Think about the last time you got a gift or a note from someone that you would describe as “meaningful” and “personal.” What made it feel that way?
Maybe they remembered something you said you liked or needed weeks ago and got it for you. Or maybe they referenced distinct memories that they know make you smile or laugh. The key is that they knew what you (and only you) wanted and what would excite you.
Now imagine you and your friend share a birthday and someone got you each the same shirt just in different colors. Sure, each gift was different, but it wasn’t truly personal.
The second scenario is exactly how most marketers approach personalization today: By taking the same experience and making it slightly different based any number of factors, all of which are about the customer. But while that creates a unique experience, it doesn’t create a personal experience.
So how do you achieve true personalization? It’s all about matching customers to the products that excite them.
Retailers Everywhere Have Been Led Astray by Technology
To date, most of us have thought about personalization in retail through the lens of the available technologies. For some, that means onsite personalization tools that present a different experience for new versus returning visitors or based on each visitor’s geographic location. For others, it may mean personalized product recommendation engines that generate suggested items by looking at common purchase patterns.
But the truth is we’ve been led astray by technology because those efforts only take into account customer data. As a result, they can only create unique experiences, not truly personal experiences that match customers to the product recommendations that excite them.
What’s missing? Product data.
The reason why achieving true personalization has proven so difficult in the past is because marketers have been limited to working with behavior-based data. Marketers haven’t had deep enough product data, let alone a way to match together product and customer data.
Consider the case of cross-sell personalization as most marketers know it today. You might recommend three products to a customer who bought a certain shirt because those are the most popular purchase combinations. But those recommendations are based on behaviors, not product attributes. Instead, what if you saw that a customer purchased two shirts, both of which were made of cotton or both of which were blue? That would change your recommendations, wouldn’t it?
Behold the Importance of Product Data
When you start thinking more deeply about why your customers buy certain products, you take on the perspective of leaders of companies like Netflix, Facebook, Spotify and Stitch Fix. Those four brands get held up time and again as prime examples of how organizations can execute a seamless personalization strategy. And it’s not a coincidence that they all share a common trait — they know an enormous amount about their products (even in cases where those products are media content or people), and they’ve built a platform that allows marketers to access those product attributes. In turn, that knowledge allows them to match customers to the products that excite them better than anyone else.
Consider Netflix: If I watch Seinfeld and Everybody Loves Raymond, Netflix won’t just recommend to me any random sitcom. Instead, Netflix will say those are New York-based sitcoms starring stand up comedians, so it might recommend Louie and Everybody Hates Chris. But if I watch Seinfeld and Friends, it’s a different story. The similarities there are ensemble comedies, with a group of 20-30 something singles living in New York, so I might get recommendations for How I Met Your Mother and Will & Grace. Netflix has tagged each show with hundreds of different attributes, and that deep level of attribution allows it to make great (re: personal) recommendations that go beyond very basic signals.
It’s the same with Facebook — the social media giant does a remarkable job of selling you (their product) to advertisers because it’s collected thousands of different attributes on all of its users. And Stitch Fix takes this model one step further, because it not only collects a variety of attributes about its products, but it also has a unique customer-stylist relationship that makes it easy to collect feedback. In turn, this feedback means that Stitch Fix knows a certain customer doesn’t like stripes or that another prefers looser fitting clothing.
Putting Your Product Data to Work
Of course once you have all of that deep product data, you need to layer on intelligence in the form of smart AI in order to figure out what’s important and make actionable recommendations.
Ultimately, this type of setup has also helped leaders like Netflix and Stitch Fix avoid some of the biggest knocks others get about personalization — the fallout from bad recommendations. Customers have built up so much trust in the personalized customer experience those brands created, that they almost gloss over the bad recommendations.
So how do you get there? It takes some work, but it is possible once you have the right infrastructure — including the ability to capture product data and match that to customer data — and a way to get that insight out the door (we believe the easiest place to start is the email channel, but the “where” is really secondary).
Intrigued? Check out our eBook on the email personalization problem for more on what it takes to create truly personal retail experiences.