Adam Sandler in 2017
PHOTO: Denis Makarenko / Shutterstock

Have you ever wondered who watches Adam Sandler movies other than adults stuck in perpetual arrested development? Ask Netflix. It knows exactly who watches his films and why they watch them. Netflix doesn’t know the names of everyone who likes Adam Sandler movies, since that’s the kind of personally identifiable information that most of us would want to keep shielded, but it knows whether you prefer Adam Sandler in silly comedies, romantic comedies, buddy comedies or, unthinkable as it is, if you like him in all three.

In other words, so long as audiences keep watching "The Waterboy" or "Happy Gilmore," we’ll have another movie starring Adam Sandler as the world’s oldest teenager to look forward to. The machines have spoken.

Using Contextual Data to Inform Product Development

Of course, I’m kidding about Adam Sandler to make a point: it’s not just data that drives decisioning, but the contextual relevance of that data. Just because you watched "Funny People," for example, doesn’t mean you’ll enjoy "Little Nicky." Maybe you like films directed by Judd Apatow. Maybe you like films about stand-up comedians. Maybe you stopped watching "Funny People" the moment you realized Adam Sandler was in it. Understanding the context behind the click is critical for Netflix to deliver the next best experience to you, rather than lumping you into a customer segment that says “Adam Sandler fan,” thus dooming you to an eternity of bad movie recommendations.

While contextual data is critical for content and offer recommendations, it’s also very important for product development. In part one of this series, I introduced the concept of audience centrism. Creating products or services for an existing audience is far more cost-effective than trying to convince customers they should buy your product. Here again, Adam Sandler is the perfect example. Netflix surprised many when it signed Sandler to multi-million dollar, multi-movie deals in 2014 and again in 2017. But Netflix was simply creating a product for an existing audience: Adam Sandler movie watchers. Its customer data clearly revealed that a substantial chunk of its audience would watch anything with Adam Sandler starring in it. 

Netflix wasn’t taking a gamble on Sandler and hoping people would tune in — it was providing a product for a paying audience that was, figuratively speaking, already in the seats and waiting for the movie to start.

Related Article: The Art of Audience-Driven Marketing: Stop Selling Products, Start Building Audiences

Getting to the 'Why' Behind the Data

You might take from this that Netflix and other audience-centric brands are simply letting data drive their business model, but it’s more sophisticated than that. They’re using new technologies like machine learning and customer data platforms to understand the data behind their data. It’s not enough for Spotify to know you clicked on the new Justin Bieber song; it needs to know why you clicked on it. Was it curiosity, peer pressure or are you a true Belieber?

Netflix, for example, organizes its data using a highly complex system of content affinities and features. A film like "The Waterboy" might have hundreds of different features: sports movies, football movies, underdog movies, Adam Sandler movies, SNL alumni movies, comedies from the 1990s, etc. Understanding which data features are driving customer actions is what allows Netflix to recommend the next best experience. Think of it as adding intent to an action. The difference between watching a movie and enjoying it is significant, as different as "Little Nicky" is from "The Longest Yard." Understanding that intent is critical for Netflix and, really, for any business where content recommendations are important.

Related Article: Decisioning — The Only Way to Accelerate Analytics to Value

The Caveat: Decision Velocity

One caveat to this model is that content recommendations require real-time decisioning. This is where traditional data marketing often fails, because it involves a lot of pre-analytic effort (e.g., data consolidation, cleansing) and, in many cases, direct involvement from the IT department. That means doubt and delay creeps into your data-decisioning process, and once that happens you lose the competitive advantage that real-time decisioning can give you.

Decision velocity is often under-discussed and sometimes even misunderstood by marketers. I’ll take a closer look at decision velocity in the final blog of my three-part series, “Can Your Data Decisioning Keep Up with Dad Sneakers?”