CHICAGO: Media planners could benefit from a new tool designed to forecast earned media that employs three different "clusters" of variables, two leading industry practitioners have explained.

Writing in the February edition of Admap, Kate Sirkin, global research director at Starcom MediaVest Group, and Seraj Bharwani, chief analytics officer of Visible Measures, the campaign tracking organisation, argue that strategically planned earned media has become an integral part of the media mix.

They introduce a new tool, dubbed Contagion, which capitalises on large normative data sets and performance on tens of thousands of earned media campaigns.

"The collaboration was intended to establish a set of normative benchmarks or indices that would help the planners/buyers plan earned-media outcomes (within a range of earned-media performance) based on input scenarios within a media planner's domain of control," the authors write.

The new Contagion tool, Sirkin and Bharwani explain, is grounded in "three different clusters of variables having high statistical significance with respect to impact on EM outcomes".

These include variables related to the branded video creative (the content); variables characterising the environment (the publishers/networks) where the videos were syndicated; and, finally, parameters representing the online media investments (bought media placements) allocated for the individual campaigns in the data set.

Contagion programs this set of variables as controls and is currently web-accessible in beta mode to approximately 200 planners and buyers, who use it at one of several Publicis Groupe agencies.

"Contagion is most relevant for planning campaigns with reach expectations far beyond what the budget can help accomplish exclusively through bought media," Sirkin and Bharwani contend.

"Planning for branded video campaigns with earned-media expectations requires more inputs than simply media budget, target, reach and frequency.

"It requires input on the characterisation of the creative content and campaign association with event and/or celebrity, preference for specific publishers to be included on the plan and other traditional media support such as TV that might be part of the campaign."

To date, they continue, "earned-media outcomes for campaigns planned through Contagion have generally been within a plus/minus 30% of the range projected through the Contagion algorithm. We expect the performance to improve as the weights are trained with more in-market campaign data supplied by the planners currently using the system."

Data sourced from Warc