Marketers tend to define behavioral targeting in two quite different ways. Digital denizens see it as a pure data play, reacting to how people behave online to determine what to do to convert them. Marketing theorists usually look at it as an exercise in psychological research aimed at learning why customers react the way they do to then using that information to shape long-term brand and marketing strategies.
There’s no question that behavioral targeting encompasses both of these definitions and that its use will continue to stretch into new areas as both data and analytic technologies continue to grow. Reacting to customers’ online behavior can deliver immediate results, so attention to digital behavioral targeting is top of the agenda. But a lack of in-depth research into people’s purchase motivations can derail the long-term effects of digital efforts by losing customers along the way for lack of relevance. Considering the disparity in what many marketers say behavioral targeting is, we asked several experts to tell us what it’s not.
MYTH 1
Behavioral targeting is easy to implement
Reality: Behavioral targeting can be expensive, and is often limited by the capabilities of the media through which marketers may be running their campaigns. So, it’s not a tactic to trot out for every digital campaign. It’s best to focus behavioral campaigns in areas where the success factor is likely to be high.
“Retargeting, for instance, is a subset of behavioral targeting that is getting a lot of attention today. But it’s a lower-funnel activity. You use it to get conversion,” says Mary Pocsik, executive director of media at Geometry Global. Pocsik notes that retargeting can be an effective approach when marketers expect a given number of leads within a specified budget.
One such Geometry Global client is an insurance company that sells a healthcare product with a long research cycle, but a short selling period. The company uses various methods to communicate to prospects throughout the year about its plans; but those plans are confusing and complex. Purchasers delay making decisions, and most transactions are done in a competitive, nine-week stretch at the end of the year.
Geometry Global works most of the year building cookie pools for the insurer on prospects when they’re doing their homework, all in anticipation of a finely tuned digital campaign aiming for maximum policy applications during a two-week free-for-all. “When you’re in such a tight window and the potential payoff is big, behavioral targeting can be one of the most cost-efficient tactics,” Pocsik says.
Often, however, marketers looking to use behavioral tactics are constrained by opportunity or budget. “Advertisers can set whatever parameters they want, but they can be limited because there’s little leeway in specific user networks,” she says, noting that publishers may not be able to provide scalable platforms for some behavioral targeting campaigns.
“What we’ve found is that it’s a strong tactic, but it’s not the be-all and end-all,” Pocsik says. “The whole opportunity of behavioral marketing is that we’re actually going to talk to segments in a more relevant fashion. But it takes more time and money, and you’ve got to have some overlap using contextual ads and content integration.”
MYTH 2
Behavioral targeting automatically improves display advertising
Reality: Many marketers don’t gain the improvements they expect when using behavioral targeting to enhance their display advertising in terms of copy, timing, and placement. The reason: These marketers are using behavioral targeting incorrectly, says Jesse Cahill, client director at Essence, a digital marketing agency.
“One of the main myths you hear about behavioral targeting [is that behavioral targeting improves display ads], but the method misses the mark in this regard,” Cahill says. “For one thing, we don’t value recency enough. We ignore the implications of frequency capping. Even when [a marketer is] using a data point that ensures an ad is relevant, it can easily go from relevant to annoying.”
Too many marketers use behavioral targeting as an add-on that’s rarely included in the pre-campaign planning process, Cahill maintains. Purchase consideration cycles can vary widely for certain products and services, so marketers need to test placement frequency to determine customers’ reactions within categories. Plus, factors such as cookie deletion and device sharing can muddy up data.
“We test the impact of recency on performance by setting up segments and looking at their interactions over the last seven days and then the last 24 hours. That’s the best way to determine retargeting against those pools [using display ads],” Cahill says. “Just because someone exhibited a certain behavior in the last seven days doesn’t mean messaging to that person at any time of day will be good. Too often, these moves are based on the intuitions of the media buyer.”
MYTH 3
Behavioral targeting is purely a real-time digital play
Reality: Customers must also be engaged over time to determine what motivates them to purchase in the first place.
Daryl Travis thinks Amazon may be getting too much credit. The e-commerce idol’s recommendation engine is the gold standard of digital behavioral marketing. But Travis, CEO of consumer insights firm Brandtrust and author of Emotional Branding, contends that although Amazon marketers might know when they’re getting it right transactionally, they’ll never know when they’re getting it wrong experientially. A customer buys a suggested product and Amazon sees a victory. But Amazon gets no direct feedback on all the wrong recommendations that customers turn their noses up to, or may even be offended by. “No one really considers the cost of getting it wrong,” Travis says.
