How To Improve Your eCommerce Search

Search on the internet has come a long way since its inception in 1993. It has been an integral part of the internet framework and still continues to stay the same. Since there has been an increasing reliance on search, the function has also evolved to deliver better results. In fact, the success rate of search results has gone from 50% in 2001 to 83% in 2017. This has a direct effect on your website as users now expect more from your search. 


Search in eCommerce has a higher prominence as users who engage with search function have a higher conversion rate than the overall conversion rate and this difference in conversion rate some estimates show is as high as 160%. 

This points to the magnitude of change in revenue a fully optimized search function could bring in. One more point to note is the increased expectations from the users from the search function. Users are now more than ever accustomed to getting their desired outcomes. 

This can be delivered only by making use of a capable search function failing to which the user experience will take a big hit. Therefore even if your specific eCommerce site is not heavily reliant on search, there is still a need to offer a good search function as it is an expectation from the users that you do not want to let down. 

And as far as any generic eCommerce site is concerned an optimized search function is not an option to have but a basic need for the business. There are two main parts to an eCommerce website: the product and listings. The product part involves all the information related to the product like its title, image, description, or any further information specific to the product. 

Listing involves all the different sellers selling the same product and a search result in eCommerce is finally a form of a combination of product and all associated listing. It is not always pure text and contains structured information about products like price, offers associated with it, MRP, delivery time, Geo-location, etc which all play as factors in the ranking algorithm. 

This nature of ecommerce search makes it different from web-search as documents have a structure and in ecommerce, it is more of a hybrid structure. Queries are also implicit. 

For example, a search for ‘white gym shoes under rupees 1000’ must be mapped to the category shoes; subcategory sports; color = white; price <= 1000. This differentiates it from the token matching kind of method in web search.

Want to Know How to create a profitable eCommerce store from scratch? check out this article here.


Ecommerce search as mentioned above is a very interesting case. In order to get a better idea of it, it’s helpful to understand the various types of ecommerce search. Baymard Institute conducted a survey on the top 60 ecommerce sites and came up with eight different search query types found in the ecommerce domain. 

  • Exact search: In this type of search the search matches exactly to the product name. example: HP E51 laptop
  • Product Type: Here users search for a type of product rather than a specific item. Example: Laptop
  • Non-Product: Here the users are searching for information not directly related to any product. example: Refund policy
  • Symptom: Here the user types in a problem or issue experienced by him looking for a product that will solve the problem. example: Laptop overheating
  • Feature: Here the desired product feature is added in the query. example 4GB RAM Laptop
  • Thematic: Here the user is looking for a product that suits a particular occasion, season, or theme. example: Gaming Laptop
  • Compatibility: Here the user searches for a product associated with another product. example: Charger for HP E51.
  • Slang & Symbols: Here the user searches for a particular product, brand, or feature using commonly used slang. example: Power Block

Given below is the graph containing the percentage of ecommerce sites in the survey providing support to different types of queries. Ideally, an optimized search function must be able to deliver support to all 8 query types. 


  • Spellcheck: Often users get excited while searching and make typos or sometimes users may not know the exact spelling of the product. Spellcheck identifies the results closest to the letters entered in the query. This happens very often and now without this feature, a lot of potential clients who made an error are not going to make a purchase.

  • Multiple language support: You could be selling locally in a non-English speaking country or you could be shipping internationally either way having multiple language search options are going to improve the chances of users not familiar with English buying. Also having an availability of multiple language inputs is going to boost the prominence of the website in the user’s perception. 
  • Natural Language Support: The search function’s capability to process natural language can go a long way in providing relevant results. The ability to integrate plural and singular keywords, slang & synonyms are key indicators to the Natural language processing capabilities of the Search Function. People use natural language to communicate often as it is comprehensible by other people and expect the same from machines as well, the ability to deliver that expectation is going to increase the site appeal and directly influence the sale. 

  • Autocomplete: This feature anticipates the search query for the user and suggests the product being looked for before the query is completed. It makes search less painful and more intuitive. This is a feature now widely used in major eCommerce sites and hence users naturally expect this feature. Autocomplete can also often avoid spelling error which would also avoid no results coming in if the spelling is far from the right one and the spellcheck couldn’t pick it up. 

