Most marketers wouldn't think of launching TV and print ads without testing them first, but online ads typically have been introduced with little or no testing.
That's changed, however, as marketers are throwing more money at the Web. Online advertising accounted for 6.5% of the ad market in 2006, but is projected to rise to 7.2% this year, per TNS.
That growth has led marketers to test spots before they are released and online media firms to start offering testing as a value-add for Internet ad buys.
Online ad execs say the change has happened over the last year or so. "As budgets go up, people apply a bit more scrutiny," said Jeff Marshall, svp and managing director at Starcom.
In addition, the Internet Advertising Bureau, New York, is planning to release a study in the next month or so showing testing Internet ads can dramatically increase their performance, said Sheryl Draizen, svp and general manager at the IAB.
There are generally two ways to test online ads: use a select group of paid respondents (usually around 200); and, discreet launches that are aimed at a small portion of the Web-surfing audience. While agencies like OTX Research and Dynamic Logic, and online media firm MSN, use the former, Yahoo! and Google advocate the latter.
OTX Research, Los Angeles, started testing Internet ads about 18 months ago and now has a roster of clients that includes Kraft, Quaker, Unilever and Coca-Cola. The company uses groups of 200 that access ads via "secret" URLs and passwords. "If you are specifically questioning about how effective the ad is itself, then the pretest is best," said David Brandt, OTX managing director.
Like OTX, Dynamic Logic uses a small group of consumers (about 100) to test ads and then tries to determine whether exposure to the ads changed the respondents' intent to purchase the product. Ken Mallon, vp-product development at the New York-based subsidiary of WPP, said the increase in demand for online testing reflects a maturing market. "Historically, people questioned the value of online. A lot of the value was 'Did it work?'" said Mallon. "Now they understand it does work and we're moving into the stage of 'How can I do it better?'"
About six months ago, MSN started pretesting ads by e-mailing them to select consumers and gauging their responses. Stephen Kim, global research director at MSN, said it's important to test online ads separately from other media. "Things that might grab attention offline behave very differently in the online world," Kim said. "Even basic things like color backgrounds and fonts can make a big difference."
Another way to test online ads is discreet launches, an approach Yahoo! and Google advocate.
Yahoo! is touting ad-testing as one of the benefits of Panama, a new ad ranking system where placements are based on the ad's price and relevancy. Panama lets advertisers test up to 20 different creative executions. Then they can let Yahoo! automatically sub in the most effective ad. "The system will pick the winner," said John Kim, senior director-product marketing at Yahoo!, who added that success is mostly pegged to click-through rates. Another Yahoo! testing option is to geotarget by IP address so that, for instance, Chicago, sees different creative than the rest of the country.
Google, meanwhile, favors testing by delineating a handful of sites that are appropriate for the target demographic before going wider. Given Google's capacity to show video (and its YouTube acquisition), the company can also test TV ads this way.
Kim Malone director of AdSense for Google, said some advertisers are even testing Super Bowl ads with such site-targeting, though she declined to name names. Overall, she said advertisers are just becoming interested in testing Internet ads. "It's the beginning of a very new trend," she said. "We're starting to see more and more interest."