A Matter Of Timing

In the past year, many leading marketers stepped up their presence on the Internet. Visa blanketed the Web with rich media ads as a major component of “Life Takes Visa,” its first new brand campaign in 20 years. IBM launched an online campaign for two new B2B platforms, “What Makes You Special?” and “Take Back Control,” seeking to reinforce its image as a technology innovator. Amgen and Wyeth Pharmaceuticals used a series of online ads to build awareness for the rheumatoid arthritis drug Enbrel. And, last fall, Microsoft lured videogame players with an interactive campaign for Perfect Dark Zero as a cornerstone of its Xbox 360 launch.

Not only are these companies spending more money on more sophisticated advertising, but they are also spending more time analyzing the success of their efforts. In discussing what works, however, marketers are learning there are various ways to measure success. What one marketer considers a reliable gauge of interactivity may have little relevance to another.

“One ad could have the goal to get people to interact; another could be to generate traffic; another, conversion,” notes Saar Safra, director of rich media at aQuantive’s Atlas Solutions, Seattle.

As marketers move away from click-throughs, richer creative and the ability to track interactions has taken them to deeper levels of measurement. As a result, the industry is starting to develop new metrics to reflect the quality of the online experience.

“We are moving past the standard measures of interactive media of impressions and clicks, to measures that take into consideration that [consumers] are time-starved and have complete control of the things they want to read and experience,” says Eric Wheeler, CEO of Neo@Ogilvy, a unit of Ogilvy Interactive, New York. “Getting your target to interact, read, explore, play and/or ask for information over several minutes is an enormous feat.”

Thanks to recent advances in technology, companies can obtain detailed data on Web user interaction within days of a campaign launch. Two of the most commonly measured variables are “timers” (how much time users spend within specific components of an ad) and “paths” (the order in which features within the ad were viewed or clicked). A banner ad from backpack maker Jansport, for example, allows users to drag a magnifying glass over a product photo to change colors and zoom in on its details; a tab allows them to fill in a sweepstakes form and send the ad to a friend. Research can tell the company how much time users spent checking out the product, the colors that were chosen in their precise order and which of those other steps were taken next.

Similarly, Warner Home Video can learn not only how many users were exposed to its ad for the Dukes of Hazzard DVD, but also how often the video clip was viewed or downloaded, whether a distributor was requested, and so on.

The process of delivering ads to publishers has become fully automated, says Young-Bean Song, vp-analytics at Atlas, allowing researchers to monitor interaction with “tags” that measure paths, timers and other metrics.

The definition of success, of course, depends on the marketer’s objectives. “Direct-response marketers will define success as a conversion: Did the user go to our Web site or buy our product?” says Safra. “Brand marketers may [ask] did users spend 30 seconds viewing an ad?—or define success as whether users explored a certain part of the ad.”

This ability to go deeper and faster into measurement has contributed to the rapid growth of optimization. The term refers to how marketers shift dollars based on a statistical analysis of the media buy (with variables including ad placement, location, size, frequency and behavioral targeting), creative strategy (ad themes, colors, animation) and pricing. Most marketers deploy agencies or third-party ad-serving firms such as Atlas and DoubleClick to use optimization tools, which in some cases can dramatically improve a campaign’s performance (see “Optimal Effect,” right).

Direct marketers, Wheeler notes, may achieve a 20 percent to 50 percent increase in response rate as a result of optimization. “We can now say to a big brand advertiser like Budweiser, ‘Your CPM on Yahoo is three times greater than on CNET,'” adds Song.

Still, experts caution against reading too much into initial optimization data or shifting gears too quickly based on its assessments. “Intelligent and diligent optimization is the key to any campaign’s success, but it takes time,” says Derek Leedy, vp/account director at Mediasmith, San Francisco. “Knee-jerk reactions can cause advertisers to make decisions on small amounts of data. There are times when you have to stay the course to find long-term partners that can generate both efficiency and volume.”

In fact, some argue, engagement is only the beginning.

“We are interested in online metrics in the sense that we want to make sure we are reaching our audience and that they are engaging with the ads. However, those metrics are not what help us to define a successful campaign,” says Jon Raj, vp, advertising and emerging media platforms at Visa USA.

The success of the interactive portion of “Life Takes Visa,” says Raj, will depend on the extent to which the multitude of ads moves the needle on key brand attributes. Visa is in the process of gleaning that information the old-fashioned way, with traditional brand impact studies. “Someone clicking on a banner doesn’t mean anything to me. I’m interested in: what effect does it have?” says Raj.

