They say that a click is a click, but according to a recent Mpire report, using AdXpose, all clicks aren't equal. Especially clicks generated from run-of-network (RON) online advertising buys. According to the report up to 95% of these clicks and up to 50% of the ad impressions are generated from fraudulent sources.
by Kristina Knight
AdXpose is Mpire's campaign analysis, verification and optimization technology. Researchers found that just over 50% of the ad impressions delivered and 95% of the clicks generated were potentially fraudulent. This traffic was hidden beneath layers of I-Frames, which are ad units that pull advertising content from another source. Researchers also found a large amount of URL padding. URL padding happens when a range of URLs are made available for an ad but only a few of the URLs are actually used.
Other interesting findings:
• Ad networks need to improve on 'above the fold' ad placement
• Mpire tested three campaigns with 11 campaign buys on 9 ad networks/exchanges
• More than 20 million impressions were served during the test
"Click fraud and impression fraud is far more pervasive than the industry has been willing to admit, yet thus far the industry has taken a laissez-faire approach to policing downstream traffic providers, leading to material waste in campaign budgets," said Marissa Gluck, principal analyst at Radar Research, the company that verified the research. "By delving deeper into site-level data, advertisers and agencies can get a better understanding of the impact of fraud on campaign ROI, and can thus take steps to focus their ad spend on the sites, networks and exchanges that deliver the most legitimate impressions and clicks."
Along with this report, comes news that ClickForensics has identified a new botnet called "Bahama Botnet". The botnet distributes malware, disguising itself as a high-quality source of advertising traffic. Just another reason for pay per click marketers to research advertising networks and pay close attention to how a campaign is performing from beginning to end.
No comments:
Post a Comment