Targeting On The Web: Are We Fooling Ourselves?
The concept of targeting has been huge on the web, at least conceptually. What's happening with the companies who have targeting technology? Some are dead and others are dying. Why is it that these companies who have based their stock in trade on profiling have yet to succeed?
First, some definitions. There are two types of targeting on the web. The first is the same as traditional target audience definitions for offline media. This involves analysis of product usage information and the development of a strategic target audience based on sociographic habits. For example, if we analyzed the patterns of the consumers of an energy bar, and found that they also backpacked, jogged, and rode bicycles to work, we might surmise that they would be candidates for cross training athletic footwear. We could also develop CPM's against various media alternatives based on the sociographic habits or definition. The other kind of targeting on the web is often called profiling. Profiling on the web has to do with identifying consumer needs or desires, either through questions asked at registration or following their surfing patterns and serving up information or advertising accordingly.
One of the issues not often discussed has to do with the limitations of profiling. First, when we establish a target audience in traditional media, it is just that, a target. Impressions are not limited to the target audience but essentially skew towards them. For example, if we place a national cable buy against Males 18-34 for the energy bar, we might buy networks like MTV, ESPN, ESPN2, Comedy Central and VH-1. Many of these viewers are not 18-34 males, but still watch one of these networks. Targeting in this manner increases your odds of sale by reaching those consumers with the highest propensity to consume your product in an efficient manner. It also reaches many other consumers outside of your target audience who have similar lifestyles to the target but do not fit the technical definition. In many cases, these consumers outside of the target can amount to 30-50% of all consumption. If they were somehow excluded, it might be very hard for the product to reach its volumetric goals. We gladly pay a premium for this targeting, but it is not significant when you take into account the bonus impressions outside of the target.
Profiling has been heralded as the thing to do. You have no "waste" and only reach those in the profile. The limitations of this include: a) leaving out those not in the profile but who could be consumers, b) the assumption that the profile guarantees appropriate consumers when it may be dated, or not reflective of what the consumer will actually do when presented with a message, and c) the premium involved with profiling, which is significant. The bottom line is that with profiling, as elegant as it sounds, is hard to pay out. In fact, I could argue that I should get a discount with profiled impressions rather than pay a premium. That would not make sense for the companies who invested in profiling though.
While Mediasmith continues to test new profiling alternatives, less than 50% of all profiling experiments have paid out.
This makes us wonder even more about the value of profiling. For, if the major sites and networks are selling a lot of inventory on a profiled basis, "skimming the cream", what are they selling us when they sell ROS? Is it truly "run of site" and does it represent a reasonable chance to reach everyone in his or her audience? Or, is it really remnant inventory, which should be discounted even more, that being the case. And, if the sites tried an experiment running all advertisers in a non-profiled mode, might every advertiser increase their ROI? The profiled advertisers could reach folks never exposed to their message and the other advertisers could get a true ROS. It might be win-win for everybody involved, including the site or network that could get higher renewal rates from more successful advertisers.
David L. Smith is President of Mediasmith, Inc., the Integrated Solutions Media Agency based in San Francisco and New York. He can be reached at firstname.lastname@example.org.