Search

Site search tracking is one of the most undervalued analytics implementations. Search queries tell you exactly what users are looking for — in their own words. Zero-result queries are a product roadmap. High-volume searches for out-of-stock products are revenue leaking out of your catalog.

GA4 has a recommended search event. Use it.

When to fire

Fire search when:

• The user submits a search query and results load
• Autocomplete suggestions are shown (optional — see below)
• Search filters are changed on a results page

Fire once per search action, not once per character typed. If your search has a submit button, fire on submit. If it triggers on Enter key, fire on keypress. If it’s an instant-search that queries on every keystroke, fire after a debounce — not on every character change.

Basic search event

// User searches for "leather jacket"
dataLayer.push({
  event: 'search',
  search_term: 'leather jacket'
});

Extended search event with additional context

dataLayer.push({
  event: 'search',
  search_term: 'leather jacket',
  search_results_count: 24,         // number of results returned
  search_type: 'full_text',         // full_text, autocomplete, filter_change
  search_category: 'all',           // which category was searched within
  search_filters_applied: false,    // were filters pre-applied?
  search_sort: 'relevance'          // how results are sorted
});

Event schema

Event Schema search

Parameter	Type	Required	Description
event	string	Required	Must be "search"
search_term	string	Required	The query string the user typed. This is the key parameter — always include it.
search_results_count	number	Optional	Number of results returned. Enables zero-result query analysis.
search_type	string	Optional	Type of search: full_text, autocomplete, barcode_scan, voice.
search_category	string	Optional	Category context if the search was scoped to a category.
search_filters_applied	boolean	Optional	Whether any filters were applied to the search.
search_sort	string	Optional	Current sort order: relevance, price_asc, price_desc, newest.

Tracking zero-result searches

Zero-result searches are your most valuable search data — they reveal what users want that you don’t have. Track them explicitly.

// Zero results returned
dataLayer.push({
  event: 'search',
  search_term: 'waterproof leather jacket',
  search_results_count: 0,
  search_type: 'full_text'
});

In GA4, create a custom report or exploration that filters search events where search_results_count = 0. Group by search_term. This is your product gap analysis.

Filter change tracking

When a user applies or changes filters on a search results page, fire another search event to capture the refined query context.

// User applies a "Sale Items" filter on search results
dataLayer.push({
  event: 'search',
  search_term: 'leather jacket',   // original query — preserve it
  search_results_count: 8,
  search_type: 'filter_change',
  search_filters_applied: true,
  active_filters: 'on_sale,brand:Heritage Co.'   // optional: which filters
});

Autocomplete selection

If you have an autocomplete/typeahead feature, consider tracking when users select a suggestion vs. type and submit manually. This helps you evaluate how useful your autocomplete is.

// User selects an autocomplete suggestion
dataLayer.push({
  event: 'search',
  search_term: 'leather jacket',   // the selected suggestion
  search_type: 'autocomplete',
  search_results_count: 24
});

Implementation patterns

Vanilla JavaScript

const searchForm = document.querySelector('#search-form');
const resultsContainer = document.querySelector('#search-results');

searchForm.addEventListener('submit', function(e) {
  const query = this.querySelector('input[name="q"]').value.trim();
  if (!query) return;

  // Fire immediately on submit — results count not yet known
  dataLayer.push({
    event: 'search',
    search_term: query
  });

  // Or wait for results and fire with count:
  fetchSearchResults(query).then(results => {
    dataLayer.push({
      event: 'search',
      search_term: query,
      search_results_count: results.total
    });
  });
});

React

function SearchPage({ query, results }: { query: string; results: SearchResults }) {
  useEffect(() => {
    if (!query) return;

    window.dataLayer = window.dataLayer || [];
    window.dataLayer.push({
      event: 'search',
      search_term: query,
      search_results_count: results.total,
      search_type: 'full_text'
    });
  }, [query]);  // Re-fires when query changes

  return (/* ... */);
}

GTM configuration

1. Create a Custom Event trigger for search.
2. Create Data Layer Variables for search_term, search_results_count, and any other parameters you push.
3. Create a GA4 Event tag for search with search_term mapped as an event parameter. In GA4, register search_term as a custom dimension if you want to filter reports by it.

Note

GA4’s Enhanced Measurement includes site search tracking, but it works by reading URL parameters (e.g., ?q=leather+jacket). This only works if your search uses URL parameters and if the parameter name is configured in GA4. For AJAX search that doesn’t change the URL, you must use dataLayer events.

Common mistakes

Not stripping whitespace from search terms. Leading and trailing spaces create spurious variations. 'leather jacket' and ' leather jacket' are different dimension values in GA4 even though they’re the same search intent. Trim search terms before pushing.

Using different casing. 'Leather Jacket' and 'leather jacket' are different values. Normalize to lowercase unless case is semantically meaningful.

Firing on every keystroke. Instant-search products fire as users type. Debounce with a 500–800ms delay before pushing, or fire only on actual result-load events.

Content Engagement Tracking what users do after they find content — article reads, video plays.

view_item_list Tracking product list impressions — the ecommerce complement to search tracking.