Top 10 ways to test internal search

Posted on December 2, 2021
7 min read


Designing experiences that meet people’s expectations is typically a high priority. And when it comes to finding things on a website, there’s no exception. People have grown accustomed to companies like Google, Amazon, and Netflix who provide exceptional search experiences by getting people to the search results they need fast and without friction. Nevertheless, many other businesses include internal search features on their websites that, by comparison, really don’t meet the mark (and expectations of many consumers). 

In this post, we’ll uncover how you can test your internal site search to make sure it's optimized for best results. But first, we’ll cover the basics. 

What is internal search?

Internal search is an optional feature on websites that allows a user to look for content within a site by typing a keyword or phrase into a search box.

definition of internal search

You’ve seen them. They often come in the shape of a small box commonly located in the upper right corner of a website. It’s usually accompanied by a magnifying glass. Despite its design—which can vary depending on a website’s style—the functionality of that feature is the primary focus of this post.

While the concept of an internal search may be simple, ensuring that it works in line with your users’ expectations is of the utmost importance.

Why is internal search important?

The internal search function on your website shouldn’t be reserved as an afterthought. According to Econsultancy, site searchers account for up to 14% of all revenue—which means a poor experience will impact your bottom line. An optimized and fully-functioning internal search can provide your business significant benefits and influence revenue.

Internal search is important because it:

  • Improves UX for site visitors
  • Improves content discoverability
  • Uncovers customer intent

How does site search improve UX for visitors?

Think about the last time you searched for something and the results page was nothing but a long list of unprioritized results. If it was a shopping experience, odds are you were frustrated that your item didn’t display, or worse, you assumed they didn’t have it. The possibilities are endless, but your opportunity to provide a positive search experience isn’t.

When designing your search feature, consider the following to provide a good user experience:

  • How and where is my search displayed?
  • What filters or categories do I provide?
  • What features, like autocomplete, suggestions, spellcheck, etc., are available to my user?
  • Are the search results relevant for most queries?

Answering these questions will start you down the right path, and prepare you for testing opportunities down the line.

How does site search improve content discoverability?

As your website grows with each new press release, blog, solutions page, or ebook, finding specific pieces of content will become more challenging for users. Some things may be easy to find by using the website's navigation, but all too often there’s useful, relevant content buried beneath too many clicks. This is where site search can help.

Having a well-oiled internal site search maximizes the chance that a user will find content related to their search terms, and in many cases improve the possibility of conversion.

How does site search uncover customer intent?

Everyone wants to understand their website visitors better. Through analytics, you can understand their behavior and over time narrow in on their interests, but with site search functionality you can enjoy a direct line to this sort of information. 

The logic behind it is pretty simple. When someone is unsure of what to do or where to go next,  they often turn to search. By this very nature, you and your business gain access to the direct expression of what your users are looking for in an exact moment and can uncover strategic insights. 

Now that we’ve got your brain jogging through all the potential benefits of a successful site search, let’s explore the ways you can test it to make sure it’s meeting the expectations for your users. 

10 ways to test internal search

1. Make your scenarios open-ended but specific 

Sometimes asking a user to search for a particular phrase is helpful, but usually, you want to have a scenario that is open-ended enough that users will use their own mental model and linguistic phrases. In other words, you don’t want them to use any supplied language that might set your site up for success. You want to find problems (or at least the most genuine user experience), so let users do their own thing.

Here’s an example of a good test question that’s open-ended, but specific:

  • Using search, try to find a pair of shoes in a color, style, and size that appeals to you.

Here’s an example of a bad test question that feeds users specific information that might be  localized to your site:

  • Search for “brown dockers.”

Here’s an example of a good test question that’s open-ended, but not specific.

  • Search for shoes.

2. Test a wide array of visitor demographics

To continue our shoe example, your site may perform great for someone searching for women’s shoes, but not for men’s. Depending on gender, how price-conscious a customer is, their age, or computer savviness, a user may use vastly different search keywords and filtering behavior to find the same product. 

This is why it’s important to choose the demographics of your panel in order to test all of the different possibilities. Large panels, like the UserTesting Contributor Network, make it easy to find relevant participants for your study.

Imagine you ask a male to find brown dress shoes. It didn’t occur to him that he needed to type "men’s dress shoes," so the first three result pages were filled exclusively with women’s shoes.  If there’s no readily apparent, above-the-fold option to filter for men’s shoes, he may assume that this site isn’t for him and abandon his purchase journey. 

3. Test ambiguous terms or terms that could have multiple meanings

When it comes to language, words can be tricky. Sometimes phrases are specific to local regions or maybe they have multiple meanings. You’ll want to test your search to see how well it can parse out results for search terms that might be confusing. 

4. Test non-product keywords phrases

If your website is for ecommerce, the obvious search queries might include product names, product types, and other product-related keywords, however, you’d be remiss to stop your site search testing there. Frequently, people may use your site search to find things like return policies, customer service, or contact information. 

With that being said, you’ll want to test phrases like “returns” to make sure that your non-product pages are able to be found via search. 

5. Test sequential searches across multiple departments or categories

From experience, we’ve discovered that people don’t usually select a department before searching for a product or product type. However, if they do choose a category, one of the biggest problems they have is getting trapped in the category and not being able to get good results, especially if category/department names are ambiguous.

Here’s a scenario that may arise. Imagine someone is looking for a gift for their mother. They choose the “Home and Gifts” category within the site navigation then opt to search for book-related topics, but there are no book results. While you may sell books on your website, you don’t sell them within this category. 

6. Test product, brand not offered, discontinued

Sometimes people will search your website for products or content that you don’t offer. In this scenario, you want to test your internal site search to ensure that you’re effectively offering relevant alternatives to their search.

This is important for customer and user experience. You don’t want to force-feed irrelevant content, but if you let the user know you don’t offer this item or piece of content but think they may enjoy these alternatives you won’t leave them wondering if they're simply unable to find it on your website. 

7. Carefully test words your internal search engine may stem

Stemming is a process that reduces words to their stem, base, or root form. In other words, your site search may use the stem of each queried term to expand the query by searching for the original term and also indexed terms that share the same root. 

For example, imagine you’re searching for a degree program in Accounting. Accounting is very different than Accounts, but it is the same as accountancy. Finding something as simple as the accounting degree on a college website should be easy, but if your site search is stemming keywords that it shouldn’t you may run into issues.

8. Test universal search results

Universal search is a search that allows your website to show images, reviews, rich snippets, and more right on the results page. Make sure you’re testing your universal search results.

You wouldn’t want an image of a plunger to display instead of an image of a pogo stick. 

9. Test internal search results frequently

Algorithms change, new products are introduced, special events and promos come and go, and users' expectations shift.

This is your simple reminder to test search results frequently—even if they’ve turned up the correct results in the past.

10. Test queries that have a high exit rate or a high refinement rate

In Google Analytics, and other website analytics platforms, you can view exit rate and refinement rate by keyword. Almost every keyword with an exit rate above your average site exit rate is a query that can be tested and optimized.

Test your internal site search diligently and often

Internal site search is not only integral for driving site conversion goals but a critical element for driving positive customer experiences. Nevertheless, site search quality is not a problem to be solved, it’s a never-ending pursuit. Regular and rigorous usability testing of search can help you meet and exceed your customers’ expectations, raising value per visit in the short term, and greatly increase the potential revenue per customer in the long term.

In this Article