When you think about it, marketing seeks to answer the same questions that keep people glued to reality shows. What did that person do? Why did they do it? What will they do next? Your customers probably behave more rationally than the characters on reality shows, but thanks to Web-based analytics, you can develop your own long-running story lines to predict behavior. "Prior to this, marketing was a throw at the dartboard - you never knew what would stick," says Tina Bean, Director of Marketing for Visistat, a Campbell, Calif.-based developer of Web analytics tools. "The Web is the most measurable media we have."
All this is predicated, of course, on the incremental build-up of information. "We're moving toward a convergence of a lot of data," says Marshall Sponder, a Web analyst for Monster.com and a spokesman for the Web Analytics Association. "In order to answer the questions, you have to use a data warehouse, because you've got Web analytics combined with demographic data and spending data." The possibilities are almost dizzying, but so is the payoff. Just as astronomers use state-of-the-art telescopes to find undiscovered galaxies, with the right data tools, you can discern patterns that may not have been visible before.
Eric Hobein, Business Intelligence manager for Paris-based Europages, a B2B directory site, concurs. "We want to explain the number of times an ad is viewed, the content of the ad, the industry sector it relates to, even the time of the year it's viewed. Thanks to data mining software, we can get key details and produce deep analysis."
You can start collecting data for trend analysis using different options (known as A/B testing). For instance, consider a banner ad that offers two different options: a hyperlink that goes straight to your home page, with a prominently placed search box so that visitors can explore for what they want; or a hyperlink that goes directly to a page for a specific product. Which gets better results, both in terms of clicks and sales? The home page option gives you the opportunity to entice the visitor with multiple products; the landing page may be more efficient from the customers' point of view.
You do have to be careful, however, not to inject too many variables into the process. For instance, if you test the home page option, you can adjust what is displayed most prominently on the page to see if that makes a difference. You can use IP information to determine geographical specificity. For example, if it's winter and the user is from Florida, show tennis rackets; if the user is from Maine, show snowshoes. "You can set up a site to display certain things based on what's in the recipient's e-mail address," says Bean. "If the e-mail is sent to someone at a company with insurance in its name, you can dynamically load industry-specific content."
The key to success, both Hobein and Bean agree, is in not only tracking what the users do, but also in constructing and tagging your Web site properly. For example, use English descriptions of URLs so you can track the pages without guessing their content.
You can also use the information in your data warehouse to help test whether past behavior is an indicator of future behavior. For instance, if a customer starts ordering more do-it-yourself items, you may want to test a promotion for more sophisticated tools. Ultimately, your campaigns can get more targeted because you're starting with more knowledge and insight about trends and patterns. That translates to more efficiency and a better return on your marketing dollar.
 |
 |
|
« Previous Article ROI for Customer Data Integration |
