Avenue A: Rewriting the Rules of Online Advertising

It's an age-old problem: You know you need to advertise, but it's impossible to measure the real effectiveness of each piece of an ad campaign. That was true enough when advertising spots were limited to print, radio and television. But the digital platform of the Internet changes the rules of the game. On the surface, the Web seems like simply another avenue to get your message out – and another venue to complicate your efforts to measure ad campaign effectiveness. On the contrary: Web advertising is customizable, it's trackable and it's successful.

At Avenue A, overseeing successful digital media ad campaigns has become big business. The three-year-old company, based in Seattle, brings together the processes and analytical know-how that are required to make online advertising and digital marketing campaigns effective and efficient, explains Michael Galgon, chief strategy officer. Avenue A serves billions of Web ads per month, helping clients like Nabisco, Best Buy and Microsoft launch and monitor successful campaigns. They do it all: negotiating the best rates, identifying people most likely to purchase, placing banner ads in the most appropriate places, even advising which ads to show to which people.

Buy Ad Space Based on Past Behavior
So how might a company identify its best prospects? Not in traditional ways, according to Scott Fasser, senior product marketing manager. Avenue A begins by putting what it calls "action tags" on critical areas of the client's site, such as the home page, registration page, thank-you page, etc. As visitors move through the site, they trigger various tags, giving Avenue A behavioral data to work with. "Every time someone comes to this site, we log the unique Avenue A ID, which gives us information on what date they visited, how many times they've come back, if they made a purchase or not, or what part of the site they showed up on. In our analytical process, we go back and match up the information we've gathered on the site with the ads we've served for the client: Did someone see the Microsoft ad on Yahoo and later show up at MSN?"

The key to tracking this online behavior is the "cookie," a unique, anonymous identifier on every computer browser to which Avenue A serves ads. Cookie files are standard tools for identifying customers and personalizing their online experiences. At Avenue A, the cookies play a key role in gathering behavioral data.

"We probably have one of the richest databases – if not the richest database – of anonymous, individual customer behavior on the Internet," says Fasser. "Every time we serve an ad, we place a cookie on that machine, unless it already has a cookie. We serve billions of ads a month to some 4,000 sites, so we've got a great spectrum of seeing cookies on the Internet."

Avenue A intentionally confines its tracking system to watching online behavior rather than collecting personal information. Fasser explains that the cookies are "machine-specific, not person-specific," as the company strives to be compliant with all current privacy legislation and policies.

Fasser says, "The best way to buy media is to base it on prior behavior. Advertisers traditionally look at demographics for media purchases, but demographics are a proxy for behavior."

Brian Biglin, director of data strategy, compares targeting on behavioral data to targeting on demographic data. "Someone may be 75 years old and be hooked on MP3s, which doesn't match the normal characteristic of the segment," Biglin says. "Behavioral data is more reliable, and that is important when targeting the right people for a company's message."

Fasser cites the example of online pharmacy drugstore.com, which used Avenue A's Web affinity analysis to develop a list of almost 1.5 million quality prospects. After gathering anonymous data on site visitors, Avenue A segmented the customers, looking at the cookie sets that purchased the most, visited most often and registered most frequently. "And then we pulled the story around customer behavior on the site. Once we did that, and once we knew which cookies represented the highest value customer, we could see where else those cookies spend time on the Internet." That information, then, helped determine on which sites drugstore.com ads should appear.

"Using software like SAS, we've developed what we call affinity scores," says Fasser. And those scores, he says, have proven more valuable than demographics in determining where to place ads.

Send the Right Message to the Right Person
Avenue A's analytic acrobatics don't end at figuring out where to place the ads and then buying the space. "We've got a new capability where we can serve a unique ad to a unique cookie segment across the Internet. So when people talk about one-to-one marketing, we now have the technology to do that. It adds the actionable part to the equation and does it all through a single data warehouse."

Suppose you own a major travel and leisure Web site, suggests Fasser. Assume Avenue A has segmented your customers based on behavior patterns by looking at people who rent cars, people who stay in hotels, people who fly on planes, those who are visiting the site for a second or third time, and those who have never been there before. "You can predict who your best customers are going to be. You say, 'Hey, if I can get someone in and get them to use two of my four services, then I have a higher chance of them coming back and being a loyal customer.'"

Avenue A uses such information to conduct segmentation schemes and probabilities, figuring future value. "We can start pulling in things like, what's the probability of moving from the first visit to the second visit to the third visit, etc. Based on these probabilities of future visitation and purchase, how can we help our clients maximize customer value by changing the message between the first visit and the first purchase?"

Armed with this information, the travel and leisure company sees the value in advertising to those best prospects – but not with just any ad or message. That company will want to communicate to each type of prospect with an appropriate message. "You don't want to give the same marketing message to a loyal customer that we give to a prospect who has never heard of you before," says Fasser.

"If the highest affinity segments have never been to your site, you might want to send them a branding message to raise awareness. There may be a segment that's been to your site once but hasn't been back since; maybe you advertise a new feature, or offer them a special promotion. And then one segment is people who have come and used only one of your services. Instead of advertising the one service that they've used, you might try to cross-sell a service they haven't used."

So, when a prospect shows up at the spot where the travel and leisure company has banner space, Avenue A shows the ad that's best suited for that prospect. By making use of this customer targeting technology, your travel and leisure company is now personalizing ad messages – and increasing the likelihood of successfully hooking a customer for life.

Biglin says the advertiser also can use this knowledge to decide not to show ads to certain people. Based on Avenue A's probability models and customer segmentation using SAS technology, some cookies clearly will not be good candidates for certain products and services. In those cases, "we suppress the ads, so as not to waste the advertiser's money," he says.

"We're not reinventing stats here, but we are using them in some creative ways," Biglin adds.

With some 90 billion records in a proprietary flat file structure, Avenue A has an immense data bank on which to base these decisions. "Data is an asset," says Biglin. "But before we had SAS, we were not able to take advantage of that data. SAS is one of the tools that allows us to make use of it."

Measure Results to Know What Works
Before Avenue A developed a technique for tracking online behavior, it was extremely difficult to gauge how successful Web ads really were. How many times might it take a prospect to see a Web ad before finally responding to it? How many people see an ad, don't click on it, but show up at the advertiser's site anyhow? How many people see Web ads over the course of weeks or even months before needing that particular product or service? How would you know that the person had been viewing your banners for some time before typing in your company's URL?

Typically, you wouldn't know these things. But Avenue A is anything but typical as it uses SAS to rewrite the rules of advertising.

Copyright © SAS Institute Inc. All Rights Reserved.