How YouTube Data Can Help Us Make Better Movies
This post was originally written for The Jungle, a weekly newsletter about the business of YouTube. Subscribe here.
By Sarah Ullman
YouTube gathers copious amounts of data about every interaction you have with each video on its platform: how long you watch, how you discovered the video, whether you rewind the video, what device you’re watching on, whether you click on the “endslate,” etc. As a professional working in the YouTube ecosystem, the ability to glean insights from YouTube analytics is one of my most important skills.
- This video has a good audience retention rate (71%); the expectation set for the viewer by the video and the title was maintained by the video content itself. In other words, people clicked for a reason and they were satisfied after they clicked.
- Every “bump” shows a place where the audience replayed a moment in the video. This is a good way to see what jokes are working, what moments people enjoyed, and what you should do more of.
- The general downward slope of the line is normal; people will drop off as the video goes on.
- The steeper downward slope at the end of the video is also normal; this is usually the place for an “endslate” or end card with clickable links.
- These insights are much more valuable in the context of the actual content and channel (which I am unable to share here).
Data is a huge part of the reason why YouTube content is so wildly successful: we can analyze and iterate.
Imagine if we had this kind of feedback for feature films! The test screening process wouldn’t be so arduous, fraught with filmmaker vs. studio battles about endings or specific edits or cuts.
However, the “lean back” nature of the feature film experience is quite different from the YouTube platform, which is engineered to compel an extended watch time session. Netflix maximizes their user data by making informed programming decisions, but most of their library of feature films were not intended to be watched on a computer screen. Furthermore, the data is only available long after it would be useful in the filmmaking process.
If studios continue to make films intended for the theatrical experience (will they eventually accept VOD day-and-date, and if so, design the content for that user experience?), there must be a way to gather information about an audience’s engagement. This query led me to an interesting Fast Company / Wired “neurocinema” rabbit hole:
- How Movies Synchronize the Brains of an Audience
- Rise of Neurocinema: How Hollywood Studios Harness Your Brainwaves to Win Oscars (Written in 2011)
- How Your Brain Can Predict Blockbusters (Written in 2013)
A cursory survey of studio exec friends revealed that most test screenings still use a paper survey rather than any of the methods described above. Of course, trailers are hosted on YouTube, and I’m sure those insights are useful when crafting marketing…but what is the feature film development version of YouTube analytics?
Theory: analytics vs. neuroscience = actions vs. reactions.
Thinking something doesn’t necessarily mean that you’ll do something about it, so if you hate a moment in a movie, it doesn’t mean you’ll take a negative action. It’s not as likely that you’ll leave the theater during a movie you don’t like, but clicking to the next YouTube video is an expected and engineered part of the YouTube experience. Analytics is a collection of information about actions taken, but emotional reactions are an involuntary human expression.
Of course, there’s an ethical rabbit hole here, too. How much information is appropriate to gather about someone without their knowledge? Are our emotions and physiological responses fair game for creators to use?
No answers this week, only more questions. There must be a way to apply compelling data analysis to the theatrical film experience…please let me know (@thesillysully) if you have the ingenious solution!
Sarah Ullman (@thesillysully) is a writer and creative consultant focusing on the YouTube ecosystem. She writes a weekly newsletter about the business of YouTube called “The Jungle” (subscribe here) and