We are all familiar with places like YouTube and Netflix recommending videos based on our viewing history. If you watched The Silence of the Lambs all the way through, there is a good chance you will enjoy Red Dragon, and so Netflix may well add it to your recommended list. Did you enjoy the US Office? Well, here are a host of other US sitcoms you might enjoy.
This strategy has been prominent on all video streaming services for a long time. But Netflix’s use of big data goes above and beyond what other companies have previously done, and the reason for this is the sheer amount, and depth, of data that they store, and how they use it.
Do you watch a lot of TV shows with a predominantly female cast? Then it’s likely you saw a trailer for House for Cards focusing on the women in the show. But, if you like David Fincher’s films such as The Girl with the Dragon Tattoo or Fight Club then you will have seen a different House of Cards trailer – one that focused on the fact that he was the director – in an attempt to appeal to your particular viewing sensibilities.
In fact, there were numerous different cuts of the trailer that were shown, each aimed at an audience with different viewing preferences.
And Netflix’s data analysis goes even deeper than just what you watched. They can see how you watched.
Did you take 4 days to watch a film because you kept pausing it to do other things? If the answer is yes, then you probably got bored with the movie, so Netflix won’t recommend films that are similar.
If you stopped midway through the film, Netflix can find out why. Did the tone of the film change? Maybe it got a lot darker at the one-hour mark and you stopped watching at that point. If so, then they know that you didn’t like it from there, and as a result, won’t recommend films that are dark.
The Netflix movie Bright is a prime example. It was hated by the critics, but the audience thought it was bang-average. Netflix used the viewing figures, demographic and preference date to plot out a roadmap of exactly how many people would watch the film, based on its plot, themes, and similarities to other films.
Which means that the people who watched it thought it was OK – just as Netflix always knew they would.
Stories like this just go to show the power big data can have and its importance when it comes to understanding our audiences, and how our products will perform