Data Segmentation: Web Site Analytics for PPC

By Eric Enge, Search Engine Watch, Oct 24, 2007

Web analytics tools enable data segmentation. Slicing and dicing visitor data provides greater visibility into their behavior patterns. Armed with this knowledge, you can modify Web site and advertising campaigns to improve the results they deliver.

Personalization can be a very powerful segmentation tactic. Many Web site content management systems dynamically display content based on an incoming visitor's identity. A visitor logs in to a Web site, for example, and sees a personalized greeting (Hello, Eric!).

There are limitless ways to segment data. First, though, clarify Web site goals before deciding how to segment. Segmentation analysis should provide insights to help you more effectively and frequently achieve those goals. To make data more meaningful, you probably want to look at data over time: the last 10 Mondays, for example.

Segmentation can be conducted on visitor behavior. Even extremely simple data can lead to interesting insights. You may find Web site weekend traffic totals half the volume of weekday traffic. Further analysis may reveal whether weekend visitors are more likely to convert than their weekday counterparts.

Who would find this data meaningful? Everyone running a large PPC campaign is probably sticking their hands in the air yelling, "Pick me! Pick me!" They know they could lower their PPC bids on the weekends based on data, then raise them on weekdays. Changing bids based on site analytics data can drive extra revenue and margin dollars.

You can also segment data by campaign. What would you do if you found two campaigns underperforming on the weekend? If they're PPC campaigns, you'd further decrease bids, or perhaps even temporarily turn those campaigns off. Similarly, if they're banner ad campaigns, you may choose to not run them on weekends at all.

Let's look at another example. A company sells red widgets and blue widgets online throughout the U.S. and Canada. West of the Mississippi, there's stronger demand for red widgets. East of the Mississippi and in Canada, there's stronger demand for blue widgets.

Armed with this data, you can tweak the marketing message based on the source of the visitor. You can showcase red widgets if a visitor is from west of the Mississippi; blue widgets, east of the Mississippi or Canada.

You might also develop a special offers box to feature a product targeted to visitors from a specific geographic location. This allows you to leave most of the Web page alone, and tweak only the special offers box to optimize your results.

Summary

Visitor segmentation can bring real insight into how visitors interact with your site. What types of segmentation make sense for you may require some experimentation to find situations relevant to you. A/B split tests and multivariate testing enable more sophisticated customer segmentation and merchandising. We'll cover this topic in an upcoming column.