Datafication and personalization
The Netflix way to extend value creation for customers
Netflix started as a DVD rental company more than 20 years back. With personalization at its core, Netflix recommended and mailed DVDs to its customers. Back then its personalization algorithms had very few data points — the past rental history, length of the time a DVD was held, and maybe some additional demographics information.
Fast-forwarding to the time when Netflix launched a streaming service, it plunged into datafication of user behavior, capturing the customer browsing history, the points where a customer pressed forward, rewind or pause, the titles added to wish list, and so on. Netflix has divided its customers into a few thousand micro clusters that are, essentially, taste communities, and each individual might be part of multiple taste communities. This datafication in the modern streaming and internet world helped Netflix better understand its customers to provide a more personalized experience - a customized homepage for each customer and a hyper-personalized “Recommended for you” and “Because you watched the ABC title”.
La Casa de Papel aka Money Heist
La Casa de Papel, a crime thriller, was created and telecasted in Spain back in 2017, placing it for a strong start with a prime time slot. However, the viewership declined consistently over its first two-season run, with the show being summarized as a flop in Spain.
In Dec 2017, Netflix bought international streaming rights for La Casa de Papel, dubbed it into English, renamed it to Money Heist, and placed it into its catalog of numerous other foreign-language shows. Money Heist was launched on Netflix without any promotion, simply subject to the mercy of algorithms possibly with the right tags and classification. This is where folks like myself, with an affinity for crime thrillers, had Money Heist added to the Netflix recommendation list.
In parallel, in early 2018, the show's actors started observing an unusual phenomenon on their Instagram profiles — their followers were just climbing and the actors were not anonymous anymore. Within four months of its global launch, the show became the most-watched foreign-language series on Netflix, and the show was renewed for another two seasons in April 2018 with a significantly higher budget. Needless to say, the next two seasons were received with immense pleasure by fans, with 34M watching it in the first week of its launch.
How it happened — the science of recommendation systems
What happened here is pure data science and machine learning. Netflix published this show with tags such as “TV thriller, Suspense, Exciting.” Some other shows with similar tags include Prison Break, Narcos, Breaking Bad. In its microclusters, Netflix would have identified customers who like TV thrillers and Money Heist started showing up on their recommendation list. With more people watching the show and providing great reviews, the show jumped up the recommendations. With all the datafication, classification, and clustering that Netflix had done on its customers, the show was recommended to the right cluster of customers. I will be writing a detailed article on recommendation systems later.
The future — data integration across companies
If in a hypothetical situation, Netflix and Instagram were to come together and leverage each other’s datafication outcomes, both could create extra value for their respective customers. Instagram, with the knowledge of Netflix’s clusters, could propose new connections, show similar advertisements to the customers in one cluster, and so on.
Netflix can identify who are the actors its customers have liked/followed on Instagram and then use the pictures of these popular actors as thumbnails on different titles, a customized picture for each customer, to gain quick attention. If on my Instagram, I have followed Álvaro Morte (the Professor in Money Heist) then his picture would show up on Mirage as a thumbnail in my Netflix recommendations.
Now if Netflix were to create a new thriller, as it is creating its own content these days, it would identify, through Instagram, the popular thriller genre actors amongst its targeted viewer base. Bingo! Netflix has the cast for its next production.
Summary
Data is the new oil! More data, better algorithms, and better products leading to more data creation — the flywheel continues to support business model innovation and value creation for the customers. Companies have to venture on a datafication journey which is systematically extracting data from activities and transactions that are naturally ongoing in the business, establishing data pipelines that enable large volumes of data with high velocity while capturing enough variety. Companies like Google and Amazon have gotten to their current status, not only through world-class products but through in-depth knowledge of their customers.
