Tips for Building a Better User Experience With Site Search

A laptop computer showing a Google browser window.

Using the search option on certain websites can sometimes be frustrating. You can type in exactly what you are looking for, but there are often either an endless number of results or none at all. At its core, a search engine is a reference guide for all of the content and services you offer. Most users turn to search due to a breakdown in discovery.

Investing in an upgraded site search will create a genuinely elegant search experience that links the knowledge your customers need right now and recommends what's next for them in the consumer journey. Search keyword analysis assists your product team in rallying around what data points people care about and organizing that data in a manner that allows your customers to control how they find information.

For instance, Wayfair, an online retail company, identifies customers’ functional needs through keywords typed by users. When you help visitors find what they’re seeking the first time around, it has a dramatic impact on sales and engagement.

According to Forrester research, up to 43% of visitors immediately navigate to the search bar when they visit a site. And according to Algolia, up to 30% of eCommerce site visitors use the search function, and site searchers account for up to 14% of all eCommerce website revenue.

Here are some ways to improve the user experience when searching on your site:

Have a Google search box in results before users have to find it on your site.

Before users even get to your site, the first place they usually turn to is Google. When someone finds your website on a search engine results page (SERP), it’s now possible to display a search input under your URL. The greatest benefit of this is that users will be further along in the customer journey/discovery cycle by the time they land on your site—potentially skipping steps or avoiding having to find the search input several seconds after your site loads. Here’s how to format your microdata to include a site search box right from the Google SERP.

Make site search blatantly visible and in an expected place.

One of the easiest ways to improve your on-page performance is to make your site search displayed in an obvious and natural position. As a best practice, position the search box or icon glyph at the top far right, where the majority of users expect it to be.

When designing for mobile, be sure to consider the position of the keyboard and leave room for other components that may appear in the UI. A good rule of thumb for sizing a search input is that, at a minimum, it should be big enough for the tip of your thumb to interact with. One might start with 42-56px in height and a minimum of 30 characters in width.

Use auto-suggest as a context tuning mechanism.

As search engines get better at understanding the indirect meanings of queries and people discover the ease of using search engines, the experience expectation is changing. The desire for using a search to get immediate feedback as a context tuning mechanism is now a minimum requirement. Welcome auto-suggest to the search party!

Auto-suggest can help users avoid typos and, when they don’t have to type out the entire query,  reduce the total time of interaction.  It also helps users to search more easily because they will have the capability to select words or phrases from the auto-suggest feature. Rather than expecting the user to make up a query that may yield no results (or bad ones), auto-suggest allows the user to choose an acceptable phrase that will yield excellent results. One of the keys to doing this effectively is to highlight what is unique, not just what is matched from the user's input while suggesting key phrases.


Try using search history as a starting point. Before the user types one character, display the auto-suggest feature in an open state, making the recent searches visible for fast access to repeat searches.


Start auto suggestion query on keyup.

This is a way to keep the server load at a minimum while also balancing the need to be immediately responsive to users' search input. The key is to not give a minimum character count for search input before starting the auto-suggestion. Execute the search on “Keyup,” or as soon as a user enters a character.

Wait 50-100ms for the user to stop typing (this is less than a blink of an eye, but it has a dramatic effect on the overall feel). Also, be sure to enable a clear status to allow the user to quickly start again where needed. This is often executed in two forms—the long press on “delete” key to clear,  or an “X” in the search box.

Leverage non-traditional search inputs.

Accessibility is a major factor in search options for any site. Keeping in mind customers who may use assistive technologies like voice-to-text. The emergence of image search is helping users with intellectual disabilities use the web more effectively and will absolutely be used more as retail and e-commerce continue to merge. 

Never yield ‘empty’ results.

No one ever wants to see that their search has no viable results. This is your chance to introduce a new jumping off point. A simple, “Did you mean...?” will help redirect a customer to a related topic.

This can be done by introducing a list of the most popular searches and pages users have recently interacted with. Amazon saves searches with no results and tracks the click stream of the user to try to connect the dots of where the initial search failed. You can also help the user recover from typos by applying a spellcheck library as the first item in the results.

Leverage user input to optimize your personalization experience.

One of the key things to pay attention to with search is whether you are tracking what users are searching, or what users are clicking. By analyzing what your users are searching for, it informs where there are breakdowns in information discovery. This is usually secondary and tertiary tier information—in essence, content that's relevant later in the customer journey.

Another way to leverage user data is to keep track of the velocity of words—the rate at which searches for certain categories of words are increasing and/or decreasing.

Click streams can also provide insightful data about users’ time on page, bounce rate, and pogo-sticking—clicking on an organic result and then quickly returning to the SERP to choose another result. Second-search rate, or the use of words to refine a search, comes in handy here as well. For example, a first search can be “Nike” and a second search, to refine the first, can be “Nike Air Max One.” 

Use custom weighting to achieve business goals.

By using weighted individual user personalization, you can capture the user’s clickstream or searches and update your indexing on a per user basis. In its simplest form, weighting is the idea of giving higher ranking to words based on a numeric value. This can be done by:

  • Event scoring - What actions do your users make that are important to the sale of your products or services?

  • Facet scoring - What product or service categories does the user click or search for?

  • Popularity scoring - What products, services, or facets do the connections/community of user search for, like, and click?

By adding weight to your search items, one can easily change the rank to be more relevant based on seasons and any other variable you might want to introduce.

Leverage basic natural language processing (NLP).

Stemming, lemmatization, and other forms of NLP while indexing your records can make them more relatable to users’ natural way of searching.

Source: Devendra Kushwah

In essence, auto-suggestion is a form of stemming, the process of translating part of words into its non-changing parts. Lemmatization, on the other hand, is the process of converting a sentence’s words to their dictionary form. These forms of breaking down words into parts are truly the first step to tuning your index for relevancy. 

**A good starting point for technical folks:

If you find yourself asking if you need an expert to help you get set up—the answer is yes. Our experienced search & user experience experts will save you time and budget by sharing enterprise data best practices bookmatched with best-in-class visual design. 

Unlock the power of site search to get more performance out of your web experience.