Ad Money

At firstStreetOnline.com, where a retailing strategy based on innovative consumer products goes hand in hand with innovative marketing, director of Internet marketing Daniel Yonts takes nothing for granted when exploring new ways to promote his company’s offerings.

Targeting baby boomers with things like easy-grip garden tools and turntables that transform vinyl records into CDs, Yonts uses strategies including developing buzz-creating special-content sites like StuffTalk.com that link to firstStreetOnline.com and running ads on political and other types of blogs. And to expand firstStreet’s marketing reach even further, this year he’s planning to launch homegrown software designed to quickly produce blogs and social networking sites to create new marketing venues for unusual products.

Putting together a comprehensive online marketing strategy is one thing. Building a strategy for measuring the impact of all that marketing is quite another. Although there is a growing number of online marketing measurement products and services, finding the right mix and putting them to effective use is a challenge, Yonts and other retailers say. “It takes time to go through marketing measurement data and make decisions,” Yonts adds.

Indeed, the very nature of online marketing—and the way it continues to quickly evolve with new forms of interactive media—call for new measurement tools and tactics that are challenging to use, experts say.

“These are complicated tools trying to solve complicated problems,” says Andy Fisher, director of analytics for marketing services firm Avenue A | Razorfish, which specializes in managing Internet marketing strategies.

Moreover, effective measurement of the impact of multiple forms of online marketing requires a complex mix of applications, he adds. Getting a handle on the impact on customer traffic and sales from a mixture of search marketing, blog ads and video ads placed in social networks, for example, can require a mix of measurement tools that look at traffic and customer activity throughout the web as well as on a marketer’s own site.

Most of Avenue A’s marketing clients measure the impact of marketing efforts using a variety of tools—server-based monitoring of web traffic, consumer panel surveys, web site analytics, content management tools, landing page optimization and multi-variable testing. “Every tool is managing something different, so the information is never consistent,” Fisher says. “So it’s important to look at it altogether as much as possible.”

Deciding which mix of tools to use or which to make a lead source of information is still another challenge, experts say. At industry marketing conferences, attendees have been packing sessions covering ways to measure the impact of online advertising on Internet traffic and sales, says Brad Bortner, an analyst who specializes in online advertising at research and advisory firm Forrester Research Inc.

“The cool thing about the Internet is that it’s inherently measurable,” he says. “But it’s still not quite clear what’s the best approach to measure stuff.”

Measuring the measurers

The measuring services themselves, meanwhile, are going through steps with marketing industry groups to see how they measure up as reliable sources of information. Both comScore Inc. and The Nielsen Co.’s Nielsen Online, two leading providers of data on the effect of online marketing on Internet traffic and consumer activity, are undergoing reviews by advertising industry groups the Media Rating Council and the Advertising Research Foundation.

The goal of the reviews is to show more transparency in how comScore and Nielsen gather and present data on consumer response to online marketing, and, in turn, provide a better understanding of why their data on Internet traffic may differ. Even if their data sets remain inconsistent, marketers need to know why they’re different in order to make the data sets more useful, experts say.

Nielsen reported, for example, that the number of Internet searches grew 39% between 2006 and 2007, while comScore reported growth of 11%. And while Nielsen has estimated the number of Facebook unique visitors at 20 million, comScore has put the number at 30 million. Part of the difference may lie in the way each company recruits panel members as well as in the way they weight their panels to reflect demographic makeup of targeted populations of surveys, Bortner says.

Nielsen randomly selects a large part of its 500,000-person global consumer panel, which includes 250,000 in the U.S., according to Manish Bhatia, president of global services and U.S. sales at Nielsen Online.

ComScore, though it doesn’t comment on its use of random selection, uses a complex system of incentives to attract a broad mix of panelists, including people in work, home and university environments, says senior analyst Andrew Lipsman. ComScore works with a global panel of 2 million, including 1 million in the U.S., though any one of its survey services may use a subset of the full panel, such as the 125,000 used by comScore Media Metrics, he adds.

The reviews for each company are expected to take another year or so to complete, though each company has openly discussed data consistency issues while making improvements and additions to their marketing measurement services. The overall effort, Bortner says, is part of an ongoing evolution of media measurement. Thirty years ago, the state-of-the-art in media audience measurement was face-to-face interviews conducted by researchers in consumers’ homes. While such interviews still are widely conducted in Europe, they eventually were replaced in the U.S. by telephone surveys that proved to be faster and cheaper.

