Our client is a eCommerce store which has catalog of about 300,000 products from different vendors. The website is based on .Net and uses MS SQL Server as RDBMS. The website was using MS SQL full text and LIKE to find matching products. For navigation based on different product attributes and categories GROUP BY queries were being used. These queries were slow and were putting a lot on load on servers. The search results were inaccurate and were heavy on over all website performance. For users, the search was most convenient way to reach to desired products but the results were not accurate so conversions from search were low. Client reached to us for help in improving the search results and speed of the search. We recommended Solr 6.x to be used as search tool instead of relying on SQL Server. Our team designed and implemented the Solr part and the client’s development team integrated it using Solr.Net library. Solr Data Import Handler was used to import all products data from SQL Server to Solr. Delta import was also implemented so any updates in products are reflected in search immediately. Once Solr was integrated for search the improvements were immediately visible. The accuracy improved along with the performance of search. Solr facts made it easy to provide attribute and category based navigation with ease. After initial implementation we did multiple reviews of the results with customer. We introduced synonyms and tuned relevancy using boost and query modifications. The final out come of this project were accurate search results, increased conversions and increased revenue for customer. You can see the search in action here.


Waseem is consultant for Elastic Stack. He is Elastic Certified Engineer. Has years of experience with Elasticsearch, Solr, Wazuh, Sphinx Search, Manticore Search, OpenSearch and full text searching.