Recommendation Strategies

Overview

Algonomy embraces the individuality of every customer by offering 100+ recommendations (For example, "people that viewed this item ultimately purchased...") and selects the most relevant recommendation type(s) for each page and customer situation. Rather than using vague language such as "May we suggest" or "You may also like," each recommendation reflects an explicit message so that customers understand precisely why something is being recommended.

The Relevance Cloud selectively displays recommendation strategies based on the "wisdom of the crowds" or the customer's own behavioral profile, leveraging what is known about his or her purchase and browsing history and where he or she is in the shopping process. As a result, the customer trusts the recommendation because it relates directly to his or her individual needs and behaviors. Further, the use of multiple recommendation types allows for multiple placements per page, offering customers a choice between different clearly defined browsing pathways. The result? Increased conversions and sales, enhanced customer confidence and engagement and reduced customer service inquiries and returns.

Why is Clear Messaging so Important?

A/B tests have demonstrated 50% greater conversion when a customer knows why he or she is being recommended an item. Unlike "black box" platforms, the Relevance Cloud is able to preserve the "why" and share it with the customer via explicit messaging such as "People that viewed this ultimately bought." The result is seamless navigation through the shopping experience via trusted recommendations.

Customized Messaging

Because customers vary across merchant sites, messaging that resonates with one retail audience may not be appropriate for another. Omnichannel Personalization enables retailers to speak to their specific audience with customizable recommendation messages.

Sample Placements with Customized Messaging:

Original

Custom

People who viewed this product also viewed:

Folks who viewed this product also viewed:

People who bought this product also bought:

This gear rides well with:

People who viewed this product also bought:

People who looked at this movie purchased:

People who viewed this product ultimately bought:

People who viewed VitaMix Turbo Blend 4500 Blender Ultimately Bought:

Strategy Families

To help strategies easier to navigate, they are divided into families based on how they are best used.

  • Personalized Strategies: Personalized Strategies leverage events stored in a user profile to recommend relevant products for that given customer.
  • Cross-Sell Strategies: Cross-Sell Strategies are used to return recommendations for products that compliment an item the customer has looked at or added to their cart. These strategies are often used deeper in the funnel with the intent to add value to an order.
  • Offer-based Strategies: Offer-based Strategies rely on attribute data to determine that an offer is available and presents those items in the recommendations.
  • Channel-based Strategies: Channel-based Strategies can offer a better view of top-selling products when a site has items that perform differently in particular channels.
  • Replenishment Strategies: Replenishment Strategies examine a customer’s previous behavior and return recommendations to replenish items previously purchased.
  • Offline/Omnichannel Strategies: Offline/Omnichannel Strategies use data from offline purchases (sent to Omnichannel Personalization in a feed). Offline strategies will use the purchase data sent via feed on its own, while omnichannel strategies are a blend of the online data that is collected by Algonomy and offline data.
  • Region-aware Strategies: Region-aware Strategies are designed for retailers with regions enabled and will return recommendations based on the behaviors of regional customers.
  • Segment-based Strategies: Segment-based Strategies operate similar to other strategies but look only at the specific segment the customer belongs to.
  • Sponsored Strategies:  Sponsored Strategies work with products sponsored via HookLogic in order to allow advertisers to buy slots in recommendations in order to raise exposure to their products.
  • Top Strategies: Top Strategies return recommendations based on the top products for the site. Top products can be defined by views, purchases and sales attributed to those products, or a combination of these metrics.
  • Similar Item Strategies: Similar Item Strategies return recommendations that are similar to the object the customer is viewing or has added to the cart. These strategies are meant to give customers more options of what they are looking at to pick the best product that fits their specific needs.
  • Search Optimization Strategies: Search Optimization Strategies use internal site search terms as the seed and build models around those search terms.

Strategies with Custom Layouts

Many recommendation strategies return results that would benefit from customized layouts. Recommendations that return categories instead of products or return product hints or percents along with the list of recommended products, benefit greatly from layouts that present these elements logically.

Category Layouts

These strategies recommend categories instead of products:

  • CategoryCP

  • CategoryCP2

  • SolrSearchToCategory

  • CategoryEV

Product Hint Layouts

These strategies return product hints, which can be displayed next to the results, preferably with formatting that distinguishes the hint from the recommendations ("People who viewed X also liked Y"):

  • BuyTogether

  • CategoryDiversePersonalizedPurchaseCP

  • LimitOnePersonalizedClickCP

  • MultiItemPersonalizedOfflinePurchaseCP

  • MultiItemPersonalizedOmniChannelPurchaseCP

  • MultiItemPersonalizedPurchaseCP

  • MultiItemPersonalizedViewCP

  • PersonalizedCategoryTopOffers

  • PersonalizedCategoryTopSellers

  • PersonalizedClickCP

  • PersonalizedClickCPInCategory

  • PersonalizedClickEV

  • PersonalizedOfflinePurchaseCP

  • PersonalizedOmniChannelPurchaseCP

  • PersonalizedPurchaseCP

  • PersonalizedPurchaseCPInCategory

  • PersonalizedPurchaseEV

  • PersonalizedTopOffersInBrandAndCategory

  • PersonalizedViewedPurchased

  • PersonalizedViewedPurchasedInCategory

Percent Layouts

These strategies also return percentage product hints, which can be displayed alongside the recommendations:

  • SearchedPurchasedPercent

  • SolrSearchToPurchasePercent

  • ViewedPurchasedPercent

For more guidance on best practices and getting the most out of your recommendations, speak with your Client Services representative.

Special Strategies

Discover and Advanced Merchandising use their own strategies behind the scenes.

Search and Browse

Discover uses custom strategies to personalize the sort order of a list of products. The strategy you use depends on whether you're sorting a list of products in a category or search results: 

  • CategorySort

  • PersonalizedCategorySortV2

  • PersonalizedSearchSort

  • SearchSort

Refer to Search and Browse Setup for more on using these strategies.