New challenges

Now, with the Internet making it easier to gather data from groups of consumers, online consumer panels are emerging as the new research standard, but with a new set of challenges. While there are more types of media to measure, including ads placed in interactive rich media, online games and social networks, marketers are demanding more complete and accurate data.

IRCE 2008 Conference Multi-Media CD-ROM

Meanwhile, with the quick pace of evolution in online marketing, it’s helpful to have different approaches to measuring the effect on Internet traffic and consumer activity, Avenue A’s Fisher says. “I see the difference in Nielsen and comScore as a good thing,” he explains. “We want innovation. We want all the measurement companies out there to come up with better ways of measuring things.”

One issue Internet measurement firms are trying to reconcile is server-based data approaches—which use software to automatically track consumer Internet traffic—with surveys that ask consumer panels how they use the web. “Data our clients are getting from their own web logs or server data tends to overstate traffic 30% to 50%,” Nielsen’s Bhatia says. One way Nielsen is dealing with this situation is by tagging its clients’ sites with tracking software, then automatically refreshing web pages every few seconds to capture traffic data, and combining results with data from panel surveys.

Nielsen Online also has expanded the scope of data it collects from meters placed on the personal computers of participants in consumer panels. Instead of tracking only when panel members turn web browsers on or off, it now tracks all personal computer activity to reveal whether a panel member is actually on the web when her browser is open or perhaps working in a Word document.

Nielsen also now is tracking time spent activating online videos as well as online games, including exposure to any ads within video or game content, Bhatia says.

ComScore, meanwhile, also is expanding its services while continuing to work on innovations, Lipsman says. The company’s new Ad Metrix Publisher service, launched last fall, shows where online display ads are viewed across the Internet and demographic characteristics of consumers who view them.

“We’ll see more innovation this year,” Lipsman says. Among the targets are widgets, tiny software applications that entice consumers to visit a particular web site. “We’re expanding our definition of widgets and including more of them in response measurement,” Lipsman says.

Nielsen Online and comScore’s primary clients are large marketers who have the staffs to manage their high volume of data and can afford their fees. The cost to use Nielsen Online, for example, starts out at about $25,000 a year but averages from $50,000 to $100,000 a year for most clients, Bahtia says.

Marketing alternatives

But marketers have alternatives. Compete Inc., for example, which works with a panel of 2 million U.S. consumers, offers some of its information for free. Hitwise provides Internet traffic data gathered from Internet service providers on a base of 10 million U.S. Internet users and more than 1 million web sites, enabling online retailers to compare visitor volume across sites in a given market.

And while Nielsen Online monitors public comments on brands across blogs and Internet message boards through the BuzzMetrics service it acquired last year, it faces competition from a new player in the market, BuzzLogic, which shows marketers the extent of influence bloggers have in particular markets and product areas. By showing the links between multiple blogs, BuzzLogic shows which bloggers are tied into networks of discussions on products a retailer might want to advertise.

At firstStreet, which has limited resources for analyzing marketing data, Yonts relies on relatively small sets of data from Quantcast and Alexa to view Internet traffic, compare how firstStreet’s web sites are garnering post-marketing traffic compared with other sites, and view how well firstStreet is attracting its targeted demographic groups.

But Yonts says he looks forward to having a broader marketing staff capable of handling the large volume of data available from firms like Nielsen and comScore.

“Nielsen and Comcast provide a more holistic approach to gathering consumer data,” Yonts says. “Moving forward that will be even more significant because of online video, consumer generated content, mobile, etc.”

Holistic approach

This more holistic approach will support further developments in marketing strategies, Yonts adds. And that will create even more challenges, he figures. As data alerts retailers to hot products and categories, retailers may need to develop new venues for marketing to those interests and, in turn, measuring the impact of those new venues, he says.

“One problem we can run into is that as we learn of markets we need to focus on or new products in demand by consumers, we may realize there is no good place to advertise because no one is really reaching those interests,” Yonts says. “So we may create that advertising place ourselves, perhaps in a new blog.” And since that blog might engage visitors with video and consumer generated content, the retailer would need a way to measure its effect on traffic and sales.

“What I expect to do is get data showing what consumers are looking for and where they’re going, broken down by demographics,” Yonts says. “Five or 10 years from now retailers will be basing their entire merchandising processes on this data—what products to sell or not sell. This will have profound implications for retail operations. By incorporating this data into retail operations sooner, it will affect merchandising before marketing. When we see a demand, we’ll find a product based on that demand instead of coming up with a product first and then trying to market it.”

paul@verticalwebmedia.com