How E-Commerce Search Engines Handle Different Types of Queries

How E-Commerce Search Engines Handle Different Types of Queries

In 2019, the Baymard Institute published an overview of types of e-commerce on-site search queries and reported that: 

“... 61% of all sites [perform] below an acceptable search performance [and] will directly misalign with the user's actual search behavior and expectations. To make matters worse, 15% of sites were found to have a downright “broken” search query type performance.”

A pretty scathing report for the top e-commerce websites, but perhaps it’s not a surprise. Search engines have a tough job: they need to correctly parse the query and determine relevance, but it can be very challenging given the many different kinds of queries. 

Sometimes customers know exactly what they want, but more often they’re searching for possible solutions. Some of the query types Baymard identified include:

  • Exact searches — “Apple iPhone 13 Pro”
  • Feature-related searches — “Bluetooth speakers”
  • Compatibility searches — “case for iPhone”
  • Symptom searches — “something to clean iPhone screen”
  • (and more)

Search engines can struggle with such a broad array of query types. 

In this article, we will focus on how retailers can build a rich search index to aid in search success, and we’ll look at how processing and UI features at the time of a query can improve search relevance.

How search engines process queries

Product type functionality works well on many search engines.

Most e-commerce search engines do a good job of delivering relevant results for exact search queries. If users search for “Greenies” on a pet supplies site, they’re almost certainly going to find the dog and cat food treat by that brand name. 

Search engines can have a tougher time when users are searching for a type of product. A search on “chewy treats” may or may not turn up the best results. 

In the case of, our search engine would split a search for “chewy treats” into a search for “chewy AND treats” — that results would need to contain both keywords in the description or metadata. It helps the engine provide better results. Not all search engines work this way, which is why results vary between providers.

What if, rather than looking for a kind of product, visitors are searching on symptoms? People often write search terms for the problems they’re addressing. 

  • Leaky faucet
  • Migraine headache
  • Stained carpet

These kinds of searches indicate that people don’t know what specific product they need. They have a problem that they want to address. 

Results will be broad because visitors are searching for a range of possible solutions. “Migraine headache” results could include pain killers, natural remedies, cold compresses, head wraps, and more. 

Every search engine handles these kinds of queries slightly differently. assigns a relevance score to each document in your index. The score ranges from 0 (no match) to 1 (perfect match) and search results are ordered starting with the highest score. 

The scoring looks at how well the search text matches the content of the documents. This takes into account spelling, synonyms, stemming, AI based word embeddings and other language specific features. It also uses business-specific criteria to adjust rankings depending on the customer’s requirements. For example, you may want to promote high margin items at the top of the search results. 

Building a search index 

Improving search conversion rate starts with a good index.

There are many factors that improve the chances your site search returns a good result. It starts with building a rich search index.

Whether you’re using a product information management system (PIM) or writing product descriptions directly into your online store CMS (such as Shopify), you’ll need to build a robust content index that includes product attributes such as:

  • Product name or product title
  • Product data
  • Product category
  • Product description and/or product type
  • Image alt tags
  • Additional metadata such as color, size, quantity
  • Product reviews
  • Related products (this could also be inferred from the category)

Too many sites lack adequate descriptive data. This information not only enhances your product pages, but it also provides powerful cues for your search engine to crawl an index the site. 

Categorization and tagging

The process of categorizing and tagging products can improve search results and also help users who are browsing the site via category navigation. Ideally, the category page and/or tag labels should be present in the data when being indexed to provide additional metadata for the search engine. Categories can be added as metadata or even inferred by your site’s URL structure. 

For example,


Search engines can extract useful category and subcategory data (“swimwear” and/or “bikinis”) from URLs like the ones above for returning the right results from the outset. Categories are also used to build filters and facets on your results pages. 

A good product taxonomy doesn’t just aid with on-site search, it helps with SEO and web search engines. Crawlers work best when provided with structured, hierarchical data. Additional tagging (e.g., adding color as an attribute) can greatly improve results. 

Some things to consider when building rich descriptive data:

  • If you sell parts, consider adding related brands or items they would pair well with
  • For vitamin or supplement brands, call out what conditions each item helps with
  • Add color, size, or compatible products

There’s a risk of over-categorization and over-tagging. It's a balance between adding just enough to be helpful, but not so much as to affect browsing.

Non-product search queries

Visitors are looking for products, rights? Well, not always. Sometimes they’re looking for shipping information, return policies, support, shipment tracking, jobs, and more. If your site search is only indexing your products, it’s missing the proverbial forest for the trees. does a very good job with this. For example, a search for “careers at amazon” turns up both sale items and a result for their jobs page.

