By providing your information, you agree to our Terms of Use and our Privacy Policy. We use vendors that may also process your information to help provide our services. This site is protected by reCAPTCHA Enterprise and the Google Privacy Policy and Terms of Service apply.
Netflix‘s intermittent, self-reported “ratings” are famously met with scrutiny. But supposing the streaming giant — or any of its established peers, such as Hulu and Amazon — did cave and regularly disclose ratings for their “binge” releases. Has anyone thought about what that data would or should even look like?
For a long while, Netflix staunchly refused to disclose any viewership info, which was fine. They don’t have advertisers, only subscribers, and as such “ratings” don’t demand public dissemination.
But with the 2018 release of the Sandra Bullock-led suspense film, Bird Box, Netflix broke its own rule and in turn opened a bit of a Pandora’s Box, as many observed back then. “Took off my blindfold this morning to discover that 45,037,125 Netflix accounts have already watched Bird Box,” read the company’s late-December tweet, which further characterized that tally as the “best first 7 days ever for a Netflix film!”
We then learned how Netflix defines a “WATCHER” — a member household that watches “70 percent of a film or single episode of a series.“ And that’s a critical parameter, especially on the TV front: 70 percent, of just one episode.
(Netflix terminology also defines STARTERS as households that watch two minutes of one episode of a series, while COMPLETERS watch 90 percent of a season of a series within 28 days.)
Much more recently, Netflix said that Stranger Things Season 3 had “64 million views” in the four weeks following its July 4 release — worldwide, it has to be noted. We were also told that more than 18 million member households finished that eight-episode season within a few days. Meanwhile, Nielsen, which reports ratings for broadcast and cable TV and has tried to independently estimate Netflix numbers, said that Stranger Things 3 averaged 19.2 million viewers — though that is in the U.S. alone and not counting all manner of devices/viewing options.
For those who knit their brows over ratings, wherein lies the most “true” number? And what exactly should that number measure, for a Netflix/Hulu/Amazon series that releases all episodes at once?
Does a household’s viewing of a fraction of a single episode mean anything? (If it does, imagine the number an ill-fated broadcast sitcom could tout. “Ten million people almost made it through the Sunnyside pilot!”) Should the rating be the average audience for all episodes, or should the viewership for each episode be reported separately, to show fluctuations/trends as broadcast and cable numbers do?
But if so, how long should a streamer give people to watch any given episode(s)? That first weekend? A week? A month…?
To be clear, the aim here is not to compare GLOW to NCIS. No amount of noodling the numbers will ever create an apples-to-apples situation. But having regularly disseminated streaming “ratings” could at least let fans of those shows know how their favorite is faring against the field. So that renewals can be better anticipated, and cancellations would come as less a shock.
If you ran a streaming service, how would you define the “ratings” for a binge release?
Because SOME type of metric could be refined over time to give a clear picture of who is watching what. Or if they are watching.
.
Or if the company is a giant money laundering operation pouring money into unwatched, yet miraculously ongoing, projects.
I see you still haven’t leaned what money laundering is. Hint: the money has to have been obtained illegally in the first place. There’s no suggestion online that Netflix do anything illegal. Simply shifting funds between jurisdictions doesn’t cut it.
You’ve been quite insistent about it though. Did Netflix run over your dog or something?
There’s more than one way to launder money. But, a company like Netflix that has a constant, large influx of “clean” money that you could use to hide the “dirty” money would be one way to do it.
I think what’s actually going on, is Netflix is taking large amounts of investor money, putting it into a big pile, and setting it on fire. Nice work if you can get it.
