So with the information on the site coming in, the next step is to redesign the site to capitalize on what patrons want. This is called website optimization. And the best way to create a optimized web page is with a multi-variate test.

Step 1: Develop Different Pages

The strategy in a multi-variate test is to create a couple of versions of the same webpage. For example we could change where the links are located on the page or  how much (or little) information is displayed on the frontpage. Change can be as little as a single picture.

Essentially, we are testing a series of hypotheses. For example, we may notice that many patrons come to our library’s youth services page using a search for “Storytime” but the bounce rate is high suggesting that they did not notice the calendar link.


We suspect that this information is too hidden. So one version of the page may add the link higher up in the hierarchy. Another version might par down the information so that high demand content (such as event schedule) is highly visible while other information is not yet presented.

Step 2: Test

748443511_e3b89339d2So after making a few changes to different pages, upload them to Google Website Optimizer. This tool will randomly send visitors to different versions of the page. Each different version of the page will have slightly different usage statistics.

As an example, check out how Tim Ferris of 4 hour workweek used the Google Website Optimizer to improve his page.

Step 3: Analyze

So with all the information gathered for different versions of the site, it is easy to determine which one is most useful. So for our storytime example, we may have noticed that promoting the link higher in the menu decreased the bounce rate for that site and increased the number of clicks to the calendar.

However, maybe still more people clicked schedule on the minimal site suggesting that our control site was leading to a paradox of choice where too many options and too much information sent people away.

Step 4: Update

This is the easy part. With the analysis in, you can turn off the optimizer and alter your site to use the best page.

Key Points to Rememer

  • Have a goal: The key to optimizing a page is that there needs to be a limited number of factors. With a focus on a specific goal (increasing the number of people who click on the events schedule), it is much easier to design alternatives
  • It’s easy: There’s no real trick. That best optimized sites are making subtle changes. Just tiny little changes in how information is presented can have big results. And since I know librarians love books. . .Check out the book Nudge for an example of the psychology of small pushes.
  • It’s efficient: Imagine if you could have every patron that walked into the library fill out a quick survey. That is the beauty part of optimization. Since it is automatic, patrons will give lots and lots of feedback without needing to be inconvenienced.

For more information on creating a optimzation strategy, check out this white paper.

Next week, I’ll start with the ever confusing problem of Social Network analytics. And, as always, drop me a ling in the comments or at twitter @joesmorgan.

Photos available under cc license: