Making Product Recommendations that Customers Really Want

BrandPost By Aaron Goldberg
Mar 03, 2020
IT Leadership

istock 1150199866
Credit: iStock

One of the most important yet often overlooked changes in the new customer experience involves the huge increase in the number of products or service options available to customers. Catalogs seem endless, and simple searches return thousands of items. The reality is that the legacy approach of providing basic level of search functionality and bombarding the consumer with choices does not work. To stand out, brands regardless of size must embrace technologies like AI, which allows them to be dynamic and help exceed customer expectations. Product recommendations must be truly useful, and merchants that recognize that fact are bringing this functionality to the fore.

Next-generation product recommendation technology does far more than traditional merchandizing. And it goes beyond merely providing other options “similar to what you looked at” or “what others are buying.” Modern, automated product recommendations must consider all the data the brand has about its customers to provide more focused and worthwhile recommendations. These recommendations will then need to be updated in real time based on how the customer reacts to the initial options.

“Research has shown that with effective product recommendations, customers can convert at up to 30% higher rate than with legacy approaches,” says John Stockton, senior director of product at Adobe Commerce. “Magento Commerce Product Recommendations, powered by Adobe Sensei, provides a more intelligent, agile and scalable path forward for merchants big and small to deliver highly-personalized product recommendations, giving them a leg up from their competitors.”

Astute recommendations leverage personalization, utilizing the same cohesive real-time input from customer interaction, Stockton adds.

Avoiding the pitfalls

Getting recommendations right is essential. Bad or misdirected product recommendations will motivate prospects to shop elsewhere. But identifying this problem can be difficult. There are many ways the recommendation process can go off the rails. Problems arise when the recommendation system isn’t linked to other key data. For example, recommending a product that is out of stock will obviously lead to a bad customer experience. Repeating recommendations after they’ve been ignored increases customer frustration. And recommendations must be an omnichannel effort. When the customer contacts the brand by more than one channel and gets inconsistent recommendations—or none at all—sales will suffer.

The product recommendation process must also be done in real time to be relevant and engaging. The starting point is to utilize everything you know about the customer to make initial recommendations. However, that’s just the beginning. Watching how customers interact with those recommendations, what filters they use, and which recommendations they explore further must happen continuously. This requires AI-based systems. The large amount of data being created must be analyzed for key triggers that are recognized immediately. AI and machine learning (ML) are essential, since there’s not enough time for humans to do the processing.

“The final piece of the puzzle for making impactful product recommendations is to ensure the recommendation system is integrated with all of the customer data the organization has developed and the product catalogs and offerings available to customers,” Stockton observes. “Comprehensive solutions such as the Adobe Experience Platform and Magento Commerce provide the underlying personalization functionality that is essential to making great recommendations.”

For more information on Product Recommendations for Magento Commerce, check out this post.