When a University of Southern California and Indiana University study suggested 15% of Twitter accounts — some 48 million — are bots rather than people, the conversation about understanding identity in social media marketing naturally came up.
The problem with a vague concept of identity in marketing, of course, is not just the inability to “know” a customer, but basic problems in knowing the traffic.
After all, it’s proved very difficult for social media platforms to establish trust for their own analytics. Facebook was found to have vastly overestimated average viewing time for video ads on its platform for two years, and YouTube is scheduled to undergo an ad metric auditing, in the middle of an ad placement controversy.
These pervasive identity issues, according to Brian Baumgart, CEO at Conversion Logic, can be attributed to not only bots, as in the case with Twitter, but cookie pools.
“Cookie pools inherently have a varied shelf-life ranging from very short to somewhat longer, and as a result there is a potentially large gap both between vendors and the volume/quality of the audience attributes they offer,” says Baumgart.
Of course, attribution is certainly one of the best approaches for accurately ascribing credit to a platform’s analytics, but fraud has and will affect all platforms and publishers to some extent. So, how can a brand or marketer avoid driving time and money into fraudulent traffic?
Well, the most obvious way is to use fraud vendors to identify and prevent fraud. And while this tactic can prove effective, Baumgart explains brands and marketers are also using other methods, such as experiential learning or an attribution vendor.
An example of experiential learning, or testing of personalization experiences, is holistic user graphing; which matches two entities in a data set and determines if they refer to the same real-world object.
“One may provide details on social networks, for instance, while another focuses on particular devices,” says Baumgart. “This ecosystem further by-passes traditional, one-platform attribution by relying on the most relevant data from the most trustworthy sources.”
Marketers and brands have also made the decision to make vendor’s data available to their attribution vendor, in order to ensure that “measurements of data are valid.”
Surely these tactics will not eliminate fraud. However, if brands and marketers are not paying attention to identity, they can’t establish “a unified view of the customer” or traffic.