How and Why Behavioral Advertising Works

If you’ve been advertising online for a long time, you may have gone through stages: sticking with banner ads at first, and then going with search engine advertising, and maybe putting your ads on a publisher network belonging to a search engine or an advertising company. Most of the time you probably tried to put your ad in a matching context. That might be the wrong approach.

I’ve written before about behavioral advertising, also known as behavioral targeting. You can read my first article about it here. If the topic of behavioral targeting intrigues you, you might also want to read about behavioral retargeting. Before I plunge into the content and focus of this article, though, let me give you a quick definition. Behavioral advertising is a form of online advertising that follows the user around. For example, a web surfer who has just priced some flights on an airline’s website might be shown a travel-related ad when he surfs to the next website in which he’s interested, which might be for the local pizza joint.

The theory behind behavioral advertising is, in a sense, pretty simple. Most people are bombarded with ads most of the time, especially when web surfing. As a result, we tune them out. Because of the usual advertising practices, we might be better at tuning out ads that are in the same context as the content we’re reading. In other words, someone reading content on a web site about where the best ski slopes are just might have completely ignored an ad for your lovely Aspen getaway. To rise above this clamor, it’s necessary to hit web surfers with a surprise, something that doesn’t fit the normal context. Think about it: aren’t you more likely to stare at someone talking into a banana than a cell phone?

That’s the theory, but it’s new enough that researchers and marketers are still doing surveys to prove or disprove it. The most recent one was conducted by BL Labs and released by ad network BlueLithium. You’d probably expect it to be self-serving, at least to some degree, but the findings showed a certain amount of diversity. Indeed, judging from the results, you might want to think twice about using behavioral advertising, depending on your (or your client’s) specific field.

“We set out to prove that behavioral targeting performed better than contextual, and we wanted to see if that was true across the board,” explained Dakota Sullivan, BlueLithium’s Chief Marketing Officer. “What we found was it’s more complicated than that, and it depends on the goal of the marketer.” Indeed, as you’ll see, it depends on the marketing category as well and probably some other factors too.

BlueLithum’s study looked at 400 million impressions served to users based on the sites they had visited. So web surfers who had previously visited entertainment sites saw entertainment-related ads at the next site they visited; visitors to travel sites saw travel-related ads at the next site; and so forth. BL Labs sifted out nine behavioral categories that had over 10 million impressions, and analyzed them for patterns across various contextual categories.

If you look at the overall numbers, they present a very strong case in favor of advertising out of context, particularly in the way that it’s done with behavioral targeting. Two metrics were measured and compared: the click-through rate (CTR) and the action-through rate (ATR), sometimes known as the conversion rate. Overall, the CTR was more than 100 percent higher for the ads that were shown out of context – that is, in a different content category than the ad itself. And the ATR was 19 percent higher than for ads shown in the same context.

Before you blow your entire online marketing budget on behavioral advertising, though, you need to know that some categories showed very significant deviations from these figures. Take the web surfers who were classified as engaging in “business and finance” behavior. Their CTR was more than 100 percent higher for ads shown in the SAME content category. Their ATR was also higher for ads shown in the same content category – 128 percent higher, to be exact.

The behavior of the web surfers in the “entertainment” category told a different story though. If you’re showing ads to them which are targeted to the same content category (as opposed to ads targeted out of context), you’ll see click-through rates that are 92 percent lower than ones shown out of context, and action-through rates that are 66 percent lower. It’s tempting to draw a connection to the focuses or attention spans of web surfers interested in different categories, but that oversimplifies the issue.

Looking more closely at specific categories can be even more eye-opening. Check out the “shoppers” category of web surfers, for instance. Shoppers clicked through most often on ads on career sites, but when it came to actually taking action, that happened most on female-oriented sites. And for the “Travelers” category, while the greatest click-through rate came from food sites, the greatest action-through rate came from career sites.

Again, it’s very tempting to speculate about the reasons why. Jason Lee Miller, commenting on the study for Webpronews, noted that “This could make sense if you think that job-seekers just dream about the new stuff they can get if they get that new job. They don’t actually have the means to buy that cashmere sweater just yet. Women with credit cards at the ready, however, are a different story altogether.”

Of course, we could overanalyze this to our heart’s content, but that would be missing the forest for the trees. What’s the main take-home point here? “Behavioral targeting is one of the most effective techniques in driving conversions, but what this study reveals is that pre-packaged audience segments and automated rules-based targeting may not capture the full benefits of behavioral targeting,” pointed out Alyson Yaffe, media supervisor for Media Contacts. “Instead, ad networks need to focus on custom segmentation and data analysis to optimize the nuances of behavioral targeting for each campaign.”

It’s worth noting that the way this study was conducted, and its findings, are more sophisticated and nuanced than previous studies that examined behavioral targeting. For example, an eye-tracking study conducted within the past year by the media company Tacoda and Next Century Media found that, on average, the same ad was looked at 17 percent more when it was first exposed to web surfers in a way that targeted behaviors rather than context. After the ad appeared a second time, web surfers gave it 54 percent more looks if it was shown in a way that targeted behavior rather than context.

It sounds impressive, but that particular study was limited. The sample was made up of 18-to-64-year-olds shopping in New Jersey and California malls for plasma TVs, new cars, and computers – and only 15 participants were tested for each targeting type. Note also that the study measured only looks and time spent, not CTR and ATR as does the BL Labs study.

It shouldn’t surprise you that behavioral targeting is trickier than it first appears (and it certainly doesn’t appear that easy to begin with!). Some marketers may have touted it as a panacea, but that’s one thing it clearly isn’t. As I said in my first article on the subject, behavioral advertising isn’t for everybody.

In that sense, though, it’s no different from any other form of advertising. As Sullivan said, it depends on the goal (and probably the budget) of the advertiser. Would you advise a hardware store to advertise in a nationally-distributed magazine? If it’s a mom-and-pop operation, probably not, but if it’s Ace Hardware or Home Depot, absolutely, assuming it’s the right audience. And if the mom-and-pop hardware store has a web site and is looking to expand its business, that magazine ad might be something to consider down the line.

The same thing holds true for behavioral advertising, but there’s another variable to work with. We now know from the BL Labs study that certain fields do better with behavioral targeting, while others do better with contextual advertising. This is something to watch for, and any advertising campaign that includes behavioral targeting as one of its elements should take this point into consideration when it is planned.

Interestingly, BlueLithium’s study may lend strength to an approach that was suggested by the Tacoda study. If you’re thinking of trying behavioral targeting, you might want to try a contextual advertising campaign first, and then follow up with the behavioral one. There’s a number of reasons for this; it will help you gain reach and some recognition, which can’t hurt. But it will also give you some base numbers to compare with the behavioral campaign, so you can get some idea of where it will benefit you most to invest your marketing dollars.

Google+ Comments

Google+ Comments