Contextual vs. Behavioral Targeting

In a study that asked adult consumers what type of targeted online ad they are likely to respond to (read: click on), 62 percent cited contextual-- ("a subject of particular interest to you"). That's more than twice the number who said they are likely to respond to demographic criteria (28 percent-- "a specific group you may be a member of"). Just 24 percent of those surveyed said they respond most to geographic targeting ("businesses in your local community"), and 18 percent said they were most apt to respond to behavioral targeting (ads based on "your past behavior on a given website").

Fully one-third of respondents weren't sure or had no opinion on what method moves them most.

Vendare, the ad network that commissioned the national study of 1,000 adults, concludes from the results that subject matter matters most in online advertising. "When you present advertising on top-of-mind topics, consumers respond," says Lynn D'Alessandro, vice president, sales, for Vendare Media's Traffic Marketplace. "Contextual advertising matches your messaging with pages on related topics, and you can't really get any more relevant than that."

Anand Subramanian, CEO of ContextWeb, agrees: "Contextual targeting multiplies the opportunity for advertisers to communicate with the user by messaging different offers to the same user at different times. There are only so many online users, but there is infinite number of online user interests or mindsets. A typical online user may look at more than 10 to 15 different types of content in any given day."

On the other hand, he says, "Behavioral scales with unique users. There are only so many online users making it supply gated, and can therefore only grow as much as the online user community grows. As online advertising gets bigger, behavioral's share of the overall online ad spend will shrink."

That's not to say that other targeting methods should be dismissed, including behavioral. Based on other results that show responses to various types of targeting vary by age groups, economic status and ethnicity, Vendare believes that a mixture of targeting practices is the smartest strategy.

For example, the research shows that those with annual household incomes above $75,000 and those with post-graduate degrees are roughly twice as likely to respond to demographic, geographic and behavioral targeting as their counterparts at the bottom income and educational levels. And women are slightly more likely to respond to contextual targeting than men (63 percent to 60 percent).

Among other findings of note:

  • Response to contextual targeting varies according to racial makeup: 72 percent of nonwhites respond to relevant subject matter, against 60 percent of whites.
  • Those employed part time are significantly more likely to respond to demographic targeting than those in any other employment category (by as much as a 15-point margin).
  • Demographic targeting is nearly twice as important to those in the 35 to 44 age bracket as to those ages 18 to 24.
  • Geographic/local targeting is second only to contextual among those in the West, while demographic targeting is relatively more appealing among those in the Northeast, the Midwest and the South.
  • Behavioral targeting was three times more popular among the 35 to 44 age group than with the 18 to 24 crowd.

"Targeting isn't an either/or decision, so it's best to mix and match targeting tactics-- especially when it comes to educated, prosperous consumers who simply won't respond to messages that aren't personally relevant," Vendare's D'Alessandro says. "If your messaging says 'we know what's on your mind,' plus 'we are you,' 'we are where you live,' and 'we are ready to handle your online needs,' you have four good shots at reaching your best customers."

A mix of targeting tactics does seem to be the best strategy. Preliminary results from a recent study conducted by Next Century Media for TACODA Systems suggests that contextual targeting should be used at the beginning of a campaign but that for subsequent frequency behavioral targeting should be used.

The Tacoda study used eye tracking to compare behavioral targeting to contextual targeting in terms of advertising awareness, branding measures and ROI.

Back to the Vendare study: "This study is interesting in how it takes into account a variety of online advertising solutions instead of focusing on just one," says ContextWeb's Subramanian. "Compared to other forms of online advertising such as search or behavioral, there have not been as many studies on the effects of contextual. I think this study is positive in its ability to look at how consumers view the technology from a user's perspective and draws on many different demographics to generate its responses."

Not everyone in the industry agrees with the study's methodology, however. Bill Harvey, founder of Next Century Media and an interactive media advisor to advertisers, agencies, entertainment and media companies, points out its flaws:

"Asking consumers directly to opine as advertising experts on their own proclivities is a mistake that the traditional media learned not to use decades ago when so taught by ARF, advertisers and agencies," Harvey says. "The reality is that consumers don't even know what behavioral targeting is (many in the industry still don't know) and that makes the consumers-as-experts error even more egregious in this instance.

"Consumers predictably always say that they prefer to be targeted based on their interests because it is obviously in their self-interest to say that. This kind of survey has been done before and nothing new is learned by this.

"The other judgment error here is the implicit assumption that clickthrough is the end-all and be-all whereas about nine out of 10 advertisers are looking for communications effects, a.k.a. branding."

Dave Morgan, CEO of Tacoda, a behavioral targeting network, agrees with Harvey's assessment. "There is a big difference between clicks and qualified audience," Morgan says. "While we have seen research that shows that behavioral can sometimes drive better response than contextual, we have seen the bigger differences when you look deeper at the quality of the users' responses. Behavioral does even better when you look at viewthroughs and conversions and, most importantly, brand metrics like engagement and favorability and purchase intent. We think that it is critical that our industry stop looking only at clicks. It is a misleading and superficial metric."

So what's the bottom line?

As with all studies, results should be taken with a grain of salt. Is there some relevancy to the findings? Of course. It's definitely helpful to know how consumers perceive they're being targeted to, and how they think they prefer to get their information.

Are there flaws to the methodology? Probably. There are always at least a dozen, if not hundreds, of ways to collect, decipher and analyze the data. But that's what provides us food for thought, and keeps us looking at the topics from many angles. We'll continue to cover the numerous studies here at iMedia, and we welcome your thoughts.