Live Search uses events to power search algorithms such as “Most Viewed”, and “Viewed This, Viewed That”. While LUMA users get eventing out of the box, headless and other custom implementations have to implement eventing for their own needs.
Since Live Search and Product Recommendations use the same backend algorithm, some events are shared by both services. Some Product Recommendations events are required to populate the Recommendations Dashboard.
This table describes the events used by Live Search strategies.
Strategy | Products | Events | Page |
---|---|---|---|
Most Viewed | Live Search Product Recs |
page view product view |
Product detail page |
Most Purchased | Live Search Product Recs |
page view complete checkout |
Cart/Checkout |
Most added to cart | Live Search Product Recs |
page view add to cart |
Product detail page Product listing page Cart Wish List |
Viewed this, viewed that | Live Search Product Recs |
page view product view |
Product detail page |
Trending | Live Search Product Recs |
page view product view |
Product detail page |
Viewed this, bought that | Product Recs | page view product view |
Product detail page Cart/Checkout |
Bought this, bought that | Product Recs | page view product view |
Product detail page |
Conversion: View to purchase | Product Recs | page view product view |
Product detail page |
Conversion: View to purchase | Product Recs | page view complete checkout |
Cart/Checkout |
Conversion: View to cart | Product Recs | page view product view |
Product detail page |
Conversion: View to cart | Product Recs | page view add to cart |
Product detail page Product listing page Cart Wishlist |
Data collection for the purposes of Live Search does not include personally identifiable information (PII). All user identifiers, such as cookie IDs and IP addresses, are strictly anonymized. Learn more.
Some events are required to populate the Live Search dashboard
Dashboard area | Events |
---|---|
Unique searches | search-request-sent ,search-response-received |
Zero results searches | search-request-sent ,search-response-received |
Zero results rate | search-request-sent ,search-response-received |
Popular searches | search-request-sent ,search-response-received |
Avg. click position | search-request-sent ,search-response-received , search-results-view , search-product-click |
Click-through rate | search-request-sent ,search-response-received , search-results-view , search-product-click |
Conversion rate | search-request-sent ,search-response-received , search-results-view , search-product-click ,product-view ,add-to-cart ,place-order |
All events require the Page
and Storefront
contexts. This should happen at the page level/storefront application layer rather than when generating individual events (for example, in a PHP storefront, the PHP application container is responsible for setting them at runtime).
Here is a sample implementation of the search-request-sent
event:
const mse = window.magentoStorefrontEvents;
/* set in application container */
// mse.context.page(pageCtx);
// mse.context.setStorefrontInstance(storefrontCtx);
/* set before firing event */
mse.context.setSearchInput(searchInputCtx);
mse.publish.searchRequestSent("search-bar");
Ad blockers and privacy settings can prevent events from being captured and might cause the engagement and revenue metrics to be under-reported.
Eventing does not capture every transaction happening on the merchant’s site. Eventing is meant to give the merchant a general idea of events that are happening on the site.
Headless implementations must implement eventing to power the Product Recommendations dashboard.
If Cookie Restriction Mode is enabled, Adobe Commerce does not collect behavioral data until the shopper consents to using cookies. If Cookie Restriction Mode is disabled, Adobe Commerce collects behavioral data by default.