Yesterday’s article may have been a let down for some folks. I promised a big, deep analysis, and it only included four charts!
Well, time to get on with it. With a ton of charts. Over twenty! But first, some more throat-clearing, explaining Nielsen’s data and dataset. And then trust me, it’s charts, charts and more charts.
(My plan for this series is to include every section/part in one long article. The first article, “The Data Is In: Theatrical Films Massively Outperform Straight-To-Streaming Films” will keep getting updated, with a giant table of contents that will allow be to quickly read the post or go back to old sections. So bookmark that page or check back in there.
As for this week’s streaming ratings report, I don’t want to overwhelm your inbox, so it will go out Monday morning.)
Part II – Data Set 1: The Nielsen Results
First up, a brief explainer on Nielsen’s data, especially for new readers. I’ll try to quickly explain the data, then dive right in!
About The Nielsen Data
Nielsen provides weekly top ten lists for film and TV, with the total hours viewed for each title in a given week. Nielsen started publishing this data in August of 2020, and I have their data (with their permission) going back to March of 2020. Starting in 2021, they released a separate top ten list for film.
My Nielsen data set covers 158 weeks. In total, since 2020, 247 films were released straight-to-streaming and 49 were released in theaters. In 2022, streamers released 107 films “straight-to-streaming” and 29 were released in theaters. (In addition, Netflix had nine “Pay 1” films that skipped theaters, basically straight-to-video releases.)
That said, a few films trip up this analysis. Again, the question is “theaters or no?” but a few films that came out in 2020 in particular had crazy release dates, including many “day-and-date” releases, such as Raya and the Last Dragon and even Halloween Ends in 2022.
Netflix has given some “in between” titles as well, such as Glass Onion which had a seven day run in theaters, so when I’m categorizing these movies, you could pick and choose which column to put them in. (Theaters? Or streaming?) For the most part, if a film appeared in more than 600 theaters with a marketing campaign, it’s a theatrical title.
As a reminder, I’ll categorize things in three “window” buckets:
First Run – Meaning the first day something is available
Early – Meaning any day between 2 and 60, but usually 18 days.
Pay 1 – Day 60 to the end of year one.
Some Nielsen Biases To Be Aware Of (And How I Will Account For Them)
Yesterday, I explained three sources of variability in this data: the films themselves, the marketing styles, and the streamers streaming the films.
How we measure performance has its own set of biases to account for. First, Nielsen uses “total hours” viewed, and for certain titles, that’s going to be naturally higher than other films. The ur-example is kids films; kids tend to rewatch the same film over and over (see Encanto), so this factor can inflate the total hours viewed.
(I’ll try to provide different data cuts to account for this. Again, I’m trying to cut the data in as many ways as possible, even if some of those data cuts don’t support my thesis.)
Nielsen has other biases too. For example, Nielsen provides their data in weekly increments, so a film that came out on a Monday could wrack up more viewership than a film released on a Friday. To account for this, I’ll use the total hours over the first four weeks. (I looked at measuring longer time periods, but 80-95% of the viewership of given films is in that first four weeks.) This should mostly handle films released on different days, given viewership decay over the weeks.
Other folks like to divide films by their run-time to account for the length of titles, but I think that implies a “completion rate” so I won’t do that. This is often called “completed views estimates” or something like that, but implying that we know a film’s completion rate is just a data leap most analysts should NOT make, in my opinion. Plus, anytime I’ve compared total hours to length of a film, the overall charts barely change! (In a few articles from now, I’ll straight up tell you the difference in length between theatrical and streaming-only films; it’s minimal.)
Okay. That’s it. Finally, onto the data.
All Films Since March of 2020
Here’s the first look of the day, the average for all films that made the Nielsen charts since March of 2020: