Site Search Analytics - Pattern Analysis to Improve Your Site

Your site search data, the phrases your website users type into your internal site search engine, is data that is swimming with insights into helping to make satisfied customers with your web site.  If you are someone that is responsible for the performance of a site, this is most likely information that you've never looked at and may have not even known existed.  But, you're going to want to become familiar with it because it's about the best place you can go online to learn what your users want.  Read the first post in this series for why.

Let's start with a major analysis type that we can use with this data to improve performance, pattern analysis.

Pattern Analysis

With this type of analysis, you look at what your queries have in common or what's odd about them.  Basically, as the name suggests, any patterns that you can identify when looking at the main phrases being used by your visitors.  If you study these groups, you should be able to gain insights into the content your visitors want and need.  Take a look at these search terms for a minute or two (click for enlarged version)...

In looking at the queries, you can clearly tell this site is in the medical field.  What kinds of interesting patterns do you see?  Here are a couple of my observations...

• the top query "enter your search term here."  It looks like many visitors are just hitting the search button without entering a query.  Why?  If not, is there another technical reason we're missing here?
• 3 of the top 25 queries are forms of "log in."  Are people having trouble finding where to login?  We can even throw #24 in this group - "patient portal."  Are patients the ones trying to log in?
• 2 of the top 25 are "ipad" and "mobile."  Looks like people are going to the site looking for apps.  Are they finding them.  Why are they searching instead of just navigating there?
• 2 of these terms are forms of "system requirements."
• "demo, training, certification, support" are all queries looking for assistance with the product.

I could keep going, but this gives you an idea of the different types of groupings that you will come up with if you just observe what the queries are.  Once you do this, you will start to see patterns emerge that will give insight into what's going on with your visitors.  This will lead you to ask a lot of questions that will start you down the road to finding answers to why searchers are acting the way they are.  That's where the gold is.

Here are some of the types of patterns you should keep your eyes open for that will help improve your site...

1. Tonal - How does the language of your visitors and your content match up?
2. Synonyms - Grouping synonyms allows you to see how popular different subjects are with your visitors and the kinds of words they use to find what they're looking for.
3. Time-based - You can look at what's important and when to deliver the right content at the right time.  Time plays a huge role in shaping searcher's needs.
4. Questions - You can cluster queries by the types of questions that are embedded in them to see what is on searcher's minds and how you are delivering it to them.
5. Answers - You can do the same with types of content they want to find that they're implying in their queries.

Sometimes thing will jump out at you immediately after you identify a pattern and other times you might have to play with the data a bit.  After you've established patterns and have identified the questions that they spark, it's time to seek out those in your organization that can answer the questions about visitor intent thoroughly.  Those answers will be a helpful guide to improving your web site performance.