Mpire Corporation, an ad optimization technology company in Seattle, released the results of its AdXpose™ testing Thursday, which showed impression fraud and URL padding in standard run of network (RON) online ad buys. This is a widespread occurrence, says Marissa Gluck, Radar Research managing partner, and she calls it "the dirty little secret of the online ad industry that no one wants to talk about," on MediaPost Thursday. The July Mpire study showed that half of ad impressions and 95 percent of clicks in RON online ad buys were fraudulent.

Two types of campaigns were tested: a managed campaign with reputable sites, pre-approved to reduce fraud concerns, and a RON campaign.

The test showed that the fraudulent traffic was hidden behind layers of nested iFrames. These iFrames pull ad content from other sources that hide URLs and other data. All of this number skew makes for unreliable data for building budgets. According to Gluck, if suspected ad networks want to be taken seriously by advertisers, they have to eliminate click fraud.

The iFrames complicate efforts to track campaigns, hide fraudulent traffic and leave advertisers blind to in-view data, the study continues. The traffic behavior of many RON buys includes fraudulent and known botted sites. Not all marketplace or exchange traffic is bad, Gluck writes, but it does include nefarious inventory (these problem sites) that for some reason ad networks could, but do not, block.

Test campaign findings also include URL padding - where a broad range of URLs are proposed for an ad, but the majority of the buy are delivered via only very few of the sites. In this test over 98 percent of traffic was delivered via fewer than two percent of the URLs.

The performance obligations seem to satisfy performance obligations, says PRNewsWire, and for this reason the fraud remains unchecked. However, by leveraging this new data, advertisers can put on the pressure to improve performance and demand charge-backs for campaign shortcomings.