How to Use Product Data to Power Retail Relevance
How do you create highly relevant retail experiences? It’s all about matching customers to the products that excite them, and doing so requires an intimate knowledge of both your customer data and your product data.
Most retailers have (or are working on) the customer data part of that equation, but very few have yet to master the product data piece of it. So how do you do it? Recognizing that you need a deep understanding of your products is one thing, knowing how to capture that data is quite another.
The Netflix Approach to Product Data
Take a trip down memory lane to the heyday of Blockbuster. Once you walked in, you likely went straight to your favorite section (comedies, dramas, thrillers, documentaries…), but from there it was up to you to peruse the aisles of videos to find the (hopefully) perfect movie.
Now think about firing up Netflix. Finding a movie to watch is a completely different experience (for numerous reasons, but let’s just focus on the selection process). Instead of wandering through a comedy or thriller section, you have seemingly hand-picked recommendations based on what you’ve watched previously. We’re not talking “viewer favorites” or anything as generic as that, we’re talking “action movies with a strong female lead because you watched Tomb Raider and Kill Bill.” And Netflix can deliver such targeted and relevant recommendations because it tags each movie with hundreds of different attributes.
What happens when shoppers arrive on your website? Which does their experience resemble more: Wandering the aisles of Blockbuster or receiving hand-picked recommendations from Netflix? For most retailers, it’s the Blockbuster experience. But delivering the Netflix experience, while certainly complex, starts with tagging your products with attributes that drive customer conversion.
What You Need to Know About Your Products
Netflix reinvented the movie selection experience in large part because it thought of a new way to organize movies. It didn’t simply constrain itself to the typical categories of comedy, thriller, drama and so on, or even any other blanket categories for that matter.
So what would happen if you weren’t constrained by product categories like “shirts, pants and shoes” or “necklaces, bracelets and earrings” or “sofas, wall art and coffee tables”? What other ways would you describe your products? And how would you get more detailed in doing so?
One way to approach these questions is to get your team in a room, gather an assortment of different products and come up with 20 unique ways to describe each one. For instance, you might assign the following attributes to a shirt: Light blue, cotton, sleeveless, patterned, open back, v-neck…
If you were to do that for each of your products, imagine all of the different connections you could make. Suddenly, people who like shirts wouldn’t only like other shirts. Instead, one shopper might have an affinity for green cotton products while another might have an affinity for woven patterned products.
There’s no black and white answer to what you need to know about your products except that you need to get as detailed as possible in how you categorize them.
Making Relevant Recommendations Based on Product Attributes
The challenge once you achieve that deep product knowledge becomes how to find all the signals in the noise. With so many different customers and so many different attributes tagged to your products, how do you uncover one shopper’s affinity for green cotton and another’s for woven patterns?
When you had simply shirts and pants, matching customers to products was far easier. But it’s not humanly possible to determine the commonalities across each customer’s purchases and then come up with the item they are most likely to want next when your product attributes go deeper than those basic categories.
Enter data science. Ever the hot topic these days, data science sounds like something most marketers wouldn’t want to touch with a ten-foot pole. But it can actually be a marketer’s best friend.
Data science is all about filling in the blanks of what you don’t know. For instance, you know that people who bought shoes from you like shoes — that’s easy enough to see. But what you can’t see are all the other signals this group of shoe buyers display that make them similar. Data science can see that though. It can find what makes members of this group similar beyond the fact that they purchased shoes and it can use that information to identify other people in your database who exhibit similar behaviors. The result? An audience predicted to have an affinity for shoes even though they haven’t bought shoes from you — yet.
Product Data is the Missing Piece of Retail Relevance
The most successful retailers are those that understand the relationship between their customers and their products. Armed with that deep knowledge of customer and product data, you can serve increasingly relevant offers to smaller and more meaningful groups of customers. In turn, this potent combination should help deliver better customer experiences and generate more revenue by increasing conversion rates, order sizes and customer lifetime value.
But most retailers struggle to achieve this relevance — and the benefits that follow — because they lack a deep knowledge of product data. It’s time to change that, and it all starts with a product-tagging exercise.
Intrigued? Check out our eBook on the email personalization problem for more on what it takes to achieve retail relevance.
