• by Masha Geller , October 23, 2001

Yesterday’s discussion of robots and spiders, those pesky things that surf the web and inflate web traffic figures, seems to have hit a nerve with some of our readers.

Judging by the feedback I’ve received so far, everyone supports the IAB’s decision to put together a list of known robots and spiders, but that’s not nearly enough to solve the problem of discrepancies that occur when the numbers reported by sites’ ad management systems do not equal those reported by an advertiser’s third-party ad serving system.

According to a recent white paper published by Mediaplex, a third-party ad server, “This issue will not be solved until uniform methodologies and counting standardization are adopted throughout the industry.”

Unfortunately, I’m more willing to bet on the Yankees winning another World Series than on the probability of the online ad industry coming to a consensus about this.

Here’s the heart of the dilemma: as Mediaplex explains, most publishers count an impression when the HTML page that contains the tags (URL directions) for the ad is sent to the user. This method equates a page view request with an ad impression – big mistake. This count occurs relatively early in the chain that results in the delivery of an ad – before the publisher page is even delivered back to the user and the ad tag is read by the browser.

Third-party ad servers, Mediaplex included, count impressions later in the process, when their server has received an ad request, selected the ad and forwarded a redirect URL to the user’s browser. This tally comes significantly later in the process than publisher counts and is – in most advertisers’ opinion - more accurate than the publisher totals.

Although most website publishers aren’t happy about it, the idea of browser-side measurement is definitely catching on with advertisers. Companies such as RedSheriff, WebSideStory’s HitBox and Webtrends are good sources of information if you’re interested in actually getting what you pay for.