I think if regular ratings/viewership data for streamers was a thing, it could never just be a single number. I think it’d have to look something like:
• how many people watched 1+ episode the day of release (speaks to level of anticipation, also if the streamer’s internal promotion, such as on people’s main pages, works)
• how many people watched 1+ episode within 3 days of release (since many big shows are dropped on Thur/Fri, and lots of people don’t have time to watch much on weeknights, I think this would be just as important as same day)
• how many people finished the whole thing within 7 days
—
The above would speak to how much a show is in the cultural zeitgeist, as it were. After that immediate release of the show, I think continuing to parse on smaller time increments stops losing its meaning (eg 5 days, 7 days), just because there’s so much content now. It seems if something’s going to hit, you’ll probably know right away. So next interesting metric might be completion rate at 4 weeks.
—
For me, the whole idea of watching 70% of something is rather meaningless. I would only be interested in completion numbers. How many people watched 100% of an episode. It might be interesting to know how many people completed 100% of a season, and for those who didn’t where they dropped off.
The 70% is likely chosen in case someone skips the credits.
Skips the title sequence, the credits, scenes with people you hate, etc. There are lots of reasons why a 70% watch would be a better category. The Neilsens actually record drop-offs.
Netflix’s actual renew/cancel algorithm is probably many pages long and they’ll never tell us what’s really going on (so Apple and Disney never find out) but it must include some metric for capturing how much a show/movie contributed to the decision to subscribe in the first place and how much a show/movie contributes to the ongoing decision not to cancel.
Probably there’s a third factor, for former subscribers who come back. Let’s say GLOW correlates strongly to a return visit from a former subscriber, who watches and then bails. Now Netflix knows that a show like GLOW can keep people engaged and the need is to keep them from bailing inbetween seasons, so…make more GLOWs and scatter them throughout the year.
The best thing about Netflix is they don’t have to care about that sweet spot the 15-45 demo or whatever it’s called since they don’t do advertising they can focus on total viewers
Neither does HBO. We still know what the viewership numbers are.
Because HBO is owned by a mainstream media company and Netflix is not. Simple as that.Releasing audience measurement is a no-win proposition for HBO. Get over it. Ratings are grease for the advertising cogwheels and nothing else.
Get over what? HBO is owned by a mainstream media company and Netflix is a mainstream media company. Your comment and argument make no sense.
It actually makes no sense that HBO ever reported viewership figures so there must be some corporate reason behind it. Maybe it has to do with the cable business model: Time Warner wants backup in negotiations with cable providers, to get more money for their shows that are valuable and highly watched. Why would HBO bother to make a Game of Thrones hit show if it can’t somehow make money off the hit status? Getting more money from cable providers is the way HBO profits, and do to that, it has to provide some proof that Game of Thrones is helping Comcast and AT&T retain their own customers.
Anyway Netflix has no advertisers or cable providers, so: no motivate to report viewership.
If they release full season at once they could count those who finished it in first week. additionally they could add how many finished it in a month or show how many viewers watch each of those episodes separately. It would be better to know how those shows are doing. Now I can only guess that if everyone is talking about the show is successful and if you hardly hear of it, it probably be canceled
Netflix’s “ratings” would probably be similar to YouTube’s view count.
The Neilsen ratings are incredibly skethchy these days – certainly something that no scientist would ever use. They’re just guessing at this point. That makes it pretty silly for anyone to point fingers.
How are they “sketchy?” They’re randomly sampled households that represent a specific number of viewers in their area.
Streamers can give accurate numbers if they chose unlike the estimates we get from Neilsen for network and cable shows. Demos mean nothing in streaming, so everyone’s tastes gets equal consideration. But, you can boast when nobody knows the truth
Netflix has no motive to tell the truth. There are no advertisers to report numbers to, and investors just care about subscriber numbers, not viewing numbers. Investors might want more data on what show/movie types attract and retain subscribers, to make sure Netflix is making the right kinds of shows and movies, but worrying about this or that specific show or movie doesn’t seem very important in the bigger scheme of things.
Anyway, you can sort of guess at this. When Netflix keeps churning out content aimed at young females, or Spanish-language content, that indicates they are seeing good results for those things.
One major difference between broadcast and streaming has got to be completion rate vs ratings. Nielsen ratings go up and down each week but I kind of doubt people would skip a random episode of Stranger Things or Ozark because it is right there. On broadcast you have to go searching for a repeat (rare these days), on demand or a streaming app if you miss an episode. That’s why the 90% completion rate is so important. They would be better off renewing a show was watched by less people who actually completed the show then a bigger one where everyone tapered off before the end.
I do not necessarily binge-watch. I watch one episode a week, but sometimes may skip a week or two for whatever reason. If a series is 10 episodes, I would release viewership info after ten weeks. Eight episodes, then eight weeks 13 episodes, then 13 weeks. Since I do skip a week or two, then maybe add about three weeks to those numbers.
Binge-watching info may be interesting to know how many subscribers actually do compared to those who do not, but a show’s fate should not be determined on that alone. That’s idiotic. Also, knowing the show is, likely, always going to be there, there is no rush to watch. I have yet to see Stranger Things season 3 and may wait for season 4 to have a longer run of episodes to pair with other shows to watch that night of the week.
I would look at overall viewers, how steady the viewership is, and/or if viewership noticeably drops after just one to four episodes and does not really bounce back after those eight, ten, thirteen weeks. I’ve yet to finish Luke Cage with four episodes left to go and I stopped some weeks ago. I do plan to finish it off at some point.
I also wonder what actually causes a streaming show to be cancelled? 10 mill, 5 mill, 2 mill or less than a mill? And what plays into them picking up a cancelled network series, which is rare.
my guess overall views from around the world.
Knowing how many people watch in the first week is important, as it certainly signifies how much interest there is in a show. But it can never be compared to non-streaming services. Because even if, say, a “mere” 8 million people watch a show in its first season, the second season could get some buzz before airing and then suddenly 10 million more people watch Season One 10 months after it aired. With a streaming service where original content is always available, that show is yet another reason for people to come to the service. Regular TV with its air-once-and-once-in-summer-if-it-was-popular-enough method, original airings are far more important.
Quote “But having regularly disseminated streaming “ratings” could at least let fans of those shows know how their favorite is faring against the field. So that renewals can be better anticipated, and cancellations would come as less a shock.”
I wonder if this article is more about TVLine’s renewal scorecard than anything else?
Furthest thing from my mind. (Streaming scorecard doesn’t make predictions.)
Yeah in the absence of advertisers, viewing numbers don’t mean much. It’s just another garbage statistic that sites like this use to get an audience for their own advertisers.
Think about it: if millions of people watch Friends because it was familiar and they were bored and too lazy to look harder, is that really the same as somebody who hears about GLOW and thinks it sounds good and checks Netflix out because of it? Friends might get a lot more views, but GLOW is working harder for Netflix in getting and retaining subscribers.
Netflix has hinted at something I’m sure they have, internally, and will never ever share with the outside world, because it’s way too crucial as competitive info they don’t want Disney, AT&T, Apple, etc to see: namely the true score they use to decide cancel/renew, and this is not the same as simply viewing numbers.
I think this score comes from two factors, which I’ll call “grabbiness” and “stickiness.” Grabbiness is how important a show or movie was to the original decision to subscribe. Stickiness is how important a show or movie is to the decision not to cancel at any given time.
Netflix should have enough data to come up with good estimates for both. For any given show/movie, does it over- or under-index as the first thing a new subscriber watches? Over-indexing should be an indication that it factored into the decision to subscribe: grabbiness.
At any given time, there’s an X% chance that any subscriber will cancel. If viewing a show/movie correlates to a lower than average cancellation following the viewing, that could indicate stickiness.
That’s just a couple of ideas how these things could be sussed out of the data. I’m sure Netflix data scientists have a sophisticated algorithm that factors in many things, including the importance of grabbiness vs stickiness (grabbiness should be more important, since it’s harder to get a subscriber than to retain one).