Be relevant, not creepy

Behavioral targeting isn’t a new concept. In fact, it’s “direct marketing 101,” according to Joe Stanhope, principal analyst at Cambridge, MA-based Forrester Research.

Behavioral targeting—the practice of using individual customers’ past actions, such as purchase history and online activity, to more precisely market products—may be a fundamental marketing principle, but collecting and analyzing customer data across  multiple channels, and the subsequent privacy concerns have never been so complex. Marketers have much to learn if they are to use behavioral targeting in a way that improves marketing relevance without seeming too invasive to prospects and customers.

For many companies this reconciliation is still a work in progress. Marketers at companies from large enterprises to smaller, more local businesses increasingly support their marketing efforts with data from social channels and by connecting touchpoints like email and websites for a comprehensive customer view. At the same time, these marketers must overcome such challenges as organizational silos and educating customers about data privacy.

Targeting with limited resources

Behavioral targeting is often associated with massive campaigns for multinational corporations, using pools of customer data as deep as the Marianas Trench. But even smaller-scale businesses have the need, and the tools, to use behavioral targeting. Taylor Brannen, Internet marketing director for South Point Hyundai and Kia, operates at a local level targeting automotive consumers in Austin, TX. Because of limited resources, Internet marketing is a modest department in such an enterprise. Brannen acknowledges that for much of the targeting and tracking techniques, “there’s nobody but me to do it.” Whereas big-budget behavioral targeting initiatives can run about “$10,000 a day,” according to Brannen, South Point Hyundai and Kia doesn’t have the funds to invest in costly technology.

Brannen notes that the best behavioral targeting platform for his needs is Google Analytics. It includes a segmentation tool, Barilliance, which has a feature that allows site owners to determine website traffic based on pre-established visitor segments like search terms, time of day, and traffic source. “[Google Analytics] is the main tool [we use] for finding out what kind of traffic we’re getting on our website, and we can break down where and who our customers are. It’s a pretty powerful tool when trying to decide where to send 20,000 mailers—and the cost is nothing.”

Brannen uses digital marketing solution IMN Email Marketing Services & Newsletter Management to assess the dealership’s customer base and observe digital trends. “IMN allows me to use different types of digital advertising within my newsletter,” he says. “I can track how people respond to texting to obtain information or pricing versus emailing, [and whether] they prefer to read informational articles about a product that sparks their interest versus something more overt like a traditional graphical advertisement.” Using IMN, Brannen says he is also able to track engagement rates of fundraisers and competitions. “The reporting tools at my disposal make it very easy to adjust to my demographic,” he says.

The burgeoning social component

Social media is another cache of rich customer data, but Brannen doesn’t tap into any social feeds yet. Instead he uses social channels as a forum for conversation and feedback within the dealership’s target demographic of consumers ages 18 to 25. But behavioral targeting has made its way into the social media realm as another source of customer data that an increasing number of marketers are interested in accessing. For example, Aaron Magness, VP of marketing for online eyewear retailer Coastal.com, is looking into featuring social sign-ins for targeting purposes on its site.

“If someone were to come onto our site using their Facebook log-in, we would have tons more access into information about them,” Magness says, adding that the Facebook social graph provides “a lot of info that isn’t readily available in general…. You can access everything [consumers] have added to Facebook, such as sex, birthday, likes, interests, etc. Of course, all of this comes with the approval of the customer. Having this rich set of data will allow us to better serve more relevant messages and offers to our customers.”

The company—which Magness says sold $40 million in eyewear in 2011—does not yet have a social log-in feature, but adding it is in the planning stages. “Facebook will be one of many [social media channels] we’ll support,” he says. “We wanted to make sure we found the right partner before we jumped into the social login realm, since the log-in is only part of the total social experience,” Magness adds.