It’s important to optimize for non-product searches as well.

Your search provider should offer a way to do this, too, through… 

  • Indexing the whole site, not just products
  • Using machine learning learning to understand the context of the search
  • Designing search results pages to display different types of content differently


Product type synonyms can provide relevant search results.

A user could type in a search for “couch” but your site may list these kinds of products as “sofas.” 

You’ll probably know many of the synonyms that people use. We recommend you leverage your search analytics, whether it’s from Google Analytics or the metrics package included with your search platform, to see what other keywords your visitors are using in their searches.

Between having a robust synonym management system and good machine learning capabilities to develop an understanding of searchers’ intent, you can offer visitors a better search experience. 

(A similar problem is when someone types in the abbreviation “in” when they mean “inch.” Developing synonyms for each and every word in a query — especially for terms that are this nuanced — isn’t really feasible. Over time, if enough users are using abbreviations or colloquial terms in queries, an AI-powered search engine can make adjustments to deliver better converting results. More on machine learning for search in the next section.)

Search AI, data, and query understanding

So far we’ve covered the importance of improving your search index. Let’s turn attention to some of the ways you can improve search relevance at query time. 

Machine learning

AI-based search offers tremendous power through continuous and automatic improvements with intelligent feedback loops: the more data that’s produced from searches and sales, the better a search engine can improve results automatically over time.

Built-in search AI learns over time what visitors are looking for and what they’re buying. By knowing which searches lead to conversions, a search engine can automatically deliver higher converting results for similar searches. 

Soon site search engines will include vectors — a mathematical approach to representing words which encapsulate meaning of text very effectively and can deliver dramatically better results than standard keyword search. 


NLP offers user-friendly search capabilities.

Search is inherently fuzzy and language is often ambiguous because a user’s query and intent is not always apparent. “Bank” is a classic example of this — does it mean a financial institution or the side of a river? 

For some e-commerce site search use cases, customers may even type in symptoms or adjectives to find answers. Without added context, it is difficult to know exactly what they need.

Natural language processing (NLP) is one of the techniques available for improving results. NLP is the process of analyzing unstructured text to infer structure and meaning. Structure means information that is highly defined, for example, a category or a number, much like fields in a database. It can also represent relationships between things. Common examples include sizes, colors, places, names, times, entities, and intent, but there are many more. NLP is most valuable when the underlying data has a lot of structure that can be mapped from the queries.


If someone visits your site and types in “warm jacket” you can deliver very different results by user via search personalization. Generally speaking, the more data you have about someone (pages visited, purchase history, gender, age, etc.), the more you can personalize search results. 

Even if visitors are anonymous, a search for “warm jacket” can still be personalized using available data such as browser type, IP location, time of day or year, mobile vs desktop, and other attributes. 

In the example of “warm jacket,” you might promote different clothing to someone in Florida versus someone in Minnesota. While you may not know anything about the user, you can still personalize products based on their IP location.


Configuring search relevancy in

Most search services offer a feature to add rules (or even adjust the search algorithm) for handling different types of searches. 

Rules can help for difficult search queries, and they offer another way to provide results that match a retailer’s criteria for displaying products — best sellers, high or low inventory levels, the user’s location, price, merchandising, buying history, etc. 

Product search usability

Adding auto-complete or instant search to your search bar offers a great search experience and can actually help visitors type in better queries.  


Autocomplete enables users to find information faster.

Autocomplete (also called autosuggest) can help steer a visitor to the right category or product page faster and signals that your site has what they’re looking for. While autocomplete doesn’t improve the ability of a search engine to handle different types of queries, it can be massively beneficial for delivering useful results, sorting out irrelevant results faster, and offering great overall site usability. 

Instant search 

Instant search on helps customers find relevant products even faster.

Ready to take it one step further? Instant search uses the top autocomplete suggestion(s) to instantly run a query and return the top results automatically.

Users will see search results appear before their eyes — before they've even finished typing their query. Users can get to their ideal search result in as little time as possible.


E-commerce websites can have many challenges when it comes to handling different types of search queries. While you can’t read the minds of your visitors, a good search engine can be almost clairvoyant. 

Sign up today for a free 14-day trial to see how handles your visitors’ queries. Or contact us for a personalized product demo.  

Similar articles

Best Practice

Best Practices to Improve Marketplace Product Search & Discovery

Best Practice
AI / Machine Learning

The End of Product Search Synonyms

Best Practice
AI / Machine Learning

Does AI Search Impact Site Performance and Cost?