Netflix, Bruce Willis and the big-data mystery not even Perry Mason could solve
A view from Russell Davies

Netflix, Bruce Willis and the big-data mystery not even Perry Mason could solve

Are you old enough to remember Perry Mason?

He was a bulky detective in one of those classic, rather slow US crime series, the ones you can watch over and over again on channels with names like Murder or Surprise. You’re probably not old enough, are you? Well, here’s why you should care anyway. The other week, Alexis Madrigal performed a splendid act of writing and coding in a post on The Atlantic website called How Netflix Reverse Engineered Hollywood. It’s a brilliant illustration of what happens when big data meets entertainment.

If you’re a Netflix user, you’ll know every film is slotted into a series of precise "micro-genres". As Madrigal puts it: "Some of them just seem so specific that it’s absurd. Emotional Fight-The-System Documentaries? Period Pieces About Royalty Based On Real Life? Foreign Satanic Stories From The 1980s?" This is one of the ways they’ve got good at recommending things. It’s a massive part of their competitive advantage and they’ve put considerable effort into generating these micro-genres.

Every film and TV show in their database is tagged by people who are specifically trained to be good at tagging entertainment. It’s especially useful because it combines what people are good at (watching a movie and deciding if it has a happy ending or not) with what computers are good at (crunching through huge numbers of calculations and finding patterns hidden from human comprehension). But there’s a little more to it than that – which is where Perry Mason comes in.

On the top of the favourite directors list, alongside Woody Allen, is Christian I Nyby II. And no-one really knows why

If you reverse-engineer all this data, you can arrive at a list of the favourite actors on Netflix, and the list is mostly as you’d expect it – Bruce Willis and Robert De Niro feature prominently. Except that, right at the top is Raymond Burr, the actor who played Perry Mason. And right on the top of the favourite directors list, alongside Woody Allen and John Woo, is Christian I Nyby II, a little-known director of some Perry Mason episodes in the 80s. And no-one really knows why he’s there.

It doesn’t really affect the system – you won’t get recommended tons more Perry Mason than you’d like – but it’s in there somewhere; it’s an artefact of all that complexity, of the humans and the algorithms working together to produce something too arcane for either of them to totally grasp. That’s what big data looks like on the ground. It might bring you greater understanding; it will also bring you greater mystery. Have a look at that piece – it’s what the future’s going to be like.

Russell Davies is a creative director at Government Digital Service