After a search result
Recommendations can display similar or additional products without replacing the original query.
Recommendations →Comparison
Search and recommendations support different moments.
Search starts with an active question from the visitor. Recommendations start from product or behavioral context and help especially with discovery, alternatives and additional products.
PartSearchRecommendations
Starting pointActive search questionContext, product or behaviour
Strong atFocused intentionDiscovery and follow-up steps
Important signalsQuery, product data, rankingContext, product relations, behaviour
KPIsCTR, zero-result searches, conversionClicks and follow-up behavior
RoleBrings demand for resultIncrease discovery after or next to search
Coherence
Design both as parts of the same product journey.
Broad orientation
Combine search with filters and recommendations when visitors are still exploring choices.
Discovery Product →Measuring
Keep search and recommendation-KPIs separate so that effects are not mixed up.
Analytics →Relevance
Use product data and context as a shared foundation, but with separate quality criteria.
Product data →Assess search and recommendations on their own role.
This will prevent two different discovery problems from being assessed with the same solution.