As much as companies talk about getting to know customers on a personal level, we are all inundated with emails touting products that are unequivocally wrong for us. 

At times I’ve found myself wondering, “How many people have to unsubscribe before marketers decide to invest in truly understanding their customers’ preferences? 5,000? 100,000?” Everyday shoppers probably ask themselves similar questions, believing that, even after interacting with a brand only once, that brand should get them.

Retailers don’t disagree. In fact, most want to perfect their one-to-one personalization. They just go about doing so the wrong way: They get to know consumers by observing them and their behaviors.

Getting to know shoppers by ... getting to know shoppers. Seems logical enough. The problem with this approach is that, while it might result in multiple data points about each individual shopper, those data points have little context if they’re not paired with insights into the products the shopper has engaged with or purchased.

Related Article: Bridging the Gap Between Data Analytics and Personalization

Lessons From the King of Next-Item Recommendations

Take Netflix. Netflix doesn’t make its “Recommendations for Jared” based solely on my age and the fact that I’m a male who lives in New York City who tends to fire up “Arrested Development” (c’mon!) around 10 p.m. on weeknights. Netflix pays close attention to the shows I watch and what those shows have in common on a granular level.

After I spend time watching even just one episode of “Arrested Development,” Netflix begins coming up with theories, such as “Jared might like other goofball ensemble comedies,” or “perhaps he likes Jason Bateman and would like to see more of him,” or “maybe Jared would be interested in documentaries about wealthy families who lost everything.”

These theories manifest as recommendations for me, and I either react positively to them by watching or I don’t respond at all. Either way, I give Netflix more information. As I watch or engage with more shows and reject others, Netflix gains more show (or “product”) data to work with and can begin identifying recurring themes based on attributes shared across shows.

This approach has allowed Netflix to refine its recommendations so acutely that many viewers believe that Netflix knows what they like better than they do. Ultimately, Netflix strives to be as relevant as is humanly and technologically possible, and it does so in ways that can be adopted by retailers.

Nuanced Product Attributes > Broad Product Categories

Netflix’s understanding of its shows and its viewers is very nuanced, and that comes through in its high level of relevance to viewers. This nuance stands in stark contrast to the approaches of many retailers, who think in terms of broad categories, like pants, shirts and shoes or sofas, chairs and tables, and so on.

With access to the above-mentioned product data, however, there’s no reason to let such blunt and broad categories inform recommendations and interactions. Instead of thinking along the lines of “Two customers bought blue sweaters,” consider a far more nuanced understanding, such as “One customer bought a royal blue, V-neck, cashmere sweater and the other customer bought a light blue, cotton, crew-neck sweater with a flowered pattern.” All of a sudden, the next product you would recommend to each individual becomes far different.

Related Article: How Marketers Can Solve the Cross-Channel Personalization Puzzle

Learning Opportunities

Putting Product Attributes to Work

Basing product recommendations on nuanced attributes of the products that shoppers interact with helps head off a common retail marketing trap: promoting irrelevant next-purchase suggestions. 

Most often, retailers fall into that trap when they use broad product categories alone because that broad categorization forces them to make recommendations based on gut instinct, the quantities they have in stock, likely margin or potentially irrelevant external trend data. 

None of those factors take into account whether the items are relevant to the customer. Sure, the sales generated by showcasing those items will boost a brand’s bottom line, but they don’t build customer trust through relevance and may even turn some shoppers away.

Related Article: The 3 C's of Personalized Customer Service

Product Data Enhances Behavioral and Customer Data

Onsite behaviors and customer data help round out the view of how people shop. In other words, the combination of product, behavioral and customer data provides insight into what individual consumers are likely to buy and what approach retailers need to take to get those shoppers to buy again.

For instance, pairing nuanced product attributes with customer-specific behavioral data allows retailers to develop a detailed picture of shoppers that not only provides information about which products to recommend, but also offers insights into, for example, the best time to reach shoppers with those recommendations — thus creating the ultimate relevant experience.

Retailers may certainly face roadblocks that may limit their ability to adopt this Netflix-style approach, but there are few excuses for them not to at least try.

The ruling philosophy should be this: Assume nothing, consider everything, and go beyond basic descriptors. With deep enough product attributes, retailers can make recommendations that reaffirm their understanding of shoppers and create highly relevant experiences.

fa-solid fa-hand-paper Learn how you can join our contributor community.