Search (image) dataset

• Label: Search (Image)
• Description: Image search analytics.
• Default group key: image_hash

​

Metrics

• requests (requests)
  • Number of image search requests.
  • Example: COUNT_DISTINCT(requests)
• results_count (results)
  • Results returned per search request.
  • Example: AVG(results_count)
• total_sales (USD)
  • Gross sales from orders traced back to an image search, credited in full to that search. Shown in the dashboard as Attributed Sales. Measures line-item price after line-item discounts, in the store’s own currency: before tax and shipping, and not reduced by refunds or cancellations.
  • Example: SUM(total_sales)
• quantity_purchased (items)
  • Number of items purchased from image search traffic.
  • Example: SUM(quantity_purchased)
• view_sessions (sessions)
  • Distinct sessions in which a product was viewed from image search results. Counts sessions, not views. Shown in the dashboard as Viewing Sessions.
• cart_sessions (sessions)
  • Sessions where a product from image search results was added to cart.
• quantity_added_to_cart (items)
  • Items added to cart from image search results.
• filtered_requests (requests)
  • Distinct image search requests where the shopper applied at least one filter. Filters injected by a request transform do not count.
  • Example: COUNT_DISTINCT(filtered_requests)
• filter_usage_rate (rate)
  • Share of image search requests with filters applied. Computed as filtered_requests / requests.
  • Example: SHOW filter_usage_rate
• filtered_zero_results (requests)
  • Filtered image search requests that returned no products.
  • Example: COUNT_DISTINCT(filtered_zero_results)
• filtered_zero_result_rate (rate)
  • Share of filtered image search requests with zero results. Computed as filtered_zero_results / filtered_requests.
  • Example: SHOW filtered_zero_result_rate

​

Dimensions

• image_hash: Hashed representation of the image query.
• num_results: Number of results returned.
• shopping_channel: The shopping channel (web or app).
• locale: Storefront locale code active for the request (e.g., en, fr, pt-br).
• currency: ISO 4217 currency code applied to the request (e.g., USD, EUR).
• session_returning: Session type — new or returning.
• customer_state: Whether the shopper is authenticated — signed_in or guest.
• customer_returning: Customer type — new or returning, based on prior order count.
• b2b_company: Shopify B2B company location identifier for the signed-in company account.
• billable: Whether this request is counted for billing.
• attribution_token: Token used to attribute downstream events/purchases.
• experiment_id: Experiment identifier.
• experiment_group: Experiment group/variant.
• device: Device category.
• os: Operating system.
• geo_country: Two-letter country code.
• geo_state: Up to three-letter province code.
• geo_city: City name as captured.
• marketing_source: UTM source.
• marketing_medium: UTM medium.
• marketing_campaign: UTM campaign.
• filter_property: The shopper-applied filter property (for example, option.color, metafield.custom.material). Transform-injected filters are excluded.
• filter_operator: The operator used for the selection (in, notIn, eq, neq, range, geo, etc.).
• filter_value: The selected value. Populated for in, notIn, eq, and neq; NULL for range and geo operators, which emit one row per request without values.
• filter_name: Display-only merchant label for the filter property, resolved from the attribute’s nickname. Group by filter_property to populate this label.

​

Examples

FROM search_image
SHOW COUNT_DISTINCT(requests)
GROUP BY geo_country
SINCE -7d

FROM search_image
SHOW SUM(total_sales)
GROUP BY device
SINCE this_month

​

Next steps

• LayersQL syntax
• LayersQL functions
• Datasets overview