ValueClick's VP of corporate strategy takes issue with the assertion that working with more than one network is a waste of time.

Much has been written about behavioral targeting, and the information about the technique seems to be proliferating as it gains in popularity. Unfortunately, as tends to happen with each exciting new development in marketing, the resultant flood of information often brings more confusion than clarity.

Such was the case with the recent article in iMedia Connection by Advertising.com's Brent Halliburton regarding behavioral targeting. To summarize the piece, Mr. Halliburton suggested that if you go with the largest display ad network there was no need to use any other networks or properties because of audience overlap. That's it -- it's that simple -- just always go with the biggest. At best, this rationale is naïve, and at worst it's deceptive and irresponsible.

In a stunning oversimplification of a complex issue, the article in essence refuted the approach that direct response marketers have refined over decades; the recognition that there are no absolutes, no guarantees, and that you must test to prove.

To follow Advertising.com's recommendation to its logical conclusion, a marketer or agency should also never test direct mail or email lists; rather, he should just go with the biggest one because, as we all know, everyone has all the same names. Also, since many consumers use multiple search engines, marketers should only use Google since it's the biggest. Or since Yahoo! is the largest portal, marketers should only buy media with it (I wonder how Advertising.com's parent company TimeWarner/AOL would like those last two assertions!). The flaw in this thinking can be used in every marketing channel, both online and offline. The fact is, bigger is only better if it can provide both the scale and the desired results, and a marketer won't know that until it is tested.

Don't get me wrong: reach is important. Indeed, ValueClick Media takes great pride in having the largest reach of all independent ad networks. But in reality, reach is only one component of a successful campaign. Advertising.com rightly notes that as ad networks have grown, a substantial number of the same visitors appear across multiple networks over the course of a month. However, each network is composed of different combinations of websites, and because every network's composition is different, and each network's capabilities are different (more about this below), each produces different results for different marketers.

More importantly, each network only sees a fraction of each visitor's total monthly internet activity. Since most networks see a large percentage of their visitors only once each month, intelligent marketers try to reach their audience multiple times, engaging multiple networks, portals and solo media properties. And Advertising.com's stated concern regarding frequency capping across placements has been easily and inexpensively accommodated for years now via third-party ad servers (which also provides the benefit of consistency of serving and tracking as well as enhanced reporting and other extras).

All effective advertising -- behavioral targeting included -- seeks to reach the target audience the optimal number of times across the optimal placements in the optimal channel mix to affect the desired result, which is significantly more complicated than simply making a placement on the largest-reach property.

In addition to the errant notion that unique reach is all that matters, Advertising.com's approach contains another fundamental error: It assumes that all other components of campaign execution outside of reach are equal. Besides insulting the intelligence of marketers and agencies with this reasoning, this viewpoint also pales in light of the overwhelming evidence against it. For instance, the following is just a partial list of key differentiators between ad networks (as well as portals and solo media properties) that intelligent marketers compare before placing a buy:

• Targeting logic: While most networks offer similar targeting options (e.g., demographic, geographical, daypart, et cetera), the logic within their targeting engines varies greatly. For instance, some context-focused networks serve ads based on guaranteed delivery. Thus, even if the ad is not relevant to a particular visitor, to meet their delivery quota they decide to show it anyway. The best targeting engines base their decision on relevancy of the creative to each visitor. 
  
• Optimization logic and frequency: The ability to refine a campaign intra-flight, the data and logic used to do so, and the speed with which this can occur needs to be taken into consideration. For instance, some networks process data only once a day, which means they are slower to improve than those networks that process data at least once an hour.
  
• Transparency: This is a hot topic these days, as networks continue to evolve and as large brands continue shifting budgets to the online channel. There is a wide disparity between networks as to the degree of site revelation they provide to marketers.
  
• Pricing model and pricing level: CPM vs. CPC vs. CPA is certainly an important aspect of a network buy that varies greatly across networks, as well as the actual level of the price itself within those pricing models. Ultimately, as noted above, the flip side of this coin is the ROI that the program achieves.
• Service: Another important consideration when choosing network partners is the customer service you will receive. We've all had occasions in which we've bought something from the biggest, most well-known provider of something, only to be greatly disappointed at being treated like a number. Size and service have no direct correlation. For instance, an important question to ask your network partner is how fast it can make a change in its system once a campaign has begun delivering. Many networks take longer than 24 hours to make the requested change to a live campaign, which may not seem like a long time but it results in wasted media dollars.
• Advanced capabilities: Do the networks being considered have additional, proprietary data they can apply to further enhance your behavioral targeting? Do they offer full rich media and/or video advertising capabilities? Can they accommodate highly personalized messaging that can be generated by compiling all available data on each consumer in real-time? These and many more capabilities further separate networks and other properties from each other, and can make or break a campaign.

Ultimately, marketers don't care if you're the biggest. They care if you're the best, which is determined by your ability to solve their marketing challenges to their satisfaction. This is accomplished by a continuous cycle of testing, learning and refining. To suggest otherwise, as Advertising.com does, is to suggest that Behavioral Targeting for Dummies is the right approach. To change that to Behavioral Targeting for Savvy Marketers, what is needed is replacing sales spin with sensible education, sound marketing strategy and, of course, performance.

John Ardis is vice president, corporate strategy for ValueClick Inc. Read full bio.