• by Phil Leggiere , June 26, 2008

The once radical premise of behaviorally targeted ad placement, using behavioral profiles to "reach the right prospect at the right time with the right ad," is now largely common wisdom. Marketers now have confidence behaviorally targeted ads can far outdrive untargeted ads to their Web sites. As Brent Hieggelke, vice president of strategic marketing at Omniture, explains below, however, most marketers still leave this newfound wisdom behind once visitors enter their Web site door.

 

BI: Omniture has made several new moves in the behavioral space in the year or so since we last spoke. How has your strategy in looking at the space changed?

Hieggelke: Our focus hasn't changed. It's on leveraging all elements of data that can optimize the site experience of consumers and the monetization potential of Web publisher content. We still think that most behavioral targeting dollars spent are concentrated on getting people to their sites, whether by display ads, paid search or natural search. Once they do that, however, targeting efforts for many sites seem to lapse. So all that effort is effectively wasted at the most important point of real engagement potential.

BI: How can Web site behavior be optimized?

Hieggelke: There are a wider variety of ways to optimize sites through behavioral data. Indeed the biggest challenge this industry faces in a way is to educate publishers that they have a wider range of options for how to go about aligning their marketing goals with behavioral targeting investments. It doesn't need to be a one size fits all.

Omniture Touch Clarity, for instance, is enabled to do one-to-one targeting. That means it leverages all of the behavioral variables each visitor brings to the site. That's one methodology that's important to some marketers, though not all marketers necessarily are ready for or even need that level of granularity. The segmentation product also enables marketers to optimize their site for specific consumer segments they've identified as important to their business.

BI: How do your segmentation tools work? How customizable are they, for instance?

Hieggelke: The segments can be as broad as developing landing pages and home pages for new vs. repeat visitors, or as narrow as customizing pages by referral source, according to specific channels that brought consumers to the site, which could be affiliate networks, display ads, email, specific offers or keywords. We can also customize site experience based on the type of products consumers have demonstrated interest in. For instance if you're a large music retailer you can segment incoming traffic by the categories of musical instruments visitors have demonstrated interest in such as drums, guitars or keyboards, and [we] optimize the content to show the most relevant to each visitor segment. You can segment by types of advertisements they've clicked on, whether by particular offers or even creative elements.

Of course it's up to the marketer to decide what segment criteria is going to be most important to their business, and to prioritize optimizing the site experience accordingly. Everybody obviously belongs to multiple potential segments. I live in Portland, Oregon, for instance. I'm a drummer, a marketing professional, a snowboard enthusiast, a father, and I have a mortgage. All of those interests are expressed in my overall Web searching, browsing and shopping behavior from time to time. So the question for marketers is, how do you creatively segment customers and then customize site page content, product and feature mixes accordingly?

BI: So I guess that's where the test and target piece comes into play?

Hieggelke: The conventional answer is to make a more or less educated guess and then hope you're right when the campaign rolls out. But once you make your guess, you're basically locked in. Testing and targeting technology is designed to move marketers finally beyond that level. The beauty of it is that you can constantly test your hypotheses on how behavioral segments can be matched with specific customized contents or offers. You can test how segment A responds to five or 10 different landing pages or, conversely, how five or 10 different page designs perform with different segments. The site can become a living breathing marketing laboratory, always working towards increasing the bottom line to the company.

The key value, and it's something that will be best understood in a tighter economy, is that targeting intelligence can drive more yield without spending more on media.

BI: Can you cite some examples of how internal site optimization by behavior increases yield?

Hieggelke: GMAC Mortgage used Omniture Test&Target to test the effectiveness of their existing home page against a newly designed home page. Among the new home page layout changes included a more prominent "Get Started" button, the elimination of multiple links and heavy content, new placement and enlargement of their phone number. Over the course of one month, the new layout consistently outperformed GMAC Mortgage's existing page resulting in a conversion lift of over 12%. Through this testing, the online marketing team realized that by reducing the number of elements on the homepage, the simpler layout highlighting key elements drew more attention, and ultimately, more leads. Basically, GMAC Mortgage had a great product offering, but the message was being lost on the page. GMAC Mortgage plans to continue with future campaigns testing different audience segments including new visitors versus returning visitors and targeting content to each relevant group.

The Dallas Morning News used Test&Target to compare their existing subscription flow design against a new design that featured simpler descriptions, an information collection page that would appear earlier in the sign-up process, and a cleaner presentation of subscription and billing options. Tests showed that the new funnel design resulted in a lift in conversion of 26.13%. Then, shortening the length of the subscription funnel from 4 pages to 3 resulted in a 64.29% lift. Running a multivariate test of several elements on the subscription and billing page added an additional 34.24% more in conversion lift. Finally, testing a one-page version of the subscription funnel resulted in another 11.83% lift.

BI: What's your top-of-mind challenge on the behavioral front going into the second half of 2008?

Hieggelke: The challenge moving forward is to increasingly give marketers deeper understanding about what is possible with this strategy, and increasing the amount of control over each element of their content pages. What this does is finally connect the technology side of analytics, which has been driven by engineering with the creative discipline of marketing.