'Ads by Google': A Billion-Dollar Brainstorm Turns 10

On June 18, 2003, Google made an announcement. Following the successful launch of AdWords, the firm was going to extend its search-based ad program with a new service that would allow web publishers to serve ads themselves -- ads that would be precisely targeted to the specific content of their individual web pages. "With Google AdSense," Google announced, "publishers serve text-based Google AdWords ads on their site and Google pays them for clicks on these ads." This, Google reasoned, was a win-win: "users benefit from more relevant ads and publishers can maximize the revenue potential of their websites."

Over the past ten years, that core idea -- giving advertisers the ability to reach across different web pages, dynamically -- has grown into a service that, in some very real senses, helps to fuel the economic ecosystem of the web. If you see an ad on a random website, there's a good chance that the ad is being served with the help of AdSense. In June 2004, AdSense launched display ads. In 2005, it teamed up with Blogger to help blog publishers earn money from their content. In 2007, it added ads for video and mobile content. In 2009, it added expandable ad formats and interest-based ads. Come 2011, more transactions were made through AdSense every day than on all the world's major stock exchanges.

AdSense "basically turned the Web into a giant Google billboard," Danny Sullivan, editor of Search Engine Land, told USA Today in 2007. "It effectively meant that Google could turn everyone's content into a place for Google ads." And that, in turn, helped to transform Google into ... Google. AdSense, after Google's 2007 acquisition of DoubleClick, is now the largest player in the display ad business, and it has provided Google with a secondary revenue stream that complements the company's income from basic search ads.

Google is quick to point out that the windfall benefits more people than Google alone. Last year, AdSense shared $7 billion with more than 2 million publishers, Google notes. There is now such a thing as an "AdSense millionaire."

AdSense, in other words, is a classic Silicon Valley thing: a simple insight with wide implications. And the woman behind it, in this case, is Susan Wojcicki, who helped to build the program from an idea into a technology back in 2003. (You might also know Wojcicki as Google's new SVP of Advertising and Commerce, or as the renter of the garage that served as Google's first makeshift headquarters.)

I spoke with Wojcicki about AdSense's growth over the past ten years -- and about what we might expect from the service over the years to come. The conversation below is lightly edited for length.

Reading that first AdSense press release, I'm struck by how closely the service of 2013 resembles the service as introduced in 2003. How do you think of it, 10 years later? What do you think are the most significant aspects of its evolution?

AdSense was really revolutionary at the time, because advertisers before that were advertising, say, on a sports section. And we could actually dynamically serve them on the exact sport that they wanted. So if they were interested in hockey, we could serve them on all the stories about hockey that were running dynamically across the Internet.

And pretty much from the first day we released AdSense, we could tell it was going to be successful with publishers. We had so many publishers signing up, and it fueled a lot of innovation for publishers in enabling them to generate revenue in all kinds of ways that they might not have imagined -- and that we didn't necessarily imagine, ourselves.

So we started with this relatively straightforward concept, which was taking amazing technology that lots of people at Google had worked hard on, and then putting it on publishers' websites. And then one thing led to another. So we developed this technology, and then we realized that publishers wanted to have all these different tools, and different controls. And advertisers wanted to have all these tools and controls.

It's amazing for me, because it started as this relatively straightforward product. But the amount of investment we've made over the past 10 years has grown it into a really sophisticated suite of products for both publishers and advertisers -- something that enables publishers to monetize their sites, and advertisers to be able to get the reach that they want.

What does that product look like going forward? How do you see it changing? 

The interesting thing about advertising is that it's business-to-business - but it needs to move as fast as consumers move, because advertisers need to be where consumers are. So if all consumers are suddenly consuming all their media online, then advertisers need to move as fast as the users just moved.

That puts a lot of interesting pressure on us, and on the industry, to innovate as fast as consumers innovate. And advertising is a huge industry. So being able to try to move that all online, and make that really effective for publishers and advertisers, is a really big opportunity for everybody - for Google, for advertisers, and for publishers.

So looking forward, we're really focused on trying to think about how to solve that: how to continue to grow the ecosystem, and how to continue to enable online advertising to be more effective for advertisers. And we're looking at the brand opportunities as more and more users are moving online.

Although we've invested really heavily in this business for the last ten years, it's an area that we plan to continue to invest in equally heavily going forward. I think the next ten years will be as exciting as the last ten years were.

Has your sense of what "relevance" means, in the advertising context, changed at all over the past 10 years?

What we pioneered 10 years ago was contextual targeting: the idea that if you have an article about cooking, or making a craft, or travel, then we would serve an ad that's relevant to that. So if you have an article about cooking, you have, say, an ad for the ingredients that go in the recipe.

What's evolved since then is the recognition that there are some types of sites where the context isn't really relevant for advertisers. And probably the most significant area in that regard has been news. You read about what's happening in the Senate, or you read about a bill being passed, or you read about a controversy that's happening at the moment - and that's not necessarily a commercial type of content. It's very important, but it's not commercial.

So what we've really invested in has been audience targeting, and really trying to understand who the users are, and to understand what they want to see.

