Buyer Targeting System Brings Accuracy and Accountability to TV Advertising
NEW YORK – PrecisionDemand, the leader in data-driven TV ad targeting, has been awarded a patent for its proprietary automated ad targeting technology.
“Our Seattle-based team of scientists has developed and fine-tuned a system that has successfully produced measurable results for all of our advertiser clients, whether the goal was increased retail sales, spikes in web traffic, call-ins or other unique metrics.”
The patent makes PrecisionDemand the only media company able to effectively help advertisers leverage Buyer Targeting — the practice of mining set top box data, as well as advertiser first-party customer data, to target TV ads to media with the highest probability of buyer.
“Receiving this patent validates PrecisionDemand’s ability to bring accuracy and accountability to TV advertising,” said Jon Mandel, President and CEO, PrecisionDemand. “Our Seattle-based team of scientists has developed and fine-tuned a system that has successfully produced measurable results for all of our advertiser clients, whether the goal was increased retail sales, spikes in web traffic, call-ins or other unique metrics.”
The patent (US 8,627,359) covers the high-dimensional demographic vector matching technology that PrecisionDemand uses as one of several methods in its set top box TV ad targeting system. The basic method consists of four steps:
- Creating a buyer high dimensional profile
- Creating a media high dimensional profile
- Calculating vector similarity
- Using the similarity score for targeting
“Directly estimating future buyer probabilities based on historical counts of buyers or view-minutes on media runs into a matching density problem – the probability of an individual person being a buyer of a product and also of being detected in the TV viewing population can be low, resulting in sparse data in TV media in real-world conditions,” said Brendan Kitts, Chief Scientist at PrecisionDemand.
“We solve this by enriching individuals with over one thousand demographic elements, and then calculating the vector match of buyers with media. Because all persons viewing the media and purchasing the product are used, the algorithm has excellent scaling and generalization characteristics. We use several ensemble estimators in Scoring Service system, but this has proven to be a key method necessary for achieving ROI in real-world campaigns,” Kitts said.
ROI results from using the technique were published last year at IEEE ICDM 2013 and are available on the company’s website.
The full text of the patent, complete with examples of vectors and matching results, is available on PrecisionDemand.com.
News of the patent first appeared in MediaPost.
About PrecisionDemand
PrecisionDemand helps TV buyers deliver quantifiable ROI for their marketer clients by combining the scale of television with the precision of online targeting and analytics. PrecisionDemand uses marketers’ own first-party customer data to understand their customers and identify the most cost-effective and highly-targeted media opportunities. Through the combination of this data and PrecisionDemand’s patented technology, marketers can now go beyond traditional measures of reach to deliver against customized and specific metrics of success. Headquartered in Seattle and New York, the PrecisionDemand team includes industry leading experts in media, technology, and data analysis.
