Mixpo Unveils VideoVerify To Sniff Out Suspicious Campaign Behavior
Mixpo, a Seattle-based video advertising provider, is expected to today announce VideoVerify, a tool meant to identify suspect domain and network activity in a campaign. The company has given RTM Daily exclusive details on the new tool.
VideoVerify uses an algorithm Mixpo developed after examining seven year of data. The idea behind analyzing that much data was to establish a baseline to know how a certain type of ad should normally perform in a certain space. As a result, VideoVerify can flag campaign behavior as "normal," "suspect," or "highly suspect" when compared to that baseline. The tool uses Big Data from Amazon Redshift, Tableau's data-visualization software, and R, the open source statistical analysis software.
Mixpo is part of the OpenVV group, but it doesn't think OpenVV does enough. Emily Carrion, Mixpo's recently appointed director of communications, said, "While we fully support OpenVV, and we think it does a great job of bringing awareness around the issue of off-page inventory and standard space measurement,…viewability is just one metric." She said that in some of the instances where ads are viewable, "it might be on a poor site, or poor quality, or maybe even fraudulent."
VideoVerify is for Mixpo clients only. Part of the reason for keeping it to themselves is to have something to make their company unique. Walter Harp, Mixpo's VP of product marketing, claimed that another reason to not open source VideoVerify is to essentially keep it out of the "bad guys" hands, as both he and Carrion believe that the "bad guys" are becoming more sophisticated with their methods. Harp estimated that 50% of a video campaign is lost when combining the issues of viewability and fraudulent inventory.
The company noted that they are able to integrate first- and third-party data from providers like Nielsen OCR and comScore in order to break down campaigns more precisely. The algorithm that supports VideoVerify uses 30 metrics to determine whether or not the campaign behavior is normal, suspect, or highly suspect. Among those metrics are viewability, view rate, interaction rate, and click-through rate.
The tool is also meant to allow advertisers to act "on the fly," or in real-time, to suspicious behavior. "Combating fraudulent inventory is a major industry challenge," Brian Cohee, Mixpo's SVP of data sciences and core systems, said in a release. He claimed that once VideoVerify flags something, advertisers can "quickly address potentially fraudulent impressions."