It seems like everywhere you look, you can’t escape ‘data.’ Once obscure arguments about the data accumulation and handling of “free” technology services like Facebook and Google have exploded into the mainstream. Data Science has become an increasingly popular field, as has the use of data-driven “analytics” in a wide ranging number of fields, from healthcare to sports, to city planning and wealth management. The primacy of data (especially “big data,” or the accumulation and synthesis of vast amounts of data enabled by increasing computing power) seems to be a near-consensus: more data will improve our lives, and make us happier, healthier, and more informed.
However, the one area where data’s supremacy remains in doubt is in the creative industries, including the arts, film and television production, and book publishing. Whereas many industries have leaned into data head-first as a key (if not sole) determinant behind decision-making, the creative industries have taken a more cautious approach, wary of confronting the artists and their creative (and oftentimes highly analog) processes and (in my view) inherently skeptical of the power of “data” to yield a breakout success. Whereas data may be helpful in determining the correct number of stoplights per square mile or the perfect amount of risk in one’s portfolio, leaving a computer to assign cultural importance, popularity, and commercial success to an unknown and volatile combination of artist, author, actor, director, designer, etc. remains a task mostly left to human beings.
Yuval Noah Harari’s 21 Lessons for the 21st Century (my review here) takes on this issue head-first, arguing that rather than an inherent inability for computers to create or select art that resonates with people on an emotional basis, it is merely a question of time and data. One of the central theses of his book deals with the merging of biotech (data about our bodies, emotions, neurology), and infotech (data about the rest of the world), and computers’ ability to synthesize that data into actionable and/or tangible outcomes.
As Harari explains:
..in the long run no job will remain absolutely safe from automation. Even artists should be put on notice. In the modern world art is usually associated with human emotions. We tend to think that artists are channeling internal psychological forces, and that the whole purpose of art is to connect us with our emotions or to inspire in us some new feeling. Consequently, when we come to evaluate art, we tend to judge it by its emotional impact on the audience. Yet if art is defined by human emotions, what might happen once external algorithms are able to understand and manipulate human emotions better than Shakespeare, Frida Kahlo, or Beyonce?
After all, emotions are not some mystical phenomenon –they are the result of a biochemical process. Therefore, in the not too distant future a machine-learning algorithm could analyze the biometric data streaming from sensors on and inside your body, determine your personality type and your changing moods, and calculate the emotional impact that a particular song–even a particular music key– is likely to have on you.”
Netflix was long seen as the leading vanguard of this data-first movement in the entertainment industry. Netflix has become famous for its fealty to algorithms in recommending content for its users to watch, and even created a million dollar prize (known as the “Netflix Prize”) to create a tool to accurately predict user’s ratings of films and television shows. Over time, this approach extended to Netflix’s content acquisition strategy – upfront investment for exclusive rights to television shows and movies. Famously, the algorithmic combination of the popularity of the UK political series “House of Cards,” director David Fincher, and Kevin Spacey led to the adaptation and subsequent break-out success of House of Cards, Netflix’s first high-profile content acquisition.
With its center of operations in the tech-based San Francisco as opposed to the power center of Hollywood, Netflix has been seen as a rebel to the traditional Hollywood studios, with its ingrained ecosystem of agents, actors, and studio heads. Further, Netflix’s reputation as a ruthlessly meritocratic place to work, with a high rate of turnover and emphasis on producing results, gave further credibility to its status as a innovative company with start-up roots. Investors eager to load their portfolios with disruptive and highly profitable technology companies have driven up Netflix’s stock price, leading to its inclusion into the FAANG acronym of ‘must-own’ stocks (Facebook, Apple, Amazon, Netflix, and Google). Comparing Netflix’s 5-year stock performance against traditional entertainment companies (Viacom, Comcast, Disney) is a telling indicator – in the minds of investors, Netflix is the future of entertainment, while its immediate industry competitors are poorly equipped to handle this ongoing ‘disruption.’
