The problem with Internet advertising isn't that there's too much of it (or, these days, less and less of it), or even that most banner ads make 30-second TV spots look like Oscar material. No, the problem is that Internet advertising just isn't smart enough.
How long have marketing executives been listening to change-the-game rhetoric from business pundits, Net companies, and online-ad agencies? We all know the jargon: "one-to-one marketing," "mass customization," "permission marketing." And yet, even today, most companies don't really know what they are paying for when they buy an online ad.
A few years back, when online advertising first became a bona fide industry, it borrowed its pricing structure from the television and print-media industries. In those media, the cost of an ad is based on how many people see it — that is, on the number of "impressions" that it makes. The digital world sets rates by measuring the number of people who either see an ad banner or actually click on an ad. But even so-called "click-throughs" have turned out to be a largely illusory way of assigning value. That's because the trail is broken once a clicker lands on an advertiser's Web site. Did an ad attract someone who actually purchased a product or signed up for a service? Or did that potential customer immediately disappear to another corner of the Web? It's difficult to answer those questions, and as a result, it's hard to justify outlays for online advertising.
Of course, the billions of dollars spent on TV, radio, print, and billboard advertising are subject to the same uncertainty. But digital marketing, by its very nature, should be more precise — and hence more accountable — than traditional marketing. Wasn't that its very promise?
The folks behind Avenue A, a 4-year-old company based in Seattle, aren't ready to give up on that promise — even if they face an uphill fight in a hostile business environment. Back in 1997, when cofounders Scott Lipsky and Mike Galgon were forming the company (which now has more than 450 employees), they approached several big advertising agencies with a simple question: Why aren't your clients spending more money on the Web? The response was that advertisers had no idea what they were paying for, since they couldn't determine whether the ads resulted in actual sales. Lipsky, now 36, had worked previously at Amazon.com, where he built a data-mining system with the goal of zeroing in on the behavior and tastes of Amazon's customers. But here was a problem of even greater complexity: "We needed to figure out a system that could serve up ads and also track sales — by following what people were doing once they clicked on the ad and landed on an advertiser's site," Lipsky explains.
Creating that system required writing new software and developing a new approach to following a browser tool through the intricacies of the Web. That innovation, developed first by Lipsky and later mimicked by several of Avenue A's competitors, makes clever use of the conventional way that ads are transmitted to a computer screen. When a user lands on Yahoo!'s main search page, for example, it triggers a signal from Yahoo! back to the browser to send an electronic request for the ads that occupy that piece of Web real estate. Each ad comes from a different computer server, and it flashes across the Internet to the browser, often in the form of a small, self-contained application that offers the same piece of animated eye candy over and over again.
Lipsky realized that the same sequence of electronic requests and deliveries could be used to track a browser after it clicks on an ad and gets linked to an advertiser's Web site. With the advertiser's permission, Avenue A places tiny ads, no larger than a single pixel, on key pages of the company's Web site — for example, any page that contains an electronic shopping cart, or the Thank You page that a shopper sees when he or she completes a purchase. These so-called "action tags" are invisible to the shopper, but they trigger the same kind of request as a full-blown ad, and that request is in turn registered on Avenue A's server, leaving a trail of electronic crumbs that leads (ideally) to a purchase. (Throughout this process, the shopper remains anonymous; Avenue A's software tracks only the actions of the shopper's browser.)
In other words, for the first time, companies can track with precision the effectiveness of their advertising, and as a result, they can calculate the real return on their investment in digital marketing.
This approach entails number crunching on a grand scale. Young-Bean Song, director of analytics at Avenue A, recently ran several million Web transactions through a program to prove a rather intuitive point — that the number of people who simply click on a company's ad has nothing to do with how many people actually buy or sign up for something on its Web site. "Eighty percent of companies still base their online-media buying decisions on click-throughs," says Song, 26, who is also director of the Avenue A Institute, the company's research arm. "But it's not about the click-through; it's about what happens afterward. That's the only way to get to ROI."
Most of the analysis that Song and his colleagues perform is more targeted, more down-to-earth. Each week, Song's team generates a fresh round of analysis for Avenue A's clients, which now number more than 100. Because it negotiates flexible contracts on behalf of its clients, Avenue A is able to dump whole classes of online ads that aren't working and to concentrate each client's marketing dollars on higher-performing ads. Using such an approach, Eddie Bauer was able to decrease its marketing cost per sale by 74% over three months. Likewise, Microsoft's Expedia travel site cut its cost per sale by 91% over eight months. "We counsel our clients that this medium is still pretty immature — that we need to continue to look for efficiencies, as well as load up on placements that we know are effective," says Brian McAndrews, 42, president and CEO of Avenue A.
Beyond the Gut Check
Once the logic of Internet advertising moves from a leap of faith to ROI, a lot of interesting strategic adjustments can take place. Consider the example of Blue Nile, an online jewelry retailer that has so far managed to avoid the fate of most other e-tailing sites. Until last summer, Blue Nile had spent most of its advertising dollars on TV commercials and costly sponsorships that gave the site preferred placement on AOL and MSN. But Blue Nile's marketing director, Kevin Keith, 30, a refugee from Procter & Gamble, thought that the company could boost revenues quickly by running online ads with the help of Avenue A. Senior executives at Blue Nile were dubious of that strategy, but they agreed to try it. During the week in July when the online campaign began, traffic to the site jumped by 25%, and today the company garners $3 in revenue for every dollar that it spends on online advertising. "The old adage at P&G was that you plan, you do all this work, and then you pull the trigger on the campaign and just hope it does well," Keith says. "You have to watch for a month or so before you know anything, and if it's not doing well, it takes another 6 months to fix it. You're working in 6- or 12-month cycles. Here, the cycles are 2 weeks."
