BtoB

Measuring online ads with fewer players, more data

By Published on .

Most Popular

The online advertising measurement industry is starting to look a bit like the telecom industry did before the breakup of Ma Bell. Within the past year, consolidation has winnowed the group down to only a handful of firms.

NetRatings acquired Jupiter Media Metrix’s ad measurement unit, AdRelevance. ComScore snapped up Jupiter’s Internet Audience Measurement service. CMRi bought Evaliant Media Resources. LookSmart absorbed technology and executives from Primary Knowledge. And DoubleClick bowed out of the measurement space, taking an equity position in Dynamic Logic, a New York research and online measurement firm.

While these developments have some industry watchers crying foul, others say it doesn’t matter because online advertising measurement is still in its infancy. What’s important, some say, is what you do with the information that’s out there, such as figuring out if your advertising investment is bearing fruit—something that’s easier said than done.

"There really isn’t a way to figure out return on investment," said Jim Spaeth, president of the Advertising Research Foundation in New York. "The measurement services that are out there just tell you the nature of the audience and the reach of your campaign—things that help you with media planning."

There seems to be a disconnect between what advertisers think measurement companies do and what they actually do, industry insiders say.

There are two types of measurement services. The first is census, or visitor, measurement, which tracks how many people visit a site each month. Web logs—records of every page request made to a Web server—also fall into this category. Web logs can be self-administered or handled by an outside provider such as NetIQ Corp.’s WebTrends.

The second type of measurement service is sample measurement, which is more quantitative because it follows specific groups of people, profiling them and watching what they do while they are online.

Unless you’re using both options, you may not be getting the big picture.

For example, the only way many advertisers can judge ROI is through their own independent auditing, regardless of the type or source of the metrics they use, one advertiser said.

"We look at reporting that gives us the actual number of clicks and impressions by keyword and then match it up with our own sales reports," said Eric Lituchy, CEO of Delightful Deliveries, an online gift basket and gourmet foods retailer. "Without both metrics you’re basically guessing."

Using the information, he’s been able to tweak his campaigns, getting rid of ads that boast plenty of clicks but no sales, while continuing those that have fewer click-throughs but high conversion rates, he said.

But not everyone has this type of capability. Delightful Deliveries creates unique URLs for every ad, so it knows how each is doing on its own. Companies that don’t have the resources needed for this type of execution need to mine their own Web logs for this information. Although it might seem time-intensive, this step is crucial—especially if the only other metrics come from the agency that’s also selling you your ad placement.

Perception problem

The other problem is perception. TV and print advertisers can’t always pinpoint exact ROI, but online advertisers expect this kind of data, and more.

"Advertisers forget [that] ROI is more than just conversion. You also want to concentrate on getting a brand across. Unaided brand awareness means an ad is effective," said Mike Ripka, an account executive with Millward Brown IntelliQuest, an Austin, Texas-based research and online ad measurement service. "But people think an ad is a failure unless a consumer sees that ad and clicks through to the site."

Today, even click-through isn’t enough for most advertisers. Now, conversion is key.

"It’s definitely not a buyer’s market anymore," said Molly Hilsop, VP-research services at Dynamic Logic. "But if the quality isn’t up to par, your clients won’t be your clients anymore. The quality of the data has to be there."

In this article: