Duke Research Advantage

For at least 50 years, television advertisements have been ordered randomly within commercial breaks. But given the spread of digital video recorders (DVR) that allow users to blip-blip past the ads in fast-forward, it’s time to change this business model to maximize ads’ effectiveness, according to a new study.

The research, by Kenneth C. Wilbur of Duke University’s Fuqua School of Business and David Kempe of the University of Southern California’s Viterbi School of Engineering, was presented by Wilbur on June 23 at the Advertising Research Foundation Audience Measurement 4.0 conference in New York City.

“Think of two very different ads: the iconic Coca-Cola polar bears commercial, and a commercial for ‘natural male enhancement,’” said Wilbur. “The Coke ad will keep the audience glued to its screen, but the other ad will annoy some viewers, causing them to fast-forward or switch the channel. If the Coke ad is placed first during the commercial break, it still delivers most of the audience to the second ad. But if the Coke ad is placed second, it gets a significantly smaller audience.”

To account for these types of ad sequencing issues, the researchers have developed the Audience Value Maximization model. This new algorithm shows how to optimally select, order, and price ads based on a mathematical formula that considers advertisers’ willingness to pay and viewers’ propensity to switch channels during commercial breaks.

“Television networks historically have managed and sold advertising time. We propose a fundamental shift, with networks managing and selling the truly scarce resource in this industry: viewers’ attention,” said Wilbur. “The Audience Value Maximization Algorithm uses principles of consumer acceptance employed in Internet advertising, which increase viewer utility of ads and reduce ad avoidance.”

Under the current pricing structure, advertisers have limited incentive to retain viewers to watch subsequent commercials by other advertisers. For example, if an auto dealer features a screaming salesman in an advertisement, the dealer may increase the effectiveness of the ad by 20% while driving 10% of viewers to change the channel. The auto dealer comes out ahead, but he has reduced the audience remaining to watch subsequent commercials. The Audience Value Maximization Algorithm would charge the dealer for the 10% of the audience his ad repelled.

“If all advertisers share the same motivation to create ads that enhance sales while retaining the maximum number of viewers, the number of people avoiding television commercials will be reduced significantly,” said Kempe. “However, if advertisers are penalized or rewarded for their ads’ audience losses or gains, they would design ads to hold viewer attention to a greater degree, enhancing overall efficiency.”

The Audience Value Maximization Algorithm is a mathematical formula that utilizes market data to help television outlets select, price, and order advertisements to maximize audience value. Complete details of the new research are available for free download via the Social Science Research Network.