However, in Netflix’s continued pursuit of subscriber growth and improved financial results, as well as growing competition for licensed content from the film and television studios that initially produced the data, as well as Amazon, Apple, HBO, Hulu, and other company’s entrance into the film and tv-streaming platform business, Netflix has increasingly moved towards producing its own content. In order to meet the growth and profit demands of its investors, Netflix could no longer rely on expensive acquisitions of existing content that included pricey royalty payments to the producer and short rights windows subject to renewal negotiations. Suddenly, Netflix began acquiring and producing hundreds of different films and tv shows, including awards-focused fare and commercially-driven projects alike. In order to sate its consumer, Netflix has chosen to focus on quantity more than quality, to grow its catalog of shows and ensure that consumers paying a monthly subscription fee would never run out of things to watch (this year, Netflix will produce ~700 “original” shows and movies). In the process, Netflix has racked up more than $6.5B in long-term debt, using bonds to finance this spree of content investment.
As a result, Netflix has transformed into a traditional Hollywood studio, mostly unrecognizable from its immediate competitors. Every week, Netflix seemed to poach executives from media and entertainment companies to join its ranks (just search “Netflix hires” on Google to bring up a spate of Hollywood press releases), bringing the entrenched and systemic mindset of Hollywood with them. In the process, Netflix has bifurcated its path forward between its religious devotion to data and innovative approach with its need to build credibility and work within the traditional Hollywood system (or generously, a balance of the two) in the continued pursuit of subscriber growth and financial returns.
This contrast and conflict between Netflix’s technology-based, data-driven origins and its increasing turn to becoming a traditional media company was reported on by the Wall Street Journal in interesting detail, in an November 10th article entitled “At Netflix, Who Wins When It’s Hollywood vs. the Algorithm?” (again, doubling as this week’s Best Thing I Read This Week.)
The article provides an interesting look at how decisions across the company are made with competing considerations pitting “the data” against the more human- and art- and business-driven parts of the company. The article leads with the decision whether or not to exclude one of the two main actresses from the sitcom Grace and Frankie in major promotional and marketing materials (“the data” said that images excluding Jane Fonda, and solely featuring Lily Tomlin resulted in more clicks), but expands into broader and more existential concerns, including whether or not to “green-light” projects or renew television shows, how to incorporate actor/director input, maximizing the effectiveness of film/show titles and trailers, and even whether or not to spend on billboards and other traditional advertising outlets. In each of these cases, Netflix has run up against the complex calculations whether or not to placate agents and their representatives (the “creatives”), or to trust the algorithms and their cold, “rational” calculations.
The Wall Street Journal piece cites several examples where the Hollywood “arm” has won out over the Silicon Valley arm, further underlining Netflix’s continued shift in focus from Northern to Southern California. As the article explains:
Some shows at risk of being canceled due to poor performance have gotten a reprieve because netflix doesn’t want to damage relationships with key producers or actors, people familiar with Netflix’s deliberations say.
At times, the efforts to appease stars don’t sit well with the company’s technology and product teams, triggering heated discussions between the Hollywood and Silicon Valley arms of the company, the people say.
While no doubt Netflix still prominently features data as central to many of its audience-focused experimentation and other key decisions, it’s interesting to see where Netflix’s seemingly all-powerful algorithms have limits. While computers may have gotten better at sniffing out under-the-radar topics and given further credence to the popularity of beloved actors and actresses, it has yet to capture the nuance and importance of interpersonal relationships, and how to navigate the dynamic, highly interconnected Hollywood environment.
On the other hand, the majority of Netflix’s kowtowing has been to established and renowned Hollywood players in an attempt to win them over and convert them to seeing the tech giant as a worthy home for its work. Recent projects by the Coen Brothers (The Ballad of Buster Scruggs) and the restoration and resuscitation of Orson Welles last project, The Other Side of the Wind, as well as the related marketing and promotional spend, are moreso vanity projects than ones expected to generate a significant return. In both cases, the films were released in theatres via short runs in select theatres, hardly the reception that would accompany a similarly high profile release by an established studio. One could argue that over time, as Netflix and other streaming-first services continue to grow in power and incidence, that there will be fewer and fewer auteurs and agents to win over, leading to a reversion to the algorithms and data that will only get better and more accurate over time.
As engineers and product heads likely argue, over the long-term, the results of the algorithm are likely to win out and result in more successful outcomes. However, it’s fascinating to think through the (temporary) limits of these algorithms, and/or whether this is yet another aspect of life that is at risk of being supplanted by machines in the future.