E-commerce search teams can do these things to improve relevance
1. Identify new and spiking queries that exceed a minimum threshold. Evaluate results, optimize as needed. Useful input for inventory expansion.
2. Identify, analyze and optimize failed queries. Can be those with no results, low CTR, low revenue, or other KPIs depending on your business.
3. Optimize and tune the results of your top queries. 1,000 queries is a good place to start. After an initial pass, most queries should be fine and so the amount of queries needing iterative improvement each month should be small.
4. Relevance testing. Test and look for issue patterns across your entire query stream and data. Provides baseline of current state of search so that future development and changes can be compared.
5. Test your search against your competitors. So useful to know what the competition is doing better.
6. Optimize queries for seasonal spikes. Can mean many things depending on your specific business. Keep a calendar of anticipated events and lean heavily on previous year’s query logs.