comScore Launches Audience Advantage(TM) Predictive Modeling Platform to Help Publishers Optimize Audiences for Digital Ad Campaigns

Company: comScore, Inc.
Published: 23 February, 2010
comScore, a leader in measuring the digital world, today announced the launch of comScore Audience Advantage, a digital audience optimization platform that enables publishers to provide advertisers and media planners with the ability to accurately reach their most valuable audiences. Using anonymized data from its proprietary research panel of opt-in Internet users that is aggregated into a variety of important consumer segments, comScore builds sophisticated predictive algorithms that can be used by publishers to determine which of their site visitors should receive a particular ad. The predictive algorithms use a robust array of descriptive consumer variables, including demographics, search activity, site visitation, online and offline transactional behavior. Early tests of the Audience Advantage platform have shown lifts in reaching the desired audience segment of 3-5x, representing a substantial gain in campaign efficiency.

"The comScore Audience Advantage platform represents a revolution in the media planning and buying process by enabling publishers to use the lookalike models to reach the audience that mirrors the advertiser's best prospects. This allows them to deliver a campaign that is on point in both message and audience," said Erin Hunter, comScore executive vice president. "Audience Advantage offers significant value creation for the online advertising ecosystem that benefits all stakeholders. Publishers will be able to demonstrate higher value in the campaigns they deliver, thereby justifying higher CPMs, while advertisers will be able to minimize wasted impressions and reduce the total cost of advertising outlays. It is truly a win-win for the digital advertising industry."

How comScore Audience Advantage Works comScore Audience Advantage uses proprietary anonymous audience scoring algorithms to determine optimal audiences for advertisers using sophisticated and predictive variables available as inputs. These variables are derived from a multitude of observable behaviors in the digital environment as measured through the comScore opt-in research panel, including site visitation, search activity, video views, advertising engagements, online purchases or any other number of behaviors that may be relevant to purchaser particular advertiser. The Audience Advantage platform does not rely on cookies to construct its predictive algorithms. Rather, they are built using the actual observed behavior and characteristics of the opt-in comScore panelists.

Importantly, the Audience Advantage platform can also leverage third-party databases that are integrated with the comScore research panel to provide anonymous offline purchase behavior as another critical and highly predictive input to these algorithms.

Using all available behaviors as inputs, Audience Advantage determines a predictive score for various audience segments. Participating publishers then use this score to anonymously identify and deliver to advertisers "lookalike" audiences from the behavior they observe among their own site visitors. Campaigns can be planned at a granular level, for example, against only 5 or 10 percent of consumers most likely to purchase a particular brand with the objective of reinforcing loyalty, or they can be planned against a much larger segment of consumers who are dual brand buyers with the objective of inducing switching to the advertiser's brands.