Email targeting

For its present targeting efforts, Coastal.com thrives on a “treasure trove of consumer data.” That data plays a critical role in how and when the company retargets a customer with offers. “We’re able to look at different customer segments based on geography and history,” says Magness, who then uses that information to create profiles of Coastal.com’s customers. Did they buy glasses and then move onto contacts? Do they wear bifocals? How old are they? Are they due for an eye exam? After processing this data, gleaned from on-site activity, Coastal.com automatically sends an email containing relevant content to each specific customer. For instance, when marketing to a 29-year-old customer who has shown interest in fashion frames, the messaging will focus on fashion.

Additionally, Coastal.com is always testing messaging tactics, and often uses past purchases to influence future email content. “We’ll also look at other variables such as timing of purchase, category—glasses versus contacts—and other branding initiatives,” he says.

As a mega online retailer of shoes and apparel, Zappos.com is well-acquainted with the fashion-conscious consumer. Darrin Shamo, director of direct and online marketing at Zappos.com, says the company uses consumer data based on site activity to assemble customer profiles and determine a “segment of products” that either complement items the consumer has viewed, or are similar to the items viewed, he says. One way the company “recommends” these items is to serve them on its site as “Alternative products for you.”

Extending that “real shopping experience,” as Shamo describes it, is a top priority for Zappos. As an online-only retailer, there is an inherent challenge in targeting consumers. Shamo’s goal is to give a customer an alternate, and more relevant, experience than a brick-and-mortar retailer is able to offer.

Online-only home goods retailer Overstock.com takes a slightly different approach. “From a one-to-one targeting perspective, we mostly focus on email and customizing the site [based on customers’ behavior],” says Tim Dilworth, SVP of marketing.

Dilworth estimates that Overstock has more than 25 million customers and that their transactional information is a “powerful predictor” of a future purchase. “Targeted emails go out to those who have shown an interest in a particular category,” he notes. For example, “if a customer has just bought a bed, their likelihood to purchase bedding accessories is higher than say, their likelihood to shop the electronic department.”

Behavioral discretion

Both Overstock’s Dilworth and Zappo’s Shamo were guarded about their respective companies’ behavioral targeting methods. Part of the reticence is that they don’t want to speak too openly about what is essentially a competitive advantage. “There is competitive concern around what secret sauce brands are using and how,” says Eric J. Hansen, CEO and founder of SiteSpect, a Web optimization company that provides behavioral targeting tools.

Another reason for the hesitance: Few brands have perfected their behavioral targeting practices. “Nobody has behavioral targeting all figured out yet,” Coastal.com’s Magness suggests.

Shamo adds, “There are a lot of things we have failed on. We’ve had groundbreaking ideas only to find they don’t resonate with the audience.” Shamo declined to specify the nature of these failures, though he indicates that they’re part of a continuing experimental process, one in which even failed initiatives might eventually be changed and reintroduced. “This space is evolving so quickly,” he says, “We have yet to really discard anything—only to repurpose it.”

SiteSpect’s Hansen emphasizes that testing is an absolute necessity for brands that want to hone behavioral targeting practices. Online movie streaming service Netflix, which uses a recommendation system based on past viewing tendencies, “literally tested different algorithms for how to achieve successful targeting,” Hansen explains. Getting the recipe right for many brands involves using a number of different vendors and algorithms, he says.

For example, an algorithm that works well in the U.S. may not be so effective in Europe. “I’ve seen retailers test different targeting products and find the right one based on geography,” Hansen says. “A cross-selling engine from vendor A may work well for their U.S. site, while the cross-selling engine from vendor B may work better for the company’s site in Europe.”

Behavioral targeting in a silo

Selecting the right customer data and testing targeting algorithms and campaigns are only two legs of a three-legged stool when it comes to behavioral targeting. If the approach is to be successful, it must be fully integrated across numerous channels. Overstock’s Dilworth, for example, brings his entire marketing team together into a room and maps out all customer interactions within the brand to ensure that they all have the same goal.

Other companies take a similar approach. “We have folks specializing in social media outbound messaging and folks responsible for automated ad configuration—but our general strategy goes across all our channels and technologies,” says Stacey Sieger, VP of product management and marketing at B2B sales acceleration company, Sales Engine International.

