that 1050 unit to have interactive modules and it should suppress all other advertising that will be on that page.” What you think the response I would have gotten from the ad servers would have been?
They wouldn’t have been very polite, perhaps, but they would have said, “We’ll add it to our product queue,” and it’ll always be next quarter. Because we own Adtech, we were able to take that product and build it. And it wins the Rising Star at the IAB. It becomes a standard portrait unit. Now 1050 is a vibrant part of many, many publishers’ web sites. Some of it’s Devil and some of it’s not but we were only able to do that because we have a vertical stack that included both a publisher and an ad server.
What’s happening is because the only company in today’s model that owns that entire stack is Google, no one else has really innovated. I think Google is mostly focused on yield, in terms of the small increments, rather than how do you actually change the paradigm? How do you add a premium format as an exchange tradable unit? How do you really work on cross format targeting? How do you think about a uniformed approach to video, mobile and display? To make those things work in the marketplace, you have to have all pieces of the stack.
As you said earlier, Ad.com always contained a programmatic element in its DNA. But the idea of it being a third party network, not an exchange, was how it was designed and how it was largely seen by the marketplace. With all that in mind, does the emphasis on Ad.com as a programmatic entity change its original nature as an “ad network?” Does the basic concept of the ad network model still have any relevance in 2012 or 2013?
“Network” is just a name, right? I was just talking about this with Scott Ferber. One thing he told me was that he was thinking about television one day and he realized that he watches CBS. But in actuality, he’s really watching multiple producers’ content on CBS. “They call it a network in television so I’ll call it a network in advertising.” So that’s how this all got started.
As I said, the network is just a word for a multi-publisher and multiple-advertiser marketplace. Networks are really managed services systems. Those managed service systems will become more open and you’ll see more of a platform model emerging. The companies that can do that well will win.
