On-site search covers both the search experience and what you learn from it. The experience includes autocomplete suggestions, typo tolerance, synonyms, filters on the results page and merchandising rules that decide which products rank first. Platforms ship a basic engine, and many stores add a dedicated search app for better relevance.
The data side is the search log: every query a shopper typed, in their own words. It shows what people expected to find, the words they use for your products and the items you do not stock at all.
Why it matters for eCommerce
Shoppers who search have told you what they want, so a failed search is a lost sale with a clear cause. Zero-result queries reveal missing synonyms (for example, “sweater” versus “jumper”), misspellings, product naming problems and demand for products you could add. The same queries also inform collection names, SEO keywords and product descriptions.
How it works in practice
In GA4, the site search option in enhanced measurement fires a view_search_results event when a results page URL contains one of five query parameters: q, s, search, query or keyword. The query is stored in the search_term parameter, which fills the Search term dimension. Shopify search URLs use q and WordPress uses s, so the default setup usually works. Searches that never load a results page, such as clicks on a predictive search dropdown, may need extra tracking. Review the top queries, zero-result queries and searches followed by an exit every week or two.
Common mistakes
- Showing a blank “no results” page with no suggestions or popular products.
- Never adding synonyms for the words customers actually use.
- Assuming searchers convert better because of search, when they simply arrive with stronger intent.
- Letting internal search result pages get indexed by search engines.