Semantic search according to Wikipedia is "Semantic search seeks to improve search accuracy by understanding the searcher's intent and the contextual meaning of terms as they appear in the searchable dataspace, whether on the Web or within a closed system, to generate more relevant results. Semantic search systems consider various points including the context of search, location, intent, a variation of words, synonyms, generalized and specialized queries, concept matching and natural language queries to provide relevant search results".
In the world of e-commerce do you need it? or at what Site Traffic do you need to think of Semantic search to get the best ROI. To get to this number, let's do some quick math evaluating the millions of search queries we process every month at Tagalys. Tagalys is a product & attribute search engine, hence anything that is semantic will not yield search results. This means every semantic search query will show in the Tagalys "No results search report" in addition to product search queries that yielded no results.
Typically less than 10 out of 100 search queries result in a "No results" message from Tagalys. Across these 8-9 queries, are searches tied to products that are not in stock, not sold by the store. Fixing these addresses about 90% of all "No search results". 1 query is usually tied to the visitor engagement like asking about their order status, promo codes, store location. This can be fixed by Tagalys Search URL redirects, which helps improve your visitor experience and may not directly translate into revenue.
At any online retailer, less than 0.5% of all search queries are semantic in nature. e.g., Dresses under $50, Rings below $999, Shoes between $75 to $150, tops for mid-30s women in the midwest, etc., and over 90% of search queries are directly tied to products or combinations of products (Dresses, Skirts, Pants, etc.,) & their attributes (Maxi dresses, Red skirts, Men's pants, etc.,). This 90% is what Tagalys helps convert to its maximum and also giving you insights into the remaining 9.5% of search queries that do not have results but can be fixed by synonyms or URL redirects.
Sites like Amazon, Walmart, etc., that capture over 10M searches a month clearly have a need for semantic search.
If you have the need for a semantic search engine, here are a few sanity checks to verify it works. Run a few sample queries using between, less than, under, etc., to test if the engine understands contextual queries. When you see the results, do a sort by price High to Low and Low to High, to verify if the engine has limited the search results, but the context requested in the query. Try this across a few different product types and prices to verify if it has not been engineered as a static identifier of the query.
Having a site search engine in your e-commerce store that yields results for semantic queries using NLP starts to become important if you are seeing over 1M search queries a month, which means roughly 5000 semantic search queries (0.5%). If your site search conversion rate is 2% as an example, you can then expect an incremental 100 orders per month from the semantic queries. And if your average order value is $50, that's $5000 in extra revenue.
Tagalys maximizes conversion rate & gives merchants visual control of products displayed in Site Search, Category Pages & Product Recommendations at their online store. To know more about our solutions and features, get in touch with us now.
.svg)
.svg)
.png)
.png)
