Machine Dreams

By Published on .

Most Popular

Call it science, or even a desperate preparation for ad judgment day. The advertising world is in a tizzy over econometrics, the latest high-stakes attempt by agencies and media-buying companies to outdo one another and win big with increasingly sophisticated clients.

As advertising has become a huge, complex, and increasingly fragmented business, accountability has emerged as a key determinant in whether agencies win accounts and keep clients. Econometrics, also known as "marketing mix modeling," is a way of building statistical models to measure and predict sales performance. Long used in industries such as travel and finance, it is now becoming part of the advertising landscape, fueled by advances in data gathering and responsiveness research and an increasing demand by marketers for quantifiable results. The magazine industry is launching similar, though less sophisticated, initiatives into accountability. (See sidebar.)

"Media has become such a big part of the overall [marketing] investment that major marketers are increasingly looking for accountability," says Daryl Simm, who last year was lured from Procter & Gamble's top media spot to become president and CEO of Omnicom's media operations, known as OMD Worldwide. "Econometrics is one credible and respected way to explain return on investment." Simm was introduced to the approach while at P&G, which, like other consumer packaged goods companies, has used scanner information to track consumer product sales for about a decade. He is widely credited with helping to shake up agencies' old-fashioned approach to media buying and planning by introducing a more result-focused client mentality. But until now, ad agencies have been generally slow to respond to the effectiveness challenge.

One jolt, many say, was delivered in 1997, when Carat-one of the world's largest media specialists-acquired Connecticut-based Media Marketing Assessment (MMA), a leader in measuring and modeling the impact of advertising and other marketing elements on sales. It was MMA's research for major clients like Kraft that established marketing mix analysis as an important tool for improving return on investment in the packaged goods industry. The London-based company, which says it spends $8 million on research annually, is championing a scientific approach to media that could give it a real competitive edge in the less adventurous American market. The MMA acquisition, many say, has led to a rethinking of priorities at the top agency levels here, and the recognition that media planning and spending need more scrutiny.

"Accountability is the final future of the media and advertising business," says David Verklin, CEO of Carat North America in New York. "Everybody wants to try to figure out where the money goes. Getting a return on investment makes that possible."

Verklin, like Simm, notes that econometrics is far more used and accepted abroad than it is in the United States, in part because the international media landscape is smaller and less fragmented, and therefore more easily measured. That explains, for instance, why WPP, the world's second-largest advertising company, has been able to integrate its media buying and planning capabilities abroad into one group, called MindShare, while it has struggled to do so in the United States.

And thanks to WPP's advanced research capabilities, London-based MindShare has access to a wide array of media techniques.

Econometric modeling makes up a big part of MindShare's research and consultancy efforts, according to Paul Dyson, chairman of its Advanced Technique Group Worldwide. Dyson, who just joined MindShare from sister research company Millward Brown, says that he hopes to recommend econometrics as "the ultimate answer" to clients who have the required input data and want to gauge maximum ad effectiveness.

This is not the first time in recent years that Europeans have influenced American media advertising. Most media planners and buyers here are still coming to grips with the optimizer, the hot and expensive software import that promised to make their work easier and more efficient by virtue of its ability to manipulate real-time audience measurement data. Available in Europe since the early '90s, optimizers, which are used to figure out how to best spread media dollars in order to maximize coverage, became the sword wielded by agencies in the 1997 battle for P&G's $1.2 billion consolidated TV buying and planning account.

Although it has since become part of major agencies' media tool chest, the optimizer is still seen by some as a fad that overemphasizes technology at the expense of people talent. And at a cost that sometimes runs into the hundreds of thousands of dollars, small- and medium-size agencies see the optimizer as a luxury.

On the other hand, cheaper optimizing choices do exist, and the optimizer hysteria has in fact helped revise outdated American media planning strategies which, until recently, relied on manual execution of thousands of media schedules and on 10-year-old reach-and-frequency models from Nielsen Media Research.

Are econometrics just another fad? Better availability of data, some say, is making the process of building the models easier and more reliable. After all, the accuracy and frequency of the data variables are the bread and butter of good models. That is why many media executives, for example, are grateful to Competitive Media Reporting for recently coming out with daily gross-rating-point data, which makes it possible to build frequent, even daily models.

In addition, since marketers themselves have started demanding results, a slew of organizations have rushed to commission econometrics effectiveness studies.

One promising effort is AdWorks 2, a joint project undertaken by two pioneers of marketing mix modeling, MMA and Information Resources, and sponsored by ABC, CBS, Fox, and NBC. AdWorks 2 examines actual sales and data for more than 1,500 brands in over 200 categories, and uses advanced models to determine incremental sales effects on TV advertising. Preliminary findings show, for example, that brands with at least 30 percent of their impressions in prime time had higher results than other brands. The final results are due this spring.

Another model is Consumer Mix Modelling, developed by Chicago-based Spectra, which claims to be the first to integrate consumer targeting and response analysis and link it to marketing execution. In Spectra's analysis of a yogurt brand, it concluded that TV advertising contributed about 5 percent of total sales. The company argues that the information was not sufficient for a brand manager to understand how key consumers responded to the brand, or how advertising could be fine-tuned in the future. Spectra claims that its model proved that while the yogurt company's advertising had an overall effect of 5 percent, the range across consumers went from almost zero in the "Downscale Urban" group to about 30 percent in the "Metro Elite" group-the company's core target. By plugging in relevant tracking data, such as Nielsen's, Spectra's model identifies the kind of shows these households watch, as well as what magazines they read.

Elsewhere, Utah-based MediaPlan is expanding its efforts to develop behavior-based optimizers, one of the first applications of its kind to use retail scanner data and other information to help advertisers target TV commercials better to consumers.

As tough as it may be to forecast the future, MediaPlan is confident that the fundamentals driving consumer response to marketing stimuli are universal, and justify building accurate econometric models. The stakes are high enough, demand is on the upswing, and so far, there seems to be no shortage of money.

In this article: