June 18, 2020

Why Intelligent Search Should Come First on Your Ecommerce Site

In a recent blog, "How to Improve Your Site Search Relevancy and Customer Experience," we talked about what happens when your site search fails to provide a seamless customer experience, and users become frustrated and jump to a competitor’s site with a better experience--never to return.  Specifically, we covered the importance of offering refined filters on your company’s website search, including relevant keywords and combinations of key phrases, to meet the unique needs of your user.

Did You Know?    

•82% of the top-grossing ecommerce sites offer autocomplete suggestions to their users as they begin typing their search query.  

•BUT only 36% of websites with autocompletion (or “search autocomplete” functionality) do not offer intelligent suggestions.    

Your Website Visitors Expect Intelligent Search Suggestions  

Let’s look at another important piece of site search functionality: intelligent suggestions that display in your website search results. Amazon and Google, with their refined recommendation engines, have set the bar high for today’s user experience and this expectation carries over to ecommerce sites.  Your visitors expect a comparable experience when searching on your website. A site search that offers “search autocomplete” with functionality that offers intelligent suggestions or “recommendations” can help your visitor, prospect or customer find exactly what they need faster and easier, giving them the best website experience possible.    

16% of eCommerce websites don’t support searching by product name or model number despite these details appearing on the product page.    

Some search autocomplete functions do more harm than good. In fact, one report titled, “The Current State of Ecommerce Search, ” (based on a Baymard report) says that 82% of the top-grossing ecommerce sites offer autocomplete suggestions to their users as they begin typing their search query. However, of those websites with autocompletion (or “search autocomplete”) functionality, 36% do not offer intelligent suggestions, creating an unsatisfactory user experience.  The search autocomplete functionality in HawkSearch allows a search to be done based on variant logic. This intelligent search is built into the site search functionality, giving users the ability to easily identify (and buy) the product they want using detailed facets of information--and with the fewest possible clicks.    

How Intelligent Site Search Works Using Variant Logic    

Let’s say a customer is searching for “shoes.” They have the option to select a specific size from the size facet or by typing the search term “size 12 Nike” in the site search bar. Depending on how your data and ecommerce system is structured, HawkSearch’s variant logic will return the specific result displaying the child SKU product in relation to the parent SKU product so the customer can see the correct price and corresponding attributes (like color, etc) of those child items. If your site does not support rendering child SKUs on pages, HawkSearch can still streamline the experience by showing the specific price and image against the parent product.  This is all managed through the HawkSearch user experience. Search was meant to deliver meaningful results, unfortunately, most search offerings do not provide this functionality. HawkSearch's workbench provides the auto-complete for top categories.      

How Search Predictions Work  

When you search on search engines, the algorithm uses search predictions that give you the ability to find information quickly. This algorithm generates search predictions without human involvement and is made to show a range of information on the web. Search predictions are possible search terms related to what you’re looking for, what other people have already searched for, and are also based on factors like popularity or similarity. When you select a specific prediction, the search results are based on that term. This all sounds straightforward, but all too often many websites ignore common search recommendation issues for their users.    

Common Issues in Site Search Suggestions    

Users typically want to know the variety of search filters instead of the variety of products within a range related to a single attribute. The following list describes the kinds of irrelevant suggestions that can display when your website comes with basic autocomplete function:    

Dead-ends:  Suggestions that send the user to a page with no products.  

Duplicates:  Duplicates of the same query suggested in the autocomplete list.  

Indistinguishable results:  Suggestions with minor linguistic-only differences (like “Sofa tables” and “Sofa table”) in an autocomplete list.  

Nonsensical suggestions:  Strange or confusing suggestions, when a completely unrelated term or phrase displays.  

Category-only suggestions: Categories like dresses, shoes, or socks display but attributes like color, material or brand are not suggested.  

Control a variety of suggestions: When a set of suggestions like “dress red”, “dress blue” or “dress color” display instead of suggestions for different product attributes like material or brand.

Incompatible suggestions: When confusing, unrelated suggestions display, such as “high low high waist dresses.”  

Seasonal or temporary relevance: Temporary suggestions that are not equally relevant at all times of the year (this takes a lot of time and effort to properly manage these results).  

The problem?

Your site search engine most likely does not have the sophisticated built-in autocomplete functionality with detailed search filters. HawkSearch has this capability built into the dashboard, as shown below.    

How You Can Provide Intelligent, Relevant Suggestions via Autocomplete  

Clearly, if you are running an ecommerce site, a robust autocomplete function is a must to help your users easily identify the precise words they need to do an efficient website search. Surprisingly, 27% of sites have an autocomplete implementation that has several severe usability issues, so overall those sites perform poorly with end-users.  Here are some dos and don’ts for effective autocomplete functionality:  

Ensure your autocomplete functionality offers attribute-based filters to provide highly relevant suggestions to your website visitors. Users generally perceive your site search suggestions to be “recommendations” and will show a bias toward selecting those over their own query. If your suggestions aren’t what the user is looking for, it can create a poor website experience.

Keep your suggestions to under ten items. When your autocomplete suggestions start to exceed more than 10 items, the users typically ignore all of the suggestions (in other words, the suggestions become mere noise) or they spend too much time reading them (effectively halting their search process).

Be cautious about auto-suggesting based on other users’ past queries because it often leads to low-quality and redundant suggestions. It’s also important to regularly check that auto-suggestions don’t lead to a dead-end URL.

Query auto-completion (QAC) is one of the most prominent features of modern search engines. The list of query candidates is generated according to the prefix entered by the user in the search box and is updated with each new keystroke.  

If you are seeking to improve your website search functionality, here are essential design patterns for the Autocomplete function that can help you deliver a better user experience.    

Top Design Patterns for An Amazing Autocomplete Function  

The following design patterns have been shown in this UX research as critical to achieving a high-performing autocomplete function:  

Keep the List Manageable (mobile & desktop)

Style Auxiliary Data Differently (mobile & desktop)

Highlight the Differences, Not What Users Just Typed (mobile & desktop)

Reduce Visual Noise Within Autocomplete (desktop specific)

Highlight the Active Suggestion (desktop specific)

Provide Visual Depth (desktop specific)

Reduce Visual Competition from External Elements (mobile specific)

Style Autocomplete Suggestions for Readability (mobile specific)

Use Text Wrapping for Suggestions (mobile specific)

Provide Visual Hints that Autocomplete is Scrollable (mobile specific)

Enable Easy Removal of Queries (mobile specific)      

Discover More Ways to Improve Your Site Search

HawkSearch provides you with complete control over the display of your user's autocomplete choices so you can make sure they are seeing the products and items they seek plus the ones you want to promote. This mini guide will show you how to hide or display top product categories, popular searches, and top product matches.  Now that you know how to up-level your user experience on your website with intelligent search suggestions,  click here to learn more.  


Latest blog posts