• by Joe Mandese  @mp_joemandese, July 12, 2011

They say the devil is in the details, and in the following interview AOL CEO and chairman Tim Armstrong provides them about Project Devil, the code name for an ambitious project to redesign the way people experience the Internet, including not only display ads but also content and navigation features.

What is Project Devil, and why is it important to the future of AOL, and the future of the online display-advertising experience?

The Internet that was designed 20 years ago wasn't designed with display advertising in mind. It was designed as an informational medium. It's been very successful at that, but if today were your first day on planet Earth, the most important trend now would probably be the connection of people to information over digital networks, and you would probably expect brands to be a huge piece of that, but the way the Internet was designed hasn't allowed for consumers to have great interactions with brands, and for brands to have great interactions with consumers. Project Devil is really meant to fix the connectivity between consumers and brands.

So it's a retrofit?

I think the Web is moving from algorithms to design. And at a broad trend level, I think that's an important part of what the Internet is going to be about. The technology has gotten better. The infrastructure has gotten better. The algorithms have gotten better. But the one piece that has not gotten better is design. And design is a huge part of what's going to happen on the Internet in the next 10 years.

So Project Devil is about design, interface and usability?

Yes. We actually started Project Devil by taking a ruler and measuring our own pages to see how much of the page was content, how much was navigation, and how much was advertising. In essence, 20 to 30 percent of our pages were ads, but on average, there were 14 ads on a page. So the average consumer was seeing 14 ads. And the average advertiser was next to 13 other ads. From a brand mathematics standpoint, that design is not going to work. A lot of people have misunderstood what we have been trying to do with Project Devil. We didn't give up ad real estate on our pages. We actually have the same amount of ad real estate. It's just one advertiser using it, instead of 14. The press look at it as, "Oh, they're putting bigger ads on the page." But there's a mathematical formula behind it, which is: one consumer and 14 ads versus one consumer and one ad. And that is a big mathematical benefit for the consumer, for the advertiser, and for the metrics that sit between them.

Where did the name Project Devil come from?

When we realized that what we had was a design problem, we decided to go talk to some of the best design and brand people in the world. One of the people we went to see was Anna Wintour at Vogue. We met with her numerous times, and after one of the meetings, we started thinking, "How would Anna Wintour think about the Web and how to design Web pages?" So we took the name from The Devil Wears Prada. Although Anna is a friend of mine, I don't think she knows that. It came from really thinking about our pages from a design standpoint. How could we make them more beautiful, but also really allow brands to tell stories? And it was the meetings we had with people like Anna Wintour that inspired us. So you spoke with a lot of aesthetic design people, not just technical design people? Our viewpoint is everything from industrial design through aesthetic design. I would expect us to keep innovating on the types of designs we are doing around advertising and page design. They are integrated. Project Devil may be perceived as an ad project, [and] it was, but in essence, it's redesigning the whole page.

How has it impacted users' experience? Have you seen any trends?

Yes, we have huge amounts of data. One point is that, on average, people spend about 47 seconds with a [Project Devil] ad, which is almost three times the time people spend with other ads. Engagement metrics are through the roof. On average, there has been more than a 300 percent increase in engagement rates for our ads. People are interacting with the ads more. They're looking at the video more. They're looking at the store locators. It's almost a full funnel marketing effect. People are dwelling on the pages more. Almost twice as long.

Obviously, you've been able to boost the effectiveness of the ads, but you've also been able to increase the CPM commensurate with them. How sustainable is that? Are you achieving the same yield by reducing the number of ads?

We actually are making more money as a company by taking ads off and putting one ad on a page than we were before. The rates have gone up by a commensurate rate to the engagement levels, and I think that's what advertisers are paying for. As an advertiser, what you're really paying for is the ability to talk one-on-one and create better levels of engagement. That was reflected in your first-quarter earnings report, which was the first in many quarters to show an increase in advertising revenues. How sustainable is that? I think there's a lot of upside. I think the challenge, as Project Devil gets up to scale, is getting all the creative agencies to scale. I think we have a lot of work to do in this industry, and we plan to share the success and metrics to help people keep innovating on what they can do creatively. But I think we've seen a big receptiveness from ad agencies, especially creative agencies.

A lot of it hinges on the portrait unit you've designed, which won the IAB's Innovators award. How important was the acquisition of Pictela in terms of fueling Project Devil?

We had launched Project Devil a few months before we purchased Pictela, but Pictela brought something very unique to the table. We at aol spent a lot of time on the design of Project Devil - the look and feel of it - and what Pictela brings is actually an operating system within the ads. In essence, in the broadest context, the way we think about Project Devil is that you're not just buying 20 percent of the page. We're investing hundreds of millions of dollars in great content and services, and what Pictela allows an advertiser to do is to program their 20 percent of our page in real-time. Instead of looking at it as a straight ad model, they're basically getting an operating system to program 20 percent of the page and to update the content in real-time. Pictela is really a content management system for advertisers.

So it's not just a content management system for advertisers. It's really an operating system for advertisers that can serve anything in real-time?

We think the lack of understanding comes from an old mentality of serve-a-page/serve-an-ad versus engaging the user. When you start from there - with a content management system for ads - you can start to focus on what the users want from advertising.

But you're not really personalizing that content at this time?

We do use data to help personalize the ads for people. But one of the interesting things with the privacy debate that is beneficial about Project Devil, which we did not set out to do, is that because you are able to program ads almost the way you would content, you don't need personally identifiable information to do that. By giving people more relevant content, we make it so they can have a better experience with it. Over time, data will just become more and more commoditized in terms of how you talk to consumers. It actually drives commoditization, which is what's happening with the marketplace right now. Project Devil and Pictela allow you to refresh content in front of somebody and attract them in a much more engaging way.

Long term, what are the implications of turning over 20 percent of your pages to advertisers?

Conceivably, if they use it right, the advertiser might get a bigger bang with that 20 percent than you're getting with the other 80 percent of your content. I would hope that at some point advertisers do as good of a job programming as we do by focusing on our great content. They already have the assets to work with. The average Fortune 500 company spends about $200 million a year creating their own content for themselves, including white papers, videos and all that stuff. Let me give you an example. [At this point Armstrong gets up and begins surfing a major technology marketer's site. He walks me through a ton of brand content, including video programming.] We were having the same conversation... They have tons of stuff, but the thing that is amazing is that they actually have their technical product managers taking you through every product in detail. It's amazing. Nobody knows what they have. I was actually taking them through their own site and showing them what they have. AOL doesn't exist in a vacuum. You're operating in an extremely cluttered online content- and display-advertising marketplace that seems to be getting more fragmented and cluttered each day.

What kind of effect do you think you can have on the rest of the marketplace?

You've seen [Terence Kawaja's] famous chart about the online display ecosystem. What we're doing is basically trying to create an experience that sits on top of that. It is based on a very simple premise, which is: what's the user experience with the ad itself, and what's the advertiser's experience with the user? What all these companies are trying to do, in essence, is optimize the experience of a user with data - one pixel to one user and optimize around that. All these companies come in and pitch us as well, but it's really hard to differentiate among them. I mean, "What's the difference between your data and the last person that was in here 10 minutes ago?" But when you sit on top of that, you can understand in a very simplistic way: 20 percent of the page targeted at one user. What that means for us is 118 million people domestically and 250 million people globally. You will be able to talk directly to them with 20 percent of our page. Now, let's talk about what you put in that space to make it really relevant to them. Yes, using data can help, but the average user doesn't get up in the morning and look at data. They get up in the morning and want to look at contextualized content. That's what they're attracted to.

Is it your job to influence the rest of the ecosystem, or can AOL pull this off just by changing behavior within its own pages?

I don't think it's our job to affect the rest of the ecosystem. I think our job is to basically have a differentiated scaled experience for our products and services. I mean, you could make the argument that as the rest of the Web is going in a more cluttered direction, we will stand out even more. Our overarching strategy has been to go in the opposite direction of Silicon Valley. We've done that with our content business, and we're doing it with our ad business. We're still very technical and Internet savvy, but we're taking a giant step back and saying, "If the Internet were created today, how would you design things for people?" Not the fact that we've been living in it for 20 years. It's a tough thing to do. When I look at our own sites where we haven't fixed things yet, you can see a big difference. I feel like, as a consumer, it's just not a good experience. And most of the sites on the Web look like this. [Armstrong loads a page for aol's Moviephone.] This is not a great design. We should be able to do this better. It's one of our older pages, but if I measured all of this, probably 60 percent of the page is not content.

