Web Analytics, a Life and Death Plot

The other day, my friend Chris pointed me to a new movie trailer. Sony Pictures is preparing to release a Web analytics movie starring Diane Lane, with a $57 million budget! Well kind of…

Actually, “Untraceable,” which is set to be released nationally next week, is about a serial killer who triggers (no pun intended) actions based on site visits and behavior. The killer kills his victims, shown live in streaming video, based on how many people log on to watch the killing. The more people who log on, the quicker the victim dies.

The idea is online behaviors, in this case visits, trigger other actions. Yes, the movie’s premise is far-fetched, but acting in real time based on visitors’ behaviors on your site isn’t.

How do you respond to different visitor behaviors on your site? Do you customize or change the experience? Do you make offline changes based on online behaviors? Most companies don’t, but a few are starting to see the power of doing so.

A few examples:

• Behavioral targeting. Using tools like Omniture’s TouchClarity, you can tune the site experience based on different visitors’ behaviors. For example, the system learns which is version of a promo to show based on past visitor experiences, where visitors came from, what they’ve clicked on, and so on. Site content adjusts based on visitor actions.
• A/B and multivariate testing. This is an obvious one. If you’re undertaking a big campaign, start the campaign by running three versions of the landing page (or many more, if you’re doing multivariate testing). The traffic is initially split equally among these versions. Based on what converts best against specific goals, roll out the winner to all campaign traffic. Or to 90 percent of campaign traffic while still testing the other two versions on 10 percent of traffic to see what works. Again, you’re making immediate decisions based on visitors’ behavior.
• Online advertising. No question, this is the most mature of the four examples. For a number of years, companies have been testing different creative banners or placements on the fly and automatically tuning them based on performance. As most larger advertisers know, this can be very beneficial, but it still isn’t done on the site itself as it should be.
• Offline changes. Think of a large computer retailer that sells online, in stores, and through catalogs and call centers. It can test out new products, configurations, and pricing models quickly online and see what people respond to, then roll that insight offline to impact its next catalog. It could test online a promo for two different bundles that come with a new computer, seeing which performs better and using that in its next catalog. While it can also test different versions of the catalog, online testing is quick and efficient and can help it try new, riskier ideas.

There are better ways to take real-time or near real-time actions (online and off-) based on your Web traffic than those in “Untraceable.” But too few companies are really leveraging them, despite their power. While not life in death, they can have a significant impact on your business.

One last thing on the movie. As a Web analytics professional, I was really disappointed to see the word “hits” used in the trailer. That word will never die!