How people relate to brands is more about feelings than facts. Brand survival in the long run, Travis says, depends more on learning their emotional involvement than their click-throughs on websites or emails. He gives the example of an auto insurance client looking to analyze whether prospects could be rated for policy approval based on their behavior on the company’s website.
“This company likes to treat everyone fairly, and so it has a policy of explaining to people the reasons why they were rejected,” Travis says. “After we finished personal interviews with a sample of people who visited the site, we got better satisfaction scores from the people who were rejected than the people who were approved. They expected to be treated rudely, and when they weren’t, they took away a positive view of the company.”
Such analyses give marketers a longitudinal look at customer behavior that can be accessed over time to build future prospect profiles rather than solely acted on within the constraints of a single campaign. Travis says such behavioral research can be done as easily—if not easier—than tracking behaviors online.
“There’s a technique we use called pulsing. We hit people with a few questions on their smartphones sporadically,” Travis says. “We can measure their energy level—positive or negative—for a brand quickly and simply and aggregate it into a full study of customer behavior over time.”
MYTH 4
Behavioral targeting is the answer to realizing true one-to-one marketing online
Reality: One-to-one marketing is a concept as old as retail. Millenia before the advent of the Web, buyers met sellers face-to-face. Ecommerce may be more convenient than a brick-and-mortar store, but it can never be more personal, even with behavioral analytics empowering marketers with personalization tools.
“These technologies allow sellers to create a unique, digital shopping experience for a buyer at two levels: one-to-many and one-to-one,” says Gartner Research Director Penny Gillespie. “While there is a lot of one-to-many personalization, few if any companies offer true one-to-one personalization because this requires incorporating all six types of personalization: [buyer-set preferences, preferred channels, product relationships, buyer’s location, behavior patterns, and knowing the customer by following his actions across all channels]. This is hard to do because there is no single solution; it’s expensive and requires multiple technologies.” In fact, Gartner estimates that currently less than 1% of digital customer experiences incorporate all six types of personalization.
According to Gillespie, although currently a differentiator for marketers, personalized digital experiences will be table stakes within five years.
MYTH 5
Behavioral targeting doesn’t need historical information
Reality: Those who use predictive analytics to power behavioral targeting believe that the tactic’s true promise is based on identifying customer intent by analyzing customer data over a period of time; that it’s not solely about reacting to what already happened, but should be used to predict what’s likely to happen. These marketers think that, when it comes to behavioral targeting, real time is unrealistic and expensive.
“Look at it this way. Say you see two people at the starting line of a 5K race and both look fit and fast. But one guy’s run five races in the past six months and the other guy’s run none,” says Glenn Pingul, VP of products and mobile strategies at Globys, a data analytics company that assists financial services companies and telecoms with behavioral targeting. “Without having the history of what someone’s done in the past, your understanding of that person is limited.”
Globys conducts a great deal of research on customers’ use of prepaid calling plans because they must actively update their accounts and, therefore, provide telecom marketers with rich behavioral data. Most prepaid customers set out to recharge every 30 days, but often they run out of minutes after 20 days. Globus analyzes tens of thousands of actions taken by these customers, aggregates the data on those actions, and then segments the customers based on their actions. This can help to inform how companies should market to those customer segments, including what incentives they should offer.
“We have a client in Asia that has well over 40 different rate plans, yet what we found after analyzing all the data from its customer base is that there are only seven distinct behaviors,” Pingul says. “Now, amazingly, this particular client chose to keep all the different rate plans, but what they could do is take the segment that is spending extra money topping up their plans and offer them a cheaper alternative that fits their true usage levels. Maybe they’ll lose $10 a month from a customer, but at least they’ll keep him.”
Pingul says if marketers are going to commit the budget needed to do behavioral targeting, they should use it to do more than anticipate conversion tendencies. In fact, he maintains that marketers can use behavioral targeting to change customers’ habits. Globys and a partner company are currently experimenting with something they call the Nudge Theory. Consider a quick-service restaurant operator as an example. The restaurant operator conducts a study and finds that 15% of patrons who show up at its locations at noon end up leaving because of long lines. The restaurant operator could use behavioral targeting to send those customers offers that could entice them to come in at 11:30 a.m. instead. In this case behavioral targeting might help to recapture lost revenue.
“Especially with mobile technology, you could do this right now at any level,” Pingul says. “You can test how [customers] respond, via SMS or email. You can test negative or positive messages.”
And, he emphasizes that deeper testing of multiple customer reactions unveils historical patterns that can be used to fashion future campaigns. “It’s longitudinal—looking at behavior over time—versus vertical, which looks at behavior at the current moment in time.”
