Recommendations
PULSE Recommendations rank a strip against the visit happening now, so a first-time shopper gets relevant picks without any purchase history. One call, alongside the page it sits on.
Most recommendation engines need history: past purchases, a login, weeks of clicks. A first-time visitor gets the bestsellers, which is everyone's strip. PULSE ranks the strip against what this visit implies, so the first session is served as well as the hundredth.
Who it is for: storefronts where most visitors are new or anonymous, which in Southeast Asian marketplaces is most of them.
The call
Ask for a strip on the same call that ranks the page. recommendCount returns that many products,
each ranked out of what is left after the page itself.
curl -X POST "https://api.thepulseapi.com/api/v1/catalog/ranked" \
-H "Content-Type: application/json" \
-H "x-api-key: pk_live_..." \
-d '{
"sessionId": "sess_abc123",
"organizationId": "org_abc123",
"categorySlug": "headphones",
"recommendCount": 6,
"includeBreakdown": true
}'The response carries items for the page and recommendations for the strip, up to 50, each with
its score and, with includeBreakdown, the reason behind it. It ranks the catalogue you have
synced to PULSE. Reference: POST /catalog/ranked.
Counts toward usage: yes, one request for the page and its strip together.
Search
PULSE Search ranks results on what the shopper is deciding, not only on matching words. Keep your search engine and rerank its results, or let PULSE rank the products you pass in. Reads Vietnamese, Thai, Bahasa and English.
Category pages
PULSE orders category and landing-page grids for whoever arrived, so a campaign that sends a thousand people to one page shows each of them the products they came for.