  • Ability to crawl PDF, Word, and Excel: This is an important feature to have in your search functions as some very detailed information of a particular product could be uploaded in the form of PDF or excel sheets. for example, the effects & ability of a certain pharmaceutical product to cure a symptom could be attached in the form of a PDF, and in this case, the user is putting in a symptom-based search query the particular product could be shown as a result of the crawler was able to crawl through the PDF containing the details.
  • Fast results: Waiting too long for the results to load is a major problem for the users as they are accustomed to having faster results. Making sure that the results load fast is very crucial for site performance. 
  • Facets & Filters: Facets and filters do a great job in increasing the appeal and functionality of the search function. Filters & Facets are options that give the users the controls to refine their search using product attributes and features. There has been a high reliance on these features in recent times as it helps the users find their desired products easily. 

  • Multi-Device Support: If you run an eCommerce site you would know how crucial it is to have your website optimized for mobile devices. The same goes for the search function as well. The search function must be optimized for mobile devices to enable clearer viewing of details in search results and make the search function more friendly to the mobile UI. This often gets tricky for mobile devices as there is less space to incorporate a search function which can be functioned and viewed efficiently. 
  • Customization & Visibility: The search bar must be clearly visible to your user and must set in well with your website theme. This can be achieved when there is an option to customize the search bar and search result appearance.
  • Merchandising Option: Merchandising option allows you to manipulate the search algorithm in different ways for example assigning different weights for different product attributes or excluding particular products from the search algorithm. This is a feature that enables you to maximize your revenue from the site. 
  • Zero Results Page: After all the optimizations there is still a possibility for a particular search query to return no results so it’s best to be prepared rather than be caught off guard. Zero results page must ideally offer an explanation as to why the search didn’t return results, apologize, and provide relevant suggestions if possible. A good zero results page goes a long way in making your users who did not get any search results stick around. 


Now that you have a clear idea of what are the things you need to look out for while optimizing your search it is time to evaluate your options to get it done. Naturally, there are two options one is to get it done in-house the other option is to outsource it. Each option has its pros and cons. You must weigh what is more important and take a call.


Building a custom site search using in-house engineers. This is often practiced in an organization with a large resource pool.


  • The high degree of customization
  • Control over all search elements 
  • No recurring subscription fees


  • High upfront cost
  • Requires engineers of high caliber
  • Constant pouring in of resource to keep search system updated


Completely outsourcing the site search function to a third party service provider. This is often done in organizations with lean teams and resources. But this trend is shifting as search optimization is getting more and more complex to deal with, many large companies are also preferring expert third party solutions. 


  • No high upfront costs
  • Updated search systems
  • Inhouse engineers can be freed up and focused elsewhere 
  • Faster setting up & fixes
  • Expertise from specialists 


  • Recurrent subscription fee


Search has a very high prominence in ecommerce, it takes your user to the desired product. Hence ignoring it is not an option. Ideally, your ecommerce website must be able to provide support to all the eight search query types that are Exact, Product, Non-product, Symptom, feature, thematic, compatibility, slangs & symbols. 

In order to provide support to all these query types your search must be optimized enough with features like spellcheck, autocomplete, NLP, multiple language support, facets & filters, ability to crawl PDF & documents. 

Having customization features, Multi-device support, merchandising features and a good zero results page will give an added auxiliary support to optimize your search. 

Finally coming to the question of outsourcing, I cannot answer this question for you, go through the pros and cons of each option, assess your position, and make a wise call. Like search, there are a lot of elements on ecommerce that require your attention and time to work right. 

If you are only starting out here or haven’t had much success then I suggest you go through a step by step guide to optimize ecommerce from scratch. 

 Happy Selling folks!

Guest Post by Expertrec

Expertrec develops a cloud-based custom search engine specialized for e-commerce.

About the author

Ra Karthik

Ramachandiran Karthik is the Digital Marketing Manager at Adnabu. He loves to help Shopify stores grow their business with Google Shopping & Google Ads, using advanced features like multi-currencies, multi-languages, Metafields. Karthik has a lot of experience in digital marketing and enjoys exploring new ways to make advertising more effective for both advertisers and customers.

By Ra Karthik

New call-to-action

New call-to-action

New call-to-action

New call-to-action