How did Keith generate such a boost in performance? A typical Blue Nile diamond buyer visits the site four times before making a purchase, so knowing which ads a customer has seen, and in what sequence he has seen them, is critical to figuring out what's driving sales. Sometimes a pattern jumps out from the data. Two recent ads sought to appeal to the same type of consumer — a groom-to-be looking to save money on a wedding ring without looking like a cheapskate. The first banner asked, "Can a diamond be a sensible purchase?" The second one said, "Have more money for the honeymoon." People from Avenue A quickly noticed that the "honeymoon" ad was drawing far more customers. They alerted Blue Nile, which then asked its creative agency to come up with more online spots like that one.
Compared with the creative leaps made by the legends and luminaries of the ad world — David Ogilvy, William Bernbach, Jay Chiat — the performance-based approach of Avenue A lacks a certain poetry. It is, at bottom, a highly rational pursuit. But it delivers genuine results. "The core skill of the future is not 'Do I have a good gut sense of what works, a single important creative idea?' " says McAndrews. "It's 'Can I come up with 10 ideas and then test them to find out which one works?' Which means that a campaign is a living, breathing thing, and analysis becomes the central part of the marketing activity."
Waste Not, Want Not?
The challenge for marketers today is to recognize which kind of customer they are addressing at any one time and then to serve that customer the most appropriate message. On the Web, meeting that challenge is an immense task, requiring a system that can test and analyze billions of records, and then act on that analysis in real time (which, on the Web, means "in less than one second"). Such a system moves a company a big step closer to one-to-one marketing, yet it is built on a foundation of financial rigor that allows the company to determine the precise cost of acquiring each customer. In other words, it moves the ROI calculation down to the level of an individual customer.
Officially, Avenue A calls this new method "customer-life-cycle marketing." But Diane McCowin likes to call it "the Matrix," after the science-fiction movie in which much of the action takes place in the disembodied realm of cyberspace. McCowin, 32, vice president of analytics, comes to Avenue A from Bank of America, where she was senior vice president for strategic information management. Petite and intense, McCowin gets excited by customer modeling and data mining — the analytic techniques that are at the core of the new approach — and her enthusiasm is infectious. "This moves advertisers from 'Where are the sites that have the demographics that I'm looking for?' to 'Where are my customers, or where are the prospects who most closely resemble my customers?' " she says.
The implications of that shift are intriguing. Instead of negotiating for space to run an ad on a Web site, advertisers might negotiate for customers that fit a certain profile. "We usually know what an ad space is worth better than the site itself does, and we can use that to our advantage in negotiating," McCowin says. "We might be willing to pay $5 for that sports fan, while Yahoo! thinks he's only worth a dollar. At the same time, we might pass on a guy who is less valuable. In negotiating a contract with Yahoo!, one term might be 'What percentage of users can we just pass back, rather than sending an ad to them?' "
Avenue A has bet the company on its new approach. Three of its clients are testing the complete system, and eight others are trying out various parts of the customer-life-cycle marketing platform. The good news for Avenue A: Its system will become even more crucial to advertisers as the Internet becomes broader in its reach and more diverse in its formats, covering everything from broadband television to wireless devices. Marketers will need advertising technology that can not only find a company's customers wherever they might be in the digital world — logged on to the Web, watching TV, or dialing in from a handheld device — but also recognize instantly which ads they have already seen and how best to move them to the next step in their relationship with a company.
The bad news for Avenue A: In the short term, at least, most of the news about digital marketing is bad — from the collapse of the NASDAQ, to the slowdown in online-ad spending, to the ongoing (and escalating) legal concerns about privacy on the Net. Not surprisingly, Avenue A's stock price has plummeted in the past year, going from a high of $89 shortly after it went public, in February 2000, to just a couple of dollars now. Despite the company's efforts to add more established companies to its client roster, dotcoms still make up roughly half of its customer base, and its revenue growth has slowed as the cash resources of Net startups have dwindled.
Meanwhile, the company has become a legal target: On November 20, the notorious (in high-tech circles at least) class-action law firm Milberg, Weiss, Bershad, Hynes & Lerach filed suit against Avenue A — and also against MatchLogic, a subsidiary of Excite@Home — charging that each company tracks customers without their permission. Avenue A is fighting back, claiming that its technology identifies only anonymous Web browsers, not individual users.
Coupled with the downturn in spending on Internet marketing, that lawsuit makes it harder for Avenue A to chart its own course as an independent company. But Brian McAndrews remains undaunted. The CEO doesn't believe that his company — or online marketing in general — will fall by the wayside during the tumultuous correction now under way in the Internet economy. In fact, he sees an ironic side to the chill that has descended upon the nascent digital-marketing industry. "Because the medium is so accountable through all of the data it generates, people are hard on it when it doesn't work as well as they thought it might work," he argues. "But advertisers follow eyeballs. They always have. There may be a lag, but eventually money will flow into this medium. We're talking about a world where not a single marketing dollar is wasted — where every dollar you spend is targeted toward somebody who is receptive to your ad. No one is going to walk away from that."
Paul C. Judge (firstname.lastname@example.org) is a Fast Company senior editor. Learn more about Avenue A on the Web (www.avenuea.com).
A version of this article appeared in the March 2001 issue of Fast Company magazine.