While most digital marketers are intimately familiar with using recommendations to promote products and other content, it’s important to look at what makes for effective—or ineffective—recommendations strategies.
Successful recommendations that deliver real, measured returns understand how products are related to each other in ways beyond just the product category. This means making recommendations based on many contextual factors, including a person’s past buying history, their location, and what items they are viewing in their current browsing session.
So what is the anatomy of a smart recommendation? Marketers need a system that can make powerful recommendations based on:
Recommended for YouWebcast: Sales and Marketing Alignment: 7 Steps To Implement Effective Sales Enablement
- Contextual similarity. When solving contextual similarity problems, a recommendations engine should take less obvious relationships into account: Do the two products belong to the same brand? Are they in the same price range? Do they have similar ratings from similar users? These types of decisions are best made by a search engine that can quickly determine contextual similarity between semi-structured data objects.
- Affinity between items. Products can also be “linked” by users’ browsing and purchase behavior. Consider the classic association-rule example of beer and diapers: during certain times of the day, male shoppers tend to buy them together, even though on the surface, the items are completely unrelated. Valuable insights can be gained from uncovering these unexpected associations. Understanding the affinity of one item to another based on observed user behavior can be a significant analytics challenge—especially when you’re dealing with massive scale—and requires the use of multiple algorithms.
- Rich data. Your recommendations will be much more powerful if the corresponding user data is rich and detailed. The more demographic and behavioral information that can be associated with a user profile, the better, since the recommendations engine will be better equipped to draw the right conclusion about what product suggestions a user is most likely to respond to.
Some users are simply waiting for the right product—by using optimized recommendations, you can uncover what they’re most interested in. It’s not a “one size fits all” approach, and one algorithm is not likely to work for every use case. For instance, the women’s apparel market is different than the user intent behavior for car buyers, which is in turn different from people looking for financial services products. Each unique shopping session has a different user mindset associated with it, which is dependent on a complex set of characteristics ranging from the vertical of the organization, nature of the offers being presented, sense of urgency that is associated with the campaign, and other factors.
The goal of personalization is to show visitors the most relevant content that has the most potential to increase your revenue per visitor or conversion ratio.
When evaluating recommendations solutions, look for one that will give you the flexibility to test algorithms against each other and let them “fight it out,” instead of just specifying rules. You’ll be glad you did.