Marketing Automation Lessons from Netflix

netflix-logo-redEarlier this month, an Atlantic article by Alexis C. Madrigal about Netflix‘s genre generator made the rounds of my various networks. Some of my friends were really just having fun with the “Gonzo” genre generator (“Love Triangle Satanic Cynical Action Movies Set in Africa For Ages 8 to 12,” anyone?); others appreciated the deep-dive into the classification system. A third set of friends were just plain excited about the article.

To this third group, Madrigal’s extensive analysis of a classification hierarchy that Netflix assigns almost instantly to its catalog was an accessible (and less scary) example of how automation is not only all around us, but also, and more importantly, beneficial to companies and consumers alike.

Netflix’s genre-assigning algorithm is not a marketing automation tool, per se, but Madrigal’s article makes it clear that there’s a lot that those of us who work in the marketing automation space can learn from Netflix. Here are a few items that particularly jumped out at me:

Use Your Data

“Netflix has created a database of American cinematic predilections,” Madrigal writes. “The data can’t tell them how to make a TV show, but it can tell them what they should be making.” Netflix has long used its data to make recommendations, but in the last few years, it has discovered how to use the same information to produce content, but guiding them toward what people want to see.

Good marketing automation, when executed correctly, should parallel this, becoming predictive rather than reactive in consumer interactions. Based on everything we know about consumers-both first- and third-party data-it is easy to automatically generate individual profiles that then inform what marketing collateral they receive. This allows companies to know that their marketing will be effective before they distribute anything-and to be confident that they’ll be releasing far less “bad” marketing.

Much as the single, yoga-practicing twenty-something who works a full-time job and the sixty-five-year-old retired golf enthusiast probably have very different Netflix profiles, so too do they fit into very divergent marketing demographics. But without high-level automated data processing, reaching them successfully based on their profiles is impossible.

Be Responsive

Although Netflix does not rely on the five-star rating system as much as it had in the past, it still uses the system, along with viewing history and other data, to determine the likelihood that a customer will watch (and enjoy) a film. If a viewer indicates a strong preference for or against a certain film or film genre, the Netflix interface will respond in real time (or nearly) to alter its recommendations.

As marketers, we cannot afford to miss out on consumers’ “likes” or ignore their “dislikes;” nor can we afford to monitor all of these behaviors and preferences manually. So if our golf enthusiast buys a Taylor Made wedge and putter, the system should automatically recommend a Taylor Made driver (assuming there’s no record of the golfer having bought one previously), and not a Callaway wedge or putter.

It isn’t possible to have the sales team report each transaction to the marketing team so that new materials may be generated; it is, however, possible for the combination of stated preferences and behavioral history to trigger an automatic reaction that customizes these same materials.

Be Open-Ended, Always

Sure, Netflix employs over 90,000 genres to classify and sort films in its database, but what’s important is they’re 90,000 highly open-ended genres. Rather than saying that a film can fit into one broad category-and only one-the highly-nuanced Netflix genres allow for overlap, so that a film might fit into more than one category or appeal to more than one sort of viewer.

Automation should never pigeonhole a company, or its consumers, into just one category or interest set. While a manual analysis might say of our yogi: “she bought a yoga mat, we should target her with Giam products from now on,” automation allows for the creation of a multi-faceted, and much more real, consumer profile. So now, she can be targeted not only for her yoga practice, but also for her collection of vinyl records and her love of cooking-all things the system knows about her based on her interaction with the brand.

People, after all, are open-ended, multi-dimensional. They’re also complicated. When taken en masse-whether Netflix‘s viewership or the customers of a certain e-retailer-it’s little wonder that it takes automation to meet their needs.

Related reading