Behavioral targeting grows

By Published on .

Targeting customers based on their behavior has long been a part of the marketing conversation. But thanks to the wealth of behavioral information they can glean from their Web site traffic and analytic applications that make these data, once the sole domain of technology specialists and statisticians, more accessible and comprehensible, growing numbers of marketers are starting to embrace the technique.

Fourteen percent of b-to-b marketers are currently using behavioral targeting in their marketing strategy, and another 8% said they are piloting or plan to use behavioral targeting this year, according to Forrester Research's Q4 Marketing Benchmark survey.

The allure is obvious: Targeting prospects based on previous actions can have significant positive impact on marketers' conversion rates, since past behavior can be used to predict a customer's future actions.

"We've been looking at segmentation and behavior on the Web," said Theresa Kushner, director-customer intelligence at Cisco Systems. Kushner directs a team of analytics specialists whose job is to increase the company's market share by applying insights gathered from customer data.

"You take the same information you always had from purchase information, and you analyze actions on the Web as well."

She explained: "We're trying to understand the behavior people take before and after whatever action we are analyzing on the Web. Sometimes it's where they come from before they hit our site. You can usually tell where they are going, too. By looking at those three stages, you at least get a segment of their time, and their behavior tells us all kinds of things."

In one case, Cisco instituted a pilot program to get people to look at a vertical application for financial services. The customers would hit the product pages first, then click to a vertical page.

The page created dealt with a subject that particular vertical segment cared about. In this case, "they care about the new payment card industry standard on how you check credit cards," she said. Then, if they clicked back to the product page again, "maybe what they get on that product page the second time is a chance to talk to someone about something they are interested in."

Kushner said that what is most important is what you do with the data. Cisco plans to use these behavioral insights about visitors to its Web site in a number of ways.

"We're looking at the kinds of people who do similar things on the pages and then try to see if there are things we can present them with to get them to engage with us further," Kushner said, noting that this becomes difficult as those segmentations are more finely sliced and diced.

"I think that is why most people are experimenting with what you do, because if now all of a sudden you have 50,000 segments [of customer behaviors] … it's like doing 50,000 individual marketing campaigns. You need to think through all of the possibilities and iterations. That is a lot of thought and a lot of upfront planning."

Forrest Leighton, senior marketing manager, product systems division at Canon USA, said that he employs behavioral targeting in campaigns, but he does not base it on buying habits since Canon's products are sold through dealers.

"What we do is based on behavior—where they respond to a previous campaign—and we are able to target more specifically to them," he said.

Leighton is responsible for Canon's high-end digital production equipment. "I want to be able to understand the pain points of my customer base and my prospect base," he said. "That can drive how we are going to respond to them." He said the messaging and offer differ based on those pain points.

"We're working on a number of different campaigns to look at how we can better target our customers," Leighton said. "We don't have a set date to launch something specific, but we're working towards even more sophisticated behavioral targeting."

Privacy concerns

Despite excitement about behavioral targeting's potential, some red flags have appeared. The problem? Privacy concerns.

The FTC last week announced it will host a two-day town hall meeting in November to bring together consumer advocates, industry representatives, technology experts and academics to address the consumer protection issues raised by behavioral targeting.

The FTC examined similar issues in 2000 and issued two reports on the practice of online profiling.

"Technology advances and the evolution of business models since that time have raised concerns by consumer advocates, privacy experts and others about the implications of data collection in online advertising," the FTC said in its announcement.

Topics at the meeting will include: how behavioral advertising works; what types of companies play a role in the behavioral advertising market; how data are used and by whom; and how behavioral advertising has changed since 2000.

It is unclear how much of an effect the privacy issue will have on the b-to-b side. Like complaints around telemarketing and subsequent do-not-call legislation, privacy may primarily be a consumer issue.

A possibly bigger obstacle is what some say is the inexact science of behavioral targeting. Observers note there will always be a certain amount of guesswork involved when dealing with human nature and making predictions about future behavior.

"The whole notion of having a technology [or process] that can accurately interpret human intention is very difficult to achieve," said Laura Ramos, VP at Forrester Research.

"The trend I'm seeing is people use Web analytics to understand `Am I getting traffic?' and `Where is that traffic going?"' Ramos said. "You want to identify individuals in that traffic and use the tools to gauge engagement: Who comes back to the Web site, what do they look at, what do they download? I haven't seen widespread evidence that b-to-b firms have figured this out."

The nut to crack, Ramos said, is how to move from general analytics to the specifics about the buyer's engagement.

"The ultimate goal," she said, "is to use technology to capture buyer behavioral events and automate the process of figuring out which ones are more engaged."

Most Popular
In this article: