eCommerce Best Practices Site Search Blog How Text Search Works on eCommerce Sites Sarah Hillel August 30, 2020 6 Min Read From verifying facts to planning for a trip, text search plays a key role in helping users find what they’re looking for on the vast World Wide Web. In eCommerce, it’s one of the primary means for shoppers to find and ultimately purchase products. The level of sophistication of text search on eCommerce sites has an impact the customer experience. In this post, we’re uncovering how text search works, as well as the challenges and opportunities in eCommerce product search. What is Free Text Search and How Does It Work? Free text search is a technique that searches documents, records, and databases containing or matching one or more words entered by users. It’s the type of text retrieval that is the foundation of all internet search engines. Almost every application comes with a free text search tool. However, there are variations in capabilities. The most common ones are metadata search and full-text search. The former analyzes only parts or the description of documents, while the latter goes through all the words in documents. Textual search on eCommerce sites usually comes as a built-in tool and features basic text-matching based on product tags. This includes indexing of product titles, descriptions, and category structure, recognizing plural queries, correcting spelling mistakes, and supporting synonyms included in the tags. When consumers type in one or more words in the free text search tool, the product results should be direct or close matches to the search terms used. However, there’s a common problem: the lack of relevance in results. What Are the Challenges of Text Search? On average, Econsultancy found that 30% of eCommerce visitors use on-site search. Shoppers who use the search bar usually have higher intent, since they know what they’re looking for. Moreover, an Episerver survey revealed that when consumers are actively searching for products, conversion rates are significantly higher compared with when they’re browsing social media or being exposed to display ads. When eCommerce product search delivers relevant and accurate results, it enhances the online shopping experience. Not only because it leads consumers to the products they want to buy, but also and more importantly, it shows that brands understand what and how shoppers think. Unfortunately, most text search engines on eCommerce sites are still not robust enough to provide what shoppers want and need. Poor Support for Essential Keyword Searches According to Baymard Institute’s large-scale usability study of eCommerce search conducted in 2014, 2017, and 2019, there’s still a “surprisingly poor support” for common and essential textual query types. Among 60 top-grossing eCommerce sites in the United States and Europe, the latest round of the study found the following: 61% of sites require users to search by the exact product type jargon the site uses, e.g. failing to return relevant products for a search such as “blow dryer” if “hair dryer” is used on the site, or “multifunction printer” vs. “all-in-one printer”46% don’t support thematic search queries such as “spring jacket” or “office chair”32% don’t support symbols and abbreviations for even the most basic units, resulting in users missing out on perfectly relevant products if searching for inch when the site has used “ or in in their product data.27% won’t yield useful results if users misspell a single character in a product title25% don’t support non-product search queries, like “returns” or “order tracking” Poor Overall Search Experience Beyond the initial text search, other functions related to on-site search contribute to and support the overall shopping experience. Brands with an online presence need to solve the following website and text search challenges to reap the benefits of a seamless search experience: Low-visibility search boxes (e.g. hidden under a hamburger menu on mobile)Lack of support for typos, errors, or common keyword synonymsNonstandard presentation of results (pagination, sorting, filtering)Poorly executed filters (irrelevant attributes, poor functionality, empty result sets) In other words, there’s still a huge mismatch between how consumers search for products and the way textual search processes the request and delivers results. Fortunately, recent developments in technology such as machine learning, natural language processing (NLP), and more are pushing the performance of text search more and more toward what shoppers expect. Why Better eCommerce Product Search Matters and How to Move Forward Search engines such as Google and Bing, and online marketplaces such as Amazon and eBay are creating high expectations for convenience that consumers bring to any platform or channel they visit—including eCommerce sites. But it makes sense. The success of eCommerce product search depends largely on the relevance, speed, and accuracy of search results. How brands understand and support shoppers’ search behavior can make or break the customer experience. To match ever-changing consumer search and shopping habits, it’s critical for brands to have an eCommerce product search engine that can automatically and continuously optimize results based on context and intent. Capabilities should, at minimum, include the following: Real-time search bar suggestionsLinguistic analysisGeo-location detectionAutocompleteSearch previewsText query and traffic analysis Innovative Text Search for Contemporary Shoppers Before exploring new features, it’s important to address the foundational issue of poor textual search experiences. That is, a growing database of inconsistent product tags that are costly and time-consuming to input and sustain. This manual, outdated, and traditional infrastructure of eCommerce product search is not going to cut it for today’s more dynamic and fast-paced shoppers and retail landscape. With visual AI, brands can augment textual search by enhancing product tags with AI-powered deep tagging. This not only automatically enhances product information with image-based tags from thousands of vertical-specific attributes and their synonyms, but with the help of natural language processing (NLP) it also translates and recognizes the language and context that shoppers use. The added layers of product details and NLP ensure that consumers see the most accurate and relevant products even if they search for keywords that don’t match with the catalog descriptions on-site. When brands provide exactly what consumers are searching for, they can make sales right away or activate shopping journeys that are frictionless and engaging. This ultimately boosts conversion and search-generated revenue while providing shoppers with a positive and memorable customer experience.