So if you have a person who is reading a lot of articles and headline news, what are the things that are relevant for them? It's trying to understand your user base a little bit better. The more you know about the users, the more you can serve something that's useful for them. So audience targeting, and serving ads to the right audience, is something that we've invested a lot in.

Do you see the Knowledge Graph, and Google's broader investment in semantic search, playing a role in its approach to ads?

The nice thing about the Knowledge Graph is that it gives you different answers [than what you would get from a standard search]. And it can understand different components in the search. Although it's probably early for us, in advertising, to think about, I think whenever you understand meaning better, you can serve better results. And since advertising is just another form of information, it means you can serve better advertising, too. But it's still early.

What about the ad units themselves? What do you think will be next for those?

I've talked before about the 5 Cs: five of the things I think are relevant for the future of Internet advertising. And two of those that I think are relevant here are Choice and Charm. When it comes to Choice, the users should be able to opt in to seeing the ad units. One of the places we've done that has been on YouTube, where we've enabled users to opt in to TrueView, which allows users to choose whether they want to see an ad or not. So an advertiser pays only if a user chooses to see that ad. And that's also how search advertising works: the users are choosing to see that ad by clicking on it, and the advertisers pay only when the users click on it.

We've also developed some other ad formats that run across the Internet, engagement ads: an ad where, if you mouse over it, it expands with a video or a catalog or something right there on the page. So, overall, we've seen a lot of engagement by users. Which makes sense. If you seen an ad, for example, for a brand you like -- say, J. Crew -- and you click on it, and that brings you to the catalog ... that's a pretty compelling experience.

The other thing we've done in the engagement format is to live-stream different events out of it. So we've had a lot of different mobile operators -- T-Mobile, for one example -- doing their launches of phones with the streaming. So you're reading an article, you see the ad, you click on it - and here's this live event. So that's a pretty interesting format.

And, overall, I think the formats are going to continue to be more beautiful, more charming, to have richer experiences for the users. But the users will also be able to choose whether they want to see them or not -- and be able to opt in or out of those experiences.

Along those lines, I'm really interested in the signals AdSense reads to determine which ads to serve, and where (and when). I do a lot of Googling and Interneting every day; only some of my searches are for subjects and items that I have a commercial interest in. How do you think about the difference -- when it comes to, say, the keywords that help inform contextual targeting -- between standard web searches and commercially driven web searches? How do you think about distinguishing between intellectual curiosity and commercial intent -- or, I guess, between different forms of interest?

I think that's an example of ads getting more and more sophisticated. How do you understand something in the context of other signals? How do you understand whether a user is interested in something, or not?

I have a third C, which is control. And one of the things I've been advocating for, and that I think our advertising has been pushing for, is more control for users. So we've released some sets of things, but I think there's a lot more that we can do. How do we enable users to tells us, "These are not the things that I'm interested in," or "I want to see more things like this?" So if you were just researching something and now you're finished researching it, and you don't really want to see anything more related to it, how do you communicate that back to the ad system?

And that's another area that I think we can really invest in more in the future. Especially because if you say you are interested in something, that's the best possible signal to an advertiser. If you're going rafting at the Grand Canyon, for example, you'd probably love to see Grand Canyon-related offers and different opportunities, because you're in market for those. But when you're not in market for them anymore, they're no longer relevant. So I'd really like to see users, in the future, being able to be more a part of that process.

I think most users tend to approach ads -- especially, though not exclusively, online -- as things that are, in some fundamental ways, opposed to their own interests. Ads are the stuff we have to endure to get to the stuff we actually want to see. I know it's implied in the idea of "relevance" that ads at their best are, in fact, in our interest ... but I'm wondering how you see that tension playing out in the future. Is is possible to align advertisers' and users' and publishers' interests so efficiently that we'll actually look forward to seeing ads? 

We've always thought about advertising as information. But advertising is just commercial information. So the challenge is understanding what's the right ad, to serve right then and there, for that user. And that's why, in some ways, it's very similar to search: when someone gives you a query "restaurant Palo Alto" - what results do you serve them? It's all about relevance. Advertising is the same thing: you have a page and you a have an ad slot. What's the most relevant thing that you can serve right then and there to the user?

Advertising does have a few different things that you should think about: there are only certain advertisers advertising at a certain time. But advertising can be very, very relevant - and very, very useful to users. And that's been our goal: to try to figure out how to make it useful to users.

One other thing that's important to mention: advertising really does fund the Internet for being free. And that is a really, really important component. You think about all the people who have access to all this information - everywhere, globally - because the Internet provides so much information that's free. And the reason it's able to do that is because of the advertising. So the advertising does fund the Internet. And that has a lot of social and economic benefits for people everywhere.

And it will just get better. As the advertising gets more relevant, as we understand the users better, as we understand the context better, it will just keep getting better. And publishers will make more revenue.

So that's our goal: to continue to support the ecosystem so publishers can continue generating good content. They need to have the income from the advertising to be able to support their sites. And we recognize that. The better job we can do with that, the better job they can do reinvesting in their site, and into the content.