Recommendation Strategies Reference

Recommendation Strategies

Reference for supported recommendation strategy types and practical selection guidance.

Audience: Personalization and merchandising teams choosing strategy logic per placement.

Critical: Strategy values must match enum-backed supported values in recommendation widget payloads.

Modules: Recommendations Recommendation Breakdown Product Cards

Who This Page Is For

Use this page when deciding which recommendation strategy best fits each page context, catalog maturity, and personalization requirement.

Quick Start (2-5 Minutes)

List available strategies

Fetch strategy options from API.

GET /api/v1/recommendations/strategies

Pick strategy by context

Map page intent to strategy objective (discovery, conversion, retention).

Use decision matrix below.

Configure strategyConfig

Provide strategy-specific parameters where needed.

{ "strategy": "category_based", "strategyConfig": { "includeCategories": ["Shoes"] } }

Preview before activation

Validate result quality with preview endpoint.

POST /api/v1/recommendations/:id/site/:siteId/preview

Monitor performance

Track per-widget breakdown and funnel metrics after launch.

Use /analytics/recommendations/breakdown and /analytics/funnels/:siteId?steps=widget_impression,click,add_to_cart,purchase.

Required Fields / Minimum Payload

Field	Required	Type	Used by events	Description
`strategy`	Required	`RecommendationStrategy`	Recommendation create/update	Core recommendation logic selector.
`strategyConfig`	Optional	`object`	Recommendation create/update	Strategy-specific tuning options.
`productsToShow`	Optional	`number`	Recommendation display output	Number of products returned in placement.
`pageTargetingJson`	Optional	`object`	Placement eligibility	Page context eligibility constraints.

Supported strategies

manual
category_based
bestsellers
new_arrivals
on_sale
viewed_together
bought_together
user_history
similar_products
trending
ai_personalized
custom_query

Event or Endpoint Decision Matrix

Scenario	Use This	Why
Need curated hand-picked list	manual	Maximum editorial control.
Need category contextual relevance	category_based	Aligns product set with current category context.
Need social-proof style winners	bestsellers or trending	Highlights high-demand catalog items.
Need PDP affinity suggestions	similar_products or viewed_together	Contextual relevance on product detail pages.
Need cart cross-sell	bought_together	Purchase affinity driven cross-sell.
Need personalization	ai_personalized or user_history	User-specific relevance prioritization.

Common Errors and Fixes

Strategy output appears generic

Cause: StrategyConfig missing or too broad.

Fix: Add tighter constraints and preview results.

AI strategy underperforms

Cause: Low embedding/profile coverage or poor cold-start settings.

Fix: Check AI personalization coverage and fallback strategy.

Manual strategy stale products

Cause: No ongoing curation cadence.

Fix: Set update schedule for manual lists.

Contextual strategies miss expected items

Cause: Catalog metadata/category mappings inconsistent.

Fix: Normalize catalog category and product metadata quality.

Production Checklist

Strategy choice is documented per widget placement.Required
Preview validation is required before activation.Required
StrategyConfig is versioned for reproducibility.Required
Post-launch breakdown and funnel metrics are reviewed.Required
Fallback and rollback strategy is defined per high-traffic placement.Required