For example, one of the rich media banners, “Life Takes Imagination,” allows users to “paint” an abstract design by moving their cursor over a blank canvas in the ad. An out-of-home installation, “Life Takes Creativity,” is built around the same concept, only the brush strokes are made by people walking over the ad. Visa has been conducting surveys to help determine the impact of the installation. That additional procedure will help assess whether the campaign is meeting its ultimate goal: to shift the perception of Visa’s brand from one of convenience (as embodied by its previous tagline, “Everywhere you want to be”) to one of inspiration and individual empowerment.

IBM, too, has supplemented the tracking of its rich media ads with awareness studies. However, its current Internet campaign has a large direct-response component, as opposed to Visa’s pure brand play. As such, the metrics shift accordingly.

Consider “What Makes You Special?” a series of online ads targeted to businesses that have been running on the home page of WSJ.com and within the Web pages of CNN/Money.com and Forbes.com. Users who click on these banners and display ads are brought to a campaign hub (www.ibm.com/innovation), where they can read about IBM’s products and services, download a podcast or register to receive more information from the company. While IBM measures the response to all of these offerings, its ultimate goal is to generate new business.

“All of the advertising is expected to link to demand generation and sales, and leads are a good way to demonstrate and quantify that linkage,” says David Rittenhouse, group planning director for IBM at Neo@Ogilvy. “We usually look at registrations and opt-ins as indicators of how a campaign is performing.”

IBM also uses optimization to evaluate its online ads. The company is learning how many users are watching the 30-second streaming video to completion, rolling over or expanding the ads, clicking through to the landing page, and more. But while Rittenhouse relies on the analytics for a “numerical story,” he believes nothing can replace the value of human interpretation. “Sometimes the most intelligent optimization is a logical, not a technical, solution. It might simply be a matter of noticing, ‘Wouldn’t this be better if we made the ad copy [harmonious] with the content on the page?'”

Conversely, the results from optimization may run counter to logical assumptions, says Brad Bender, vp, optimization solutions at DoubleClick, New York. “Sometimes the best-performing [audience] segment is non-intuitive.”

Bender points to a packaged goods company for whom DoubleClick ran an ad test on a popular Web site, concentrating the ads on the site’s home and garden section; the marketer assumed that this audience would be most receptive to its message. Optimization showed, however, that it was actually the worst performing segment; ads placed in the hobby section got the most clicks. In a follow-up campaign, DoubleClick diluted the concentration of ads in the home and garden section and spread the campaign more evenly across the entire site. The result: a 30 percent lift over the initial test, says Bender.

Another tool helping marketers understand user demographics is Nielsen Net/Ratings’ NetEffect, an extension of the research firm’s Yahoo Consumer Direct service established in 2004. AC Nielsen (which, along with Adweek Magazines’ IQ, is owned by VNU) compiles data from 140,000 households who scan the UPC bar codes of products purchased at the grocery store. Participants download a software meter that tracks their online activity, revealing frequency of exposure to online advertising across multiple publishers. Nielsen overlays the two sets of data, comparing the purchasing behavior of groups that were exposed and not exposed to a product’s online advertising to determine its effect on that product’s sales. Follow-up surveys via e-mail determine whether a brand’s objectives were achieved, querying aided versus unaided recall, purchasing intent and a host of other issues.

Measuring engagement is particularly important to the pharmaceutical industry, where companies are shifting more resources online, while patients are increasingly turning to the Internet to seek out information on treatment options and side effects. Amgen and Wyeth, which co-market the arthritis drug Enbrel in North America, have been running rich media ads for the drug on sites including Mayoclinic.com and Webmd.com. A click on a request for treatment options links to a survey form where users can sign up to receive more information on Enbrel, which is indicated for multiple types of arthritis.

In this case, acting may mean filling out the registration form or requesting an information kit. But it may also mean talking to a doctor about the drug, a behavior that is less easily measured. “My job is to create qualified leads to the sales force, so everything I do is intended to move [consumers] from awareness to conversion,” says Joe Shields, Wyeth’s product director for Enbrel. “It’s difficult for us to know when a banner ad has led to a sale, but we’re learning that the combination of certain messages in certain places over a period of time has given us the results we’re looking for.”

Enbrel looks at hard data such as cost per lead, and optimizes its creative as part of the media buy. It also digs deeper into engagement with click-density studies that determine how long it takes users to fill out registration forms. However, the measures of the campaign’s success go beyond the initial purchase to include areas such as patient retention and compliance. Thus, online activity metrics are only a part of the story. “We can measure in a broad way. We do see the results from the ads, but it’s an iterative process,” says Barri Hollander, vp/director of marketing at Modem Media, San Francisco.