“The fundamentals of targeting remain the same across all marketing endeavors,” Forrester’s Stanhope says. In other words, although targeting strategies vary across email, websites, and customer service, each division should access the same pool of customer data. This data integration will help to improve the accuracy of behavioral targeting, as well as create more consistent marketing messaging across channels.

Assuaging customer concerns

Encouraging the type of collaboration that breaks down organizational silos and allows for shared customer data is no easy feat. Another difficulty for brands is optimizing behavioral targeting initiatives while ensuring that customers feel that their privacy is being respected. One way to improve customer acceptance of behavioral targeting is through open discourse and transparency. “People have a right to be concerned about things they do not understand,” says Forrester’s Stanhope, who wrote the June report, “Behavioral Targeting Powers: Customized Content and Increased Conversion.”

Stanhope insists that it’s the responsibility of the brand to make a clear, comprehensible policy. “Companies need to be upfront and proactive,” Stanhope continues. “[As a brand], speak in plain
English and give consumers an option to opt out of targeting, and at the same time, explain simply that you can save them time, get them relevant offers, and find them the products they want. It’s not scary, but you have to be transparent and show them that.”

Overstock.com aims to be clear about its use of customer data. “We do not share consumer data with third parties,” Dilworth explains. “We have a stringent privacy policy that we are always reevaluating—and it may evolve.”

Still, as Dilworth notes, it’s difficult to predict the extent to which consumers will accept behavioral targeting—transparent or not. “If behavioral targeting gets bad press and becomes a big deal, that could make it come to a halt,” Dilworth says. “But marketers will listen to their consumers. It is fully in the hands of the customer.”

The degree of paranoia surrounding behavioral targeting may depend in part on a consumer’s comfort with the online sharing phenomenon that lives in social media. Those already fully acclimated to it, and generally unquestioning of the Internet, could help behavioral targeting flourish. “Younger people who are into Facebook and Twitter are generally more accepting of behavioral targeting,” Sales Engine’s Sieger explains. “Older people are generally less comfortable when they first hear [the term], but are [pleasantly] surprised when they find out what it could mean for them.”

Hansen of SiteSpect agrees that younger generations are generally less questioning of behavioral targeting. In fact, in Hansen’s experience, some young people are so comfortable with the behavioral targeting nature of online shopping that they depend on it for their shopping convenience.

“Personalization technologies and targeting bring an advantage to the online space and make it more efficient for users to shop,” Hansen explains. “[Brands] try to simulate the experience of going down the street and having store employees work with you, and they do it so well that a lot of these [young consumers] don’t even need—and may not even know—a brick-and-mortar experience.”

A juggling act

The predictive power of behavioral targeting is incontestable, but marketers should remember that there is a healthy balance between tracking customer behavior in the effort to predict a future purchase, and letting consumers’ nature take its course. It’s important to respect the randomness with which consumers may act, and note that not every move can be plotted. “The idea that we can look at the actions of our prospects and infer what materials they’re interested in, and what pain points they’re experiencing, and get content that is personalized for them is important, but take it with a grain of salt,” Sieger says.

Putting all one’s focus into forecasting what a buyer is most likely to do, “takes over [one’s] ability to learn what a consumer would do naturally,” she adds. “When we get too focused on behavioral targeting, we get tunnel vision. It’s important to expose—not push—our prospects.”

Coastal.com’s Magness follows a similar thought, saying that much as he believes in targeting, “if every ad was served as something that interested a consumer it would create noise.”

Ultimately, as Sieger observes, mastering behavioral targeting is a bit of a juggling act. “Marketers need not only be a marketer, but a statistician, a technologist, and be able to understand and analyze a lot of data,” Sieger says.

To be all these things, marketers need a three-pronged approach, Hansen says. “First, brands need a measurement of function in place with Web analytics,” he says. From there, brands can go on to testing. This is where, “you’ll see how to fix problems and improve your [approach].”

Not until analytics are in place can marketers begin testing; and not until they begin testing can a brand effectively move into targeting. “It’s walk, jog, run,” Hansen adds. “You can’t improve what you can’t measure, so start measuring, then experiment, and then start with the fun stuff: targeting.”