So you've started with your AOL home page, and your other top pages, and you're gradually working your redesign through everything published by AOL?

Everything. The first day we integrated Huffington Post, it had Project Devil on it.

Who else do you think is doing it right?

Hearst is doing an interesting job of looking at the brand-ad space. Some of the Conde Nast stuff is very good. Look at the new Vogue; they've done a good job. I'm not a huge fan of pushdown ads.

It's admirable that you're giving the user more control.

Users are smart. We start with the premise that if you put something up that is actually reasonable and well designed, users will understand it. We shouldn't be tricking users into being exposed to ads or engaging with stuff. One of the things we've just started working on is Project Devil for video.

Same 20-percent concept?

Yes. [He plays a sample showing a Project Devil video prototype.] The idea is to have a fifth of someone's time. We're measuring it to find out what works best, but it's the same idea with video.

So in that case, it's duration, not just spatial?

No, it's also spatial. It's time and space. We're also connecting Devil locally on Patch. For example, local display ads haven't been designed for car companies. Car companies have their advertising broken up by national, super-regional and local. Project Devil enables them to put all three components on the page, based on the way they break it up nationally and locally. It allows them to show the design of the car, show what the pricing is for that region, and show what dealer you can go to. And they can connect it all in the same unit. [Armstrong opens a Google search bar to begin looking for samples of Project Devil campaigns.]

So you still use Google to search for things within AOL? Is the user focus of Project Devil a backlash against the algorithmic experience you had at Google?

Well, it's not a backlash. Search works very well for what it was designed to do. AdSense launched out of my group at Google so that we could bring relevant ads to you all over the Web. But it was all designed around that lower [marketing] funnel, and it always seemed like the top of the funnel was the largest opportunity for where the Web is going in the future.

And Project Devil is a good test case, because you've managed to improve your CPM by improving that experience.

Some people say content on the Web isn't going to work. But it's not the content that's the problem; it's the ads that are the problem. And that's what we've been trying to solve. [He pulls up an example of a Sprint ad for the "first eco-friendly phone," which is capable of being recharged with solar energy.] Let's just say Sprint came to us, and we put up a tiny banner to show this. What it would say is, "Buy this for $19.95. It's the first eco-friendly phone," and the user might think they are going to be tricked into something, versus what this is: the ability to actually demonstrate what the phone is, how it works, and why it is different. And then tap into our other assets, like MapQuest, to find a store locator, and find a Sprint store near you.

Is there any other technology you need to enable what you're trying to do?

I think we're pretty good on technology right now. We'll probably keep adding technologists and more engineers, and we'll make more acquisitions if it makes sense, but for now, we feel like we've done a pretty good job of getting the right talent at the company.

While you've been building out your redesign and technology overhaul, you've simultaneously been making content plays. How has this influenced the display advertising story you're trying to tell?

I think it is important for us to have really good content brands. It's analogous to American Express sponsoring the u.s. Open. Marketers want to be in areas where their brands are going to get an uplift from the environment. One of the things we have to do to create better brand advertising experiences is make sure our content experiences are great. And right now, we are probably one of the largest investors in high quality content on the Web. And I think we will probably continue to do that. It is tied very closely to this project.

So basically, you've been zigging while other people are zagging online?

A lot of people still don't understand what we're trying to do. They think it's "bigger ads, bigger ads," but it's about creating a better user experience. This is powerful. It's still a very, very small percentage of our total inventory, but it is growing. But eventually it will become the majority of our inventory. We need to spend time with the creative agencies to get there. And you have to sit down with advertisers to say, "You have to treat this like content." And that is going to take more work, but the results are going to be better.

Aren't they moving in that direction anyway? All we hear about these days is the mix of "paid, earned and owned."

It is. Most of the time when we sit down with marketing people at a company their content people are in the meeting. And the content person might spend $50 million a year creating things like white papers and videos that they're not actually deploying as part of their brand content. So there is a productivity issue that we are trying to solve, too.