Using Web Analytics And On‑site Search

January 3, 2006

I often compare web analytics to the scale in my bathroom. If I get on every day and weigh myself, but don’t go on a diet, what makes me think that I will lose weight? Ditto with web analytics — why would anyone measure just for the sake of measurement? We measure so that we can make decisions based on the data.

One of the often overlooked analytics areas is on-site search. On-site search appeals to the kind of individual who walks into a department store and looks around to find a salesperson. “Where do I find the Petites department?” she asks. On your site, she types in “Petites”.

So the data from on-site search provides some insights into the kind of keywords your customers would like to use before they get to you. After all, why type in “Women’s Clothing” if you can type in “Women’s petite clothing” and get a more targeted SERP? (Surfers are getting much better about typing in multi-word phrases to get focused responses.) This leaves you now with at least two Google-type opportunities: optimize pages for “Petite” and/or purchase petite-type keywords. Women’s petite clothing. Petite clothing. Petite clothes. Petite dresses. Etc.

On-site search also creates opportunities for you to increase your conversion rate. Of course, you should start with best practices:

  • Make sure it handles misspellings and plurals.
  • Make sure that it brings back a manageable number of hits.
  • Make sure that it never says, “Sorry, we can’t find that,” but instead says, “We can’t find exactly what you want, but here are some ideas that might help.”

Once you’ve got that down, mine your analytics for the handful of terms that customers ask for month after month, and create specialized landing pages for those items — instead of a list of search items. Then put your analytics to really good use and measure those landing pages. If you have the opportunity to do A/B testing, so much the better.