Suggested Recommendations Experiences

The guidelines for creating an experience that leverages recommendations are the same as with any experience. They require you to define the WHY, WHO, WHAT, WHEN, and in the case of Automated Personalization experiences, HOW parts of the experience.

Potential Experiences

Presented here are some potential experiences with recommendations that you can use to get started. They are sorted into three categories: experiences with different algorithms in each variant, experiences with the same algorithms but different filters in each variant, and experiences that are ideally suited for home page or landing page recommendations.

Three Variants, Different Algorithms in Each

  • Viewed and Also Viewed
  • Purchased and Also Purchased
  • Top Selling by Purchase Count
  • Top Selling by Gross Revenue

Same Algorithms but Different Filters in Each Variant

  • Any algorithm with a dynamic value filter with product type as the attribute matching the product type of the product currently being viewed
  • Viewed and Also Viewed recommendation algorithm with a dynamic value filter with product type as the attribute excluding the product type of the product currently being viewed
  • Viewed and Also Viewed recommendation algorithm with a dynamic value filter with gender as the attribute
  • Viewed and Also Viewed recommendation algorithm with a dynamic value filter with color as the attribute
  • Static filters

Home Page or Landing Page Recommendations

  • Top Selling by Purchase Count
  • Top Selling by Gross Revenue
  • Purchased and Also Purchased recommendation algorithm based on the last item viewed or purchased
  • Viewed and Also Viewed recommendation algorithm based on the last item viewed or purchased
  • Viewed and Later Purchased recommendation algorithm based on the last item viewed or purchased

Additional Experience Tips

  • With Dynamic Testing experiences and Automated Personalization experiences, engagement-based metrics are best for up-funnel experiences.
  • You can append a parameter on click-through in the recommendations actions and then build an event that looks for that parameter.
  • For a down-funnel experience (for example, one that targets a cart page), you may be able to use revenue per session as the goal metric, especially with a Dynamic Testing experience.
  • Ensure that multiple recommendation experiences are all from the same catalog to avoid unexpected results.