Last month in the Wall Street Journal, WPP's Sir Martin Sorrell called econometrics the "Holy Grail" of advertising.
According to the Columbia University Press Encyclopedia, econometrics is a "technique of economic analysis that expresses economic theory in terms of mathematical relationships and then tests it empirically through statistical research. Econometrics attempts to develop accurate economic forecasting and to make possible successful policy planning."
The feeling among a growing group of theoreticians is that use of this method can replace older vehicle- and audience-based advertising metrics.
Once again, it sounds like the measurability of the web has moved the traditional media world to action. After all, the web has proven itself in the area of ROI, with ongoing CPW ("cost-per-whatever" the client is trying to measure) available to many clients using third party ad servers in a direct response model. And, the use of tools like Dynamic Logic, Insight Express and XMOS studies have provided payout evidence in the branding area.
It is clear that questions regarding the use of econometrics remain for many advertisers who have not either used the tools or are unsure about the long-term benefits to their advertising.
In this piece, I'll try to provide some useful information about econometrics and how advertisers might be able to use them.
Will econometrics replace tools like reach and frequency for most advertisers? Probably not in the short term, but it does have long-term potential. In fact, according to Abbey Management Services, "because professional econometric models do measure long-term effects, they typically yield estimates of the contribution of advertising several times greater than the often inadequate analyses which have been published by advertising agencies."
This is understandable, as agencies are typically limited in their payout or ROI analyses to using short-term metrics. Clients want to know what happened this week, this month or this quarter. The long-term metrics are more elusive, requiring more faith and significant investment to even establish a track record that could be applied to econometrics.
It takes a brave and far-seeing brand manager to start the process given the close eye on immediate success in an era where quarterly earnings seem to mean everything.
Where do you start if you are interested? Well, you need a track record and the willingness to devote at least several hundred thousand dollars a year to the answers. This alone eliminates many advertisers. But for those with big budgets, the cost can actually be quite reasonable. Most experts agree that you need several years' worth of marketing and sales data to even start the analysis, and they believe that the inputs should not be limited to advertising.
In contrast, these experts believe the inputs should include all marketing programs including sales promotions, retail incentive programs, product placements, et cetera. Companies like MMA, IFX and Hudson River Group are prepared to take your data (and your money) and provide you with modeling that can be the basis for future planning. You could also hire your own staff to create the modeling in-house. These inputs then can be provided to the media planning systems like Telmar, which ports the information into their ROI tool called MediaPlannerROI.
If you have the faith and the budget to do the above, you can then plan media on a whole different level than before. But, if you have not really used media mix and have experience in media you want to model, don't count on econometrics to tell you about media you have not run. After all, the modeling is based on history and you are limited by the history of the brand you are tracking with the model. And, you need to keep in mind that you need to do this all over again for every brand or line extension.
So, you cannot really make a lot of econometric use for new media or new products, the lifeblood of many companies.
The concept of econometrics is great in theory. It puts a different wrapper on something that traditional media has been trying to do for over 40 years. FCB did it in the 1970s and '80s with Stage 1 and Stage 2 (methedologies FCB had to computerize some media decision making processes); General Foods did it in the '60s by putting effectiveness weights on various media types and TV dayparts.
However, the fact remains that this Holy Grail is elusive. To this date, more money is spent on copy testing than media effectiveness testing. Net/net, there is probably not enough data or budget for any but the biggest spenders on traditional media to make this work right away. It will take time for the inter-media comparison data of econometrics to solidify.
Once again, the web and its trackability have a big head start on traditional media.
And, if you need data on the viability of a concept or product quickly and inexpensively, nothing beats the web for learning.
David L. Smith is CEO of Mediasmith, Inc., a Media Agency and Media Consultancy based in San Francisco.
