Why this matters
Zero-result searches are explicit demand with no available supply. They tell marketplace teams what inventory, geography, taxonomy, or search behavior needs attention.
What you get
- Top zero-result queries by category and geography
- Result-count distribution for high-demand searches
- Whether users recover with another query
- Supply or taxonomy actions ranked by potential demand
Walk through it
You
What are people searching for that we do not have?
Agent
I’ll find searches with result_count equal to zero and group them by query, category, and geography.
POST /v1/projects/:project_id/analytics/query {
project_id: "default",
hogql: "SELECT properties.query AS query, properties.category AS category, properties.geography AS geography, count() AS searches, count(DISTINCT distinct_id) AS users FROM events WHERE event = 'marketplace_search_performed' AND toInt(properties.result_count) = 0 AND timestamp > now() - INTERVAL 30 DAY GROUP BY query, category, geography ORDER BY users DESC LIMIT 25"
}The output
The agent returns demand gaps ranked by user count and search volume. It should distinguish supply acquisition, category taxonomy, synonym handling, and geography coverage.
Setting it up
Track every marketplace search with query, normalized category, geography, and result count. Include follow-up listing views or booking starts to measure recovery.
Variations
- “Show zero-result searches by city.”
- “Which categories have demand but no supply?”
- “Publish weekly marketplace demand gaps.”