Leveraging Online Analytics with Audience Profiles

How to serve up dynamic content according to visitor interests.

By Bob Bly
Monday, November 9, 2009

---

The more you know about your Web site visitors, the more you can tailor online content to their interests and information needs.

And the more relevant and compelling the content the visitor sees, the greater your subscriber satisfaction, site stickiness, repeat visits, page views, conversion rates, revenues, and Web site ROI.

The most efficient method of identifying reader interests and information needs is to assign visitors to different groups according to audience profiles.

“Focus on unique visitors,” says Greg Krehbiel, director of marketing operations at Kiplinger Washington Editors, Inc. “Separate them into categories and see how your metrics vary.” He advises comparing organic versus paid, new versus returning, and visitors from different sources; e.g. text versus display ads versus emails.

Audience profiles can also be based on areas of interest as determined by which pages people view and what content they download. On a site for a food magazine, for example, these audience profile groups might include people looking for great restaurants, working moms who need to cook quick meals in limited time, gourmets, diabetics, barbecue enthusiasts, and wine collectors.

There are two basic methods of creating audience profiles. The most common is to assign visitors one or more user profiles based on their answers to surveys or questionnaires—as is done on qualification cards for controlled circulation publications.

The drawbacks of survey-based user profiles are twofold. First, most site visitors won’t bother to complete your questionnaire or survey. Unless they do, they can’t be placed within any profile.

Second, when user audience profiles are based on surveys, you are classifying visitors according to what they say they are interested in, and not by what they actually do while on your site.

Dynamic Profiling
A more effective method of targeting users is with dynamic audience profiling, in which readers are placed into profiles based on actions they take while on your site.

There are a couple of advantages to dynamic audience profiling. First, since every user on your site takes actions while there, most if not all users are quickly assigned one or more user profiles. A visitor can belong to multiple audience profiles. For instance, a food magazine reader might be interested in both dining out and collecting wine.

Second, dynamic profiling is based on what the reader’s interests are as determined by her actions on your site, and not what she says she is interested in.

“Dynamic profiling places Web site visitors into user groups based on two sets of criteria: (1) technical and demographic data and (2) triggered events,” explains Brett Zucker, chief technology officer at Bridgeline Software.

With technical and demographic data, the analytics software can assign Web site visitors to audience profiles based on such criteria as: Geographic region, new versus repeat visitor, browser, operating system, use of Java Script, Flash capable, and screen resolution.

“The analytics software can also assign audience profiles based on events trigged during the reader’s visit to the site,” says Zucker. “These actions can include viewing an article, making a purchase, downloading a premium, or submitting a form.”

Audience profile assignments can be made on any combination of these factors. In the food magazine example given earlier, you might place into the “dining out” audience profile anyone who reads a restaurant review online.

Driving Relevant Content
Assigning visitors to audience profiles lets you know more about them. But the power and effectiveness of audience profiles is multiplied when you use them to drive relevant content to each reader based on his specific interests and needs.

To drive content to Web site visitors based on their dynamically created audience profiles, your analytics package must be tightly integrated with your Web site’s content management system (CMS). A stand-alone analytics package not connected to the CMS cannot push relevant Web content to readers during their site visit.

When analytics and the CMS are fully integrated, both applications share a common data set. In addition, communication between the analytics package and the CMS takes place automatically and in real-time, shrinking what Forrester research calls “the action chasm.” This is the delay between when metrics are recorded and analyzed and when site improvements based on those reports are made.

Indexing and Page Tags
To drive persuasive content to readers in real time based on dynamically created audience profiles, a set of index terms is associated with the profile. The analytics software assigns these index terms to every reader placed into that profile.

For instance, within the analytics package, index terms assigned to readers placed within the “dining out” user profile group might include “diners,” “family restaurants,” and “five-star restaurants.”

Within the CMS, the IT department or Web administrator assigns page tags to each page on the Web site. Many of these pages contain dynamic content areas that can be varied with each page view.

“Dynamic content” refers to sections of the Web site that can be changed based on the audience profile or other data (e.g., sources of traffic). Static content refers to sections of the Web site that do not change."

Dynamic content can include: Web pages or content, text, images, banner ads, marketing messages, Flash promotions, case studies, PDF documents, related pages, menus, links, file folders, files, and sidebars.

When a reader lands on a page with a dynamic content section, the CMS serves the appropriate dynamic content by filtering based on the index terms assigned to that reader according to her audience profile.

For instance, www.savingstreet.com publishes content on buying, moving to, and improving a new home. Based on pages viewed, the analytics software places visitors into appropriate audience profiles such as “planning to move.”

As they browse the site, any dynamic sections with tags related to moving are served to them by the CMS on the appropriate pages, according to Mike Gavelek, general manager of Savingstreet.com.

For instance, when visitors placed into the “planning to move” audience profile click on the “Resources” link on the top menu bar, they are taken to a page of content where the first category choice is “Helpful Moving Articles.”

Audience Profiles and ROI
Once the analytics software assigns audience profiles to site visitors, the CMS can extract the profile assignment in real-time and drive appropriate content to those users as they are logged onto the site. The “wine collector” might be served a page with the most recent wine reviews, while the “barbecue investor” gets a page with grilling tips.

The result is a more meaningful, engaging, and relevant experience for the Web site visitor. Driving dynamic content based on audience profile can also improve content readership, e-mail alert sign-ups, forum or community registration, and purchases.

Bob Bly is a freelance copywriter specializing in direct response and online marketing. His clients include Kiplinger, McGraw-Hill, Medical Economics, and Agora Publishing. He can be found online at www.bly.com.

The Six Components to Matching Content to Audience Profiles

1. The reader is automatically assigned a profile by the analytics package based on who he is and what he does on the site.

2. That profile is associated with a set of index terms indicating the topics that people within that profile are most likely to be interested in.

3. Those index terms correspond to page tags indicating what content on the site – articles, banners, web pages, text, images, forms, links – are directly related to those index topics.

4. The analytics package automatically and instantly notifies the CMS that a given visitor belongs to one or more audience profiles.
5. The CMS reads the index terms connected to those audience profiles.

6. When the reader lands on a section of the site with dynamic content, the content most relevant to those index terms is automatically served to that user.