On December 12, 2023, for the first time, Netflix released viewing hours statistics to the public covering the first six months of the year for thousands of titles, not just global and regional top 10s during the past week. For those of us, ahem, of a certain age, these numbers look very different from the Nielsen ratings published in our 20th Century local newspapers. Back then, it was all about the “time slot.” TV audience members watched specific episodes of specific series at specific times. Sure, it was possible to record content using a VCR for viewing later (and there were attempts to measure such consumer behavior) but, in many living rooms, the VCR was blinking 12:00 and only used to watch movies rented from a nearby video store.
Nielsen still releases data to the public on a weekly basis, but it doesn’t mean the same thing. Back in the 1980’s, viewers who watched Season 3, Episode 4 of “Family Ties” last week were probably going to watch Season 3, Episode 5 this week. Furthermore, time slots have gone the way of the dodo. A “Punky Brewster” viewer tuned into their local NBC station was almost certainly excluded from watching the first half hour of “Ripley’s Believe it or Not!” on ABC. Again, there were VCRs, but even the best models were tough to program, and playback was at significantly lower resolution than over-the-air reception.
So, in today’s “binge anytime” viewing environment, and with Nielsen still around, what can we really learn from Netflix’s stats? Well, one of the first insights came from an article in The Hollywood Reporter entitled, “Inside the Netflix Data Trove: ‘Wednesday,’ ‘Suits’ and 10,000 Titles (Almost) No One Watched.” That got me thinking. We hear a lot today about income inequality, but what about content inequality? Is Netflix’s success based on just a few titles while much of its library languishes?
According to Netflix’s report, the top 1% of titles garnered 27% of total viewing hours. Zooming out, the top 10% accounted for 71% of viewing hours, and the top 20% reeled in 86%. This statistic could lead to questions regarding acquisition costs for bottom 80% of Netflix’s library, but maybe the situation is unavoidable.
You may have heard of the Pareto principle, colloquially known as the “80/20 rule.” It has spawned some popular sayings including, “You wear 20% of your clothes 80% of the time,” and “20% of the employees in any given enterprise are responsible for 80% of its output.” What if the 80/20 rule has always applied to content libraries, but nobody noticed until now? It could very well have been true at Blockbuster, or Borders, or Tower Records.
Next, what do the Netflix stats say about the global entertainment industry or, more specifically, the entertainment industry outside the English-speaking world and where it stands relative to Hollywood? What would happen if I ran through the list of titles, detected those with non-English characters, and totaled up their viewing hours?
It turns out that by my estimate, 21% of viewing hours are spent on non-English language films. This is not insignificant. It means that not only is foreign-language content gaining a foothold in English-speaking markets (think “Squid Game”), but viewers in non-English speaking markets can choose between watching content from Hollywood and content from elsewhere even, in some cases, content produced locally. Five years from now, where do you think this number will be?
Finally, for all you Korean Wave and K-Drama fans out there, I found a Hangul detection function and used it to estimate that Netflix subscribers spent 8.4% of their view time on Korean content. This is a massive cultural impact considering South Korea has only 0.6% of the world’s population.
So, what’s next? Even if Netflix doesn’t add any data fields to its next six month viewership report, it should be possible to generate a decline curve for content, similar in concept to that for an oil well, which will provide a formula for how individual title viewership decreases over a significant period of time. Stay tuned!
