Measure Internal Searches

Sites that have deployed an internal or on-site search application have a great opportunity to leverage their web measurement application to learn a great deal about their visitors.

Think about the last time you went to a site and had trouble finding what you wanted. Did you try to locate the product or information by navigating through the site, or did you go straight to the on-site search? If you tried the on-site search, did you get the results you were expecting immediately, or did you have to search more than once?

Far too often, both the navigation and the on-site search fail to deliver, resulting in frustrated visitors who bail out–clearly not a desired behavior. They may even leave for a competitor’s site and feel that you have wasted their time, neither of which benefits your business or your brand. This is unfortunate because if you’ve deployed internal search technology, you have a powerful ally that, properly maintained, can help you delight and convert web visitors.

Understanding How Visitors Search Your Site

The following are a set of metrics to evaluate the performance of on-site search tools. Understanding these measurements will help you improve the effectiveness of your search functionality, which can improve overall conversion rates and the success of the site.

Percentage of visits using search.

Do people automatically start with the onsite search when they land at the site, or do they move through the site and rely on search when they can’t find what they want? Some sites will record only a small percentage of search use, while others see a vast majority of visitors relying on search. Calculated as the number of visits seeing at least one “search results” page divided by the total number of visits, this percentage can help determine how high of a priority optimizing your on-site search should be. Once you review the other metrics below and see how your search is reporting, it may become a high priority to optimize it.

Searches per search visit.

This is a measure of the number of distinct searches an average visitor conducts. The ideal is one search per search visit, meaning people search once and find what they want immediately. Unfortunately, many sites record much more than one, often three or more searches per search visit. This is calculated by dividing the number of distinct searches by the number of search visits.

In general, the greater the number of searches executed in a visit, the greater the confusion on the part of the visitor. If you have to keep changing the words you search for because you’re not getting the results you’re looking for, eventually frustration sets in and you leave unsatisfied.

If you are using page tagging to track pages through your web measurement tool, each time someone explores one of the links on the results page and then returns to that results page, it may look like another search. In this case, you want to be careful that you’re measuring distinct search terms, not just additional page views for the same search term.

Percentage of exits from the search return page.

One indication of a failed search is when a visitor exits the site from the search return page, the page that lists the search results. Ineffective searches can be frustrating if visitors follow a number of links and still don’t find what they want. They may simply exit the site from the results page. If visitors find what they want, they’ll link off the results page and continue the visit. This can often be an alarming metric to view if you have not reviewed it before. Think of how those people feel about your site as they move on to a competitor.

Conversion of search visits to sales, leads, or other desired conversion.

It’s important to understand how your search visitors convert on key metrics, such as sales and lead generation [Hack #39] , compared to non-search users. Depending on the difference, you may want to point more or fewer people to the on-site search. If search visits convert at a higher rate, it may be worth analyzing the content that visitors see when coming through search pages. You can then use that information to drive the non-searchers to that content. You may have hidden the content that really drives conversion.

Average items per order for search visits versus non-search visits.

Like the conversion above, you may see a difference in the average number of items per order or the average order value. Again, you can adjust how and where the content is being presented to either group.

Percentage of searches with no results (“zero results” searches).

What percentage of visitors use on-site search and get no results? What are they looking for? What do they do when they don’t get any results on the search return page? This is another big one, and obviously has a significant impact on those search return exits.

The recommended strategy to deal with zero results searches [Hack #65] is to carefully examine the search terms in question. Often you’ll realize that you should have returned results and that something is wrong with your search index. At other times, you’ll learn that searchers are using different language to look for information than you use. In the latter case, you should be able to modify your search index or pages to generate results for those searches, taking advantage of the information your visitors are freely giving you.

Percentage of searches that generate no clicked links (“zero yield” searches).

Often, searchers will get a list of results that may or may not be correct, but for whatever reason, the list doesn’t look right. In this case, it is not uncommon that a visitor will not click any link and either back up and search again—driving up your searches per search visit metric—exit your site, or otherwise use your navigation. In this case, the search has failed your visitor, so the list of zero yield searches should be carefully examined looking for patterns or problems.

Top search terms.

What do people search for? Understanding the most common search terms can help you determine why visitors come to your site. With this information, you can improve navigation or call-outs so people don’t need to rely as heavily on the search. Make it easier for visitors to find what they’re looking for on your site. Keep in mind that how and what visitors are searching for on your site will change over time and even seasonally. It is imperative to monitor this on an on-going basis. This also gives you a good indication of what people are most interested in when they come to the site. This offers more opportunities to improve not just your on-search, but also the site overall.

As you can see, these metrics can tell you a great deal about not only the search portion of the site, but the entire site. You can gain insight into the priorities and desires of your site visitors and learn more about the motivators that get them to perform your desired behaviors.

—Jason Burby and Eric T. Peterson

Get Web Site Measurement Hacks now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.