By contrast, those metrics were the primary source of information in weighing the success of the Xbox Perfect Dark Zero campaign, which used e-mail, mobile and the Internet to pique the interest of gamers during the 360 console launch last November.

“The Xbox audience spends a huge amount of time on the Web, reading reviews, on message boards,” says Rikki Khanna, account director for Xbox at AKQA, San Francisco. “We created microsites as experiential destinations for game players . . . and we carefully measured the response to those destinations.”

AKQA unleashed a sequence of events that mirrored the action of the game: Traditional ads directed users to a microsite where they were asked to input their cell phone number and the name and e-mail address of a friend. The friend, in turn, received an e-mail with a link to a video, in which that person becomes part of the game. The user then received a phone call from one of the game’s characters.

Each of those steps is built around a measurable component. Of particular interest to Xbox marketers was the “pass-along” rate, or percentage of people on the site who referred a friend (about 30 percent). The work was also judged on traditional metrics including unique site visitors and number of wallpaper downloads. To gauge the campaign’s reach and efficiency, AKQA multiplied the average time spent on the site by the size of audience, and compared the cost with an equivalent TV buy (in this case, about $2.9 million).

Khanna says that while this metric was specific to the Xbox campaign, researchers could indeed derive a standard “unit” of engagement. Here’s how: Start by measuring the time spent with a brand and level of involvement in a campaign, and factor in items such as volume of e-mail signups, downloads, Web site and call center traffic. Multiply this weighted engagement metric with the size of the audience, and add the total cost of the campaign to determine the efficiency of the engagement. “Although this wouldn’t be affordable for every campaign, a test study might throw some light on the engagement debate,” she says.

Not everyone thinks it’s a good idea to have a standard of measurement. With rich media ads carrying many different objectives, says Atlas’ Safra, a unified measurement would be “apples to oranges.”

Furthermore, the rules that govern online ad measurement are changing along with technology. AJAX, a new system of Web applications, for example, will soon create “live” content on portals like Yahoo and MSN. Customized pages will refresh automatically, along with the ads, requiring a new formula for measuring unique visits and ad exposure.

“The Web is a dynamic environment,” says David L. Smith, CEO of Mediasmith. “When technology for the building blocks changes, it affects your ability to gain a proper measurement. You have to stay on top of it.”

Safra agrees. “There are a lot of derivative technologies that are changing the rules of measurement,” he points out. “In the next three months, you will be able to do things you could never do before.”



Michael Applebaum is a former senior features editor at Brandweek.

Optimal Effect



These days, Napster is a small player in an online music business dominated by Apple. But had it not been for a successful effort to optimize the ad campaign behind the launch of its paid subscription service in 2004, the brand might not be around at all. And its ability to reverse costs per acquisition (CPA) through optimization provides lessons other companies can use in their own marketing campaigns.

In 2004, an ad campaign by Venables Bell, San Francisco, sought to introduce Napster 2.0 as a legal source for downloadable music while retaining the maverick spirit of the original brand. Teaser ads heralding Napster’s return were filtered through music tastemaker sites, and by the time Napster re-launched, the campaign had rolled into mainstream sites on ESPN.com, Cnet and Rolling Stone.com.

By mid 2004, though, Napster’s cost per acquisition for a trial user was way too high. The onus then shifted to its media buying agency, Mediasmith, San Francisco, to optimize the campaign and get that number down within striking range of $25—or seriously hamper the company’s ability to stay in business.

“We were constantly monitoring which sites and banners worked on a click or view-conversion basis,” recalls Derek Leedy, account director at Mediasmith.

The messaging was split into two camps: one group of ads that focused on specific music genres; and another that touted the benefits of the service, such as the large number of available tracks. According to Leedy’s internal optimization data, the strategy was not working to full effect. “Though we paired the genre ads with contextual placements, placing rap banners on rap sites, for example, they never did as well as certain evergreen messages,” he says. The reason, supported by outside focus group research: “Though people may visit a site or section of a site that concentrates on a specific music genre, most everyone considers themselves to be eclectic in their music taste.”

The solution was to load up on the generic messaging and expand the campaign’s reach by buying into performance networks such as Ad.com, Burst Media and Blue Lithium. “We introduced many new concepts into the rotation to avoid burnout,” notes Leedy. “We soon learned that we needed even more creative to maintain a fresh and compelling presence.” And the result? “The volume of activity increased dramatically, lowering the CPA by over 300 percent” in less than three months.

Results like these demonstrate the statistical power of optimization. And, while there are limits to the insights that the technique can provide—optimization cannot, for example, explain why certain users behave the way the do, because it does not identify its users—it does give marketers a sense of what direction to take a campaign.—M.A.