Item Page Best Practices

Overview

The objectives of recommendations on the Search Page are to drive consideration of the incumbent product and promote alternative options and/or cross-sells.

Quantity: 2-3 placements
Location: We strive for one placement above-the-fold with the remaining interspersed within the page content. Remember, the main objective of the item page is to drive consideration of the current product—something which recommendations cannot help, only hurt. As such, be judicious in the positioning of recommendations so as not to disrupt customer absorption of salient content, which often becomes a vertical-specific consideration. For example, within Customer Electronics and Appliances (functionality-centric verticals), product specifications and descriptions are an important part of customer consideration. Algonomy avoids recommendations between the core product content (image, price, etc.) and the descriptions, instead opt for an above-the-fold vertical placement in the right margin. This is not the case in many softlines verticals such as Apparel.
Strategies: Recommendations are commonly a mix of Similar Items, Cross-sell, and Personalized strategies to support the various merchandising objectives of the page. This will usually result in similar products being presented in the top-most placement with the other recommendation types falling to the remaining zones.

Diagram

Description automatically generated

At times, it makes sense to fulfill the cross-sell objective with more curated recommendations (for example, “Complete the Look” or “Frequently Bought Together”) through Advanced Merchandising. Provided the retailer can supply the requisite attributes, this is an effective tactic across a number of verticals/categories—hardlines and softlines—where pairings are important: Apparel, Customer Electronics, Appliances, Home Furnishings, etc. Advanced Merchandising recommendations do not require a net new placement; rather they should replace the existing placement designated for cross-sells.

Strategy	Family	Description	Comments
ClickCP	Product Viewed Viewed	People who viewed this item also viewed Products other customers have viewed given they viewed the current product
ClickEV	Product Viewed Viewed	People who viewed this item also viewed Same as ClickCP but bias toward higherpriced items
CategorySiloed ViewCP	Product Viewed Viewed	People who viewed this item also viewed Same as ClickCP but results are filtered to match the category of the seed item	Only works if a category hint is instrumented or the product has only one category
ViewedPurchased	Product Viewed Bought	People who viewed this item ultimately bought Products other customers have bought given they viewed the current product
CategorySiloed ViewedPurchasedCP	Product Viewed Bought	People who viewed this item ultimately bought Same as ViewedPurchased but results are filtered to match the category of the seed item	Only works if a category hint is instrumented or the product has only one category
ViewedPurchasedPercent	Product Viewed Bought	People who viewed this item ultimately bought Same as ViewedPurchased but returns seed product and % of customers that have purchased each item	Requires layout that exposes seed and conversion %. Only use if seed clearly delineated as customers will confuse it for a new item.
PurchaseCP	Product Bought Bought	People who bought this item also bought Products other customers have bought given they bought the current product	Also enable the V2 version of this strategy which de-prioritizes general top-sellers
PurchaseEV	Product Bought Bought	People who bought this item also bought Same as PurchaseCP but bias toward higher-priced items	Also enable the V2 version of this strategy which de-prioritizes general top-sellers
SessionPurchase CP	Product Bought Bought	People who bought this item also bought Same as PurchaseCP but purchases must be from same session
CategoryDiverse PurchaseCP	Product Bought Bought	People who bought this item also bought Same as PurchaseCP, but does not recommend products belonging to primary category as seed and only recommends one product from each unique category	Includes name of category to which recommended product belongs; can be exposed in layout
CategorySiloed PurchaseCP	Product Bought Bought	People who bought this item also bought Same as PurchaseCP but results are filtered to match the category of the seed item	Only use in verticals where same-category cross-sell make sense, e.g. Apparel but not Consumer Electronics
PersonalizedClick CPInCategory	Personal- ized	People who viewed items you browsed also viewed Products other customers have viewed given that they viewed a product the customer recently browsed. Uses category context, if available, to show only products in same category as seed.
Personalized ViewedPurchased InCategory	Personalized	People who viewed items you browsed also viewed Products other customers have viewed given that they viewed a product that the customer recently browsed. Uses category context, if available, to show only products in same category as seed.
MultiItem Personalized ViewCP	Personalized	People who viewed items you browsed also viewed Products other customers have viewed given that they viewed products that the customer recently browsed	This can recommend items outside of the current category which may not be desired by client
RecentHistorical Items	Personalized	Recently Viewed Items Items viewed during current and previous sessions	This can return items outside of the current category which may not be desired by client
CategoryTop Sellers	Category Best Sellers	$STRATEGY_HINT$ Best Sellers Top products in the category based on unit sales
CategoryTop Products	Category Best Sellers	Top Products in $STRATEGY_HINT$ Top products in the category based on views, clicks and purchases
PopularProductsInCategory	Category Best Sellers	Popular Products in $STRATEGY_HINT$ Top products in the category based on # of line items * total revenue