Knowing what you're buying can dispel confusion and increase ad investment. Revenue Science's chief revenue officer provides the questions to ask.

A note from Editor in Chief & Chief Content Officer Brad Berens: It's worth pointing out that at the time of publication Revenue Science is sponsoring our Behavioral Marketing coverage, but Jeff's byline is in no way a part of that sponsorship. You can find a more detailed explanation of our editorial policy here or contact me directly with questions.

Although it's still early, 2007 is already shaping up to be an incredible year for behavioral targeting (BT). A wide range of marketers are insisting that behavioral targeting is incorporated into their overall strategy and we're finding agencies, advertisers and publishers are getting more and more creative in their use of BT. Moreover, the industry has acknowledged behavioral targeting as a truly unique discipline.

The flip side of this situation is something we also often see in our space: Because behavioral targeting is so "hot," it seems hardly a week goes by without some (both established or newly minted alike) company issuing a press release touting its brand of BT. The truth, however, is that the actual offerings and the quality of those offerings vary greatly, and that's a potential landmine for everyone.

Confusion about behavioral targeting is nothing new. But today's confusion is perhaps more insidious, as aggressive marketing and merchandising programs make the differences in behavioral targeting harder to see and understand. Agencies and marketers have to be sure they are getting behavioral targeting that will meet their specific objectives, and that is only done by drilling down on the two biggest factors in generating high-value audiences: quality and flexibility.

Know what you are buying
The first question an agency, marketer or publisher must ask when evaluating a behavioral targeting offering is, "What am I really getting?" Insist on a full explanation of how audiences are created: the specific data that's used, how it is gathered, how it is leveraged in the targeting process, how often it is updated, how privacy standards are met, et cetera. If you don't ask these questions early on, you may be in for a rude awakening.

Creating quality
The promise and power of behavioral targeting is usually -- maybe not 100 percent of the time but almost -- proportional to the depth of data used to create the audiences. At Revenue Science, for example, we can create audiences by looking at frequency of visits, what content is viewed (down to individual words on a page), postings made, keywords searched, online tools used, and geographic and demographic data, to name just a few of the many available data sources. What actually gets used depends on an advertiser's objectives. However, the overall ability to draw from a wide range of sources means that the audience targeted can be as specific as the campaign objectives' mandate. 

One-dimensional behavioral targeting may still be BT, however, the limitations of utilizing only one indicator of interest and intent can exponentially affect the qualitative and quantitative results of your campaign.

Say that a campaign is targeted to a financial audience by including everyone who visits a mutual fund page once. What is the value of that individual to the advertiser? Now, say you narrow the definition to everyone who visited the mutual fund page twice in the past month. It's getting better. How about visited the mutual fund page twice in the past month and read an article about investing in the stock market? You can quickly see how much more relevant you can be when you combine criteria to truly determine the right audience for your message.

Flexibility to meet marketer goals
The other main critical element of BT, hand-in-hand with quality, is the flexibility that a true behavioral targeting platform provides. That is, how the various data can be dynamically combined to create audiences that match the marketer's objectives. To use BT to its highest potential, behaviorally targeted audiences must be customized to the specific needs of the advertiser.

Using advertiser site visits or pre-defined audiences may have value, but they only scratch the surface of BT, and may have limited appeal for a marketer who wants to make BT an ongoing part of the media plan. Some publishers and BT providers will claim that 10 to 20 pre-defined segments are more than enough to satisfy their advertisers. Take a minute to examine this claim and it quickly breaks down.

Consider how many behaviorally targeted audiences a publisher would need for the automotive category. First, there are obvious brand-price levels: i.e. the difference between a Ford, a Lexus and a Ferrari. Each brand has a range of price points, from entry-level vehicles to top-of-the-line, all with a distinct customer target. What does the audience for the BMW convertible sedan with turbocharged engine look like compared to the audience for the BMW SUV? There are also brand attributes that come into play in determining the behaviors that should be looked at: Cadillac and BMW are both luxury vehicles, but the customers are very different.

This example shows that one, or even 10, pre-defined audiences are not sufficient to meet the needs of all automotive advertisers. It's actually interesting to see how anyone can fit 100+ million uniques into such small buckets. How would the agency for Lexus react to the idea that it's buying the exact same audience just sold to Cadillac, BMW and Jaguar?

Clearly, the flexibility to build customized audiences is key, as is a targeting platform's ability to create these audiences on the fly and adjust them on demand as marketers' needs evolve.

Knowing what you are buying by understanding the impact that quality and flexibility have on BT campaign performance -- and asking tough questions about them -- is the best way to dispel the confusion surrounding BT to get more from your advertising investment. Combine this kind of insight with scale and you have the right way to reach the right people at the right time, all the time.

Jeff Hirsch is chief revenue officer of Revenue Science.  Read full bio.