If its ratings don’t really matter, why is Netflix canceling many shows? | Hacker News Reader

If its ratings don’t really matter, why is Netflix canceling many shows?

marketwatch.com

My assumption is that for Netflix (and similar) it isn’t “ratings don’t matter” but rather “competitive ratings don’t matter”. Traditional tv/cable has to compare ratings for multiple shows at the same air time. If a giving show has a lower rating than other shows at the same feasible time - they need broad appeal because the opportunity cost involved in burning a specific time slot.

Basically say you have a few shows, 50% of people like A, 40% each like B, C, and D - note that there’s an overlap and underlay. If there’s a a single time slot available, the only “correct” solution is to run show A for the 50% popularity. If all the people who like D do not like A, then that whole group (40% of the market) have no compelling reason to pay for your service.

Netflix doesn’t have the time slot problem which means the shows are not competing. So they can provide A and D and both groups are happy, and B and C are also available to everyone.

But importantly making a show does cost money, so there’s still a minimum popularity you need to justify that cost.

Agree.

Since it's all video on demand, ratings don't matter, but popularity still matters. Production cost also matters.

As long as there's still a segment of the demographic that is watching, and it's comparable with the production cost, the show shouldn't get canceled.

Traditional TV doesn't have the strict time slot problem either. Ratings are counted as day+3 - meaning live + 3 days on demand.

You can think of Netflix as following a third generation of ratings.

The first generation of ratings was just raw data: how many TV’s were tuned into a given show at a given time? More viewers, more ad revenue. Simple as that.

The problem was that some viewers are more lucrative than others. The second generation of ratings added demographic data. This happened sometime in the late 60’s/early 70’s, just in time to discover that the recently-cancelled Star Trek, despite its low ratings, appealed to a very lucrative demographic.

Netflix has user-by-user, viewer-by-viewer ratings. Not just statistics per the old Nielsen sample, but the entire population of data. Maybe even minute-by-minute (unless you doze off). This is way more powerful than demographics, because instead of targeting shows to “18-35 middle class white urban women”, for example, you can target shows to “people who binge watch space operas” or “people who enjoyed only the first two seasons of <<some show>>” or any other large enough group you can tease out of there. And Netflix might only be scratching the surface of what is possible with this data.

> Netflix has user-by-user, viewer-by-viewer ratings.

(Minor nit: They don't necessarily know how many viewers are watching a screen)

But maybe that's because they don't care, since, does it really matter? If they're not running ads, isn't every unit of screen-time worth essentially the same amount, ignoring the variance in subscription prices for different regions and resolution options.

Actually, I'd go so far as to argue that viewer time is worth, incrementally, a negative amount. I'd expect their goal is to have just enough content to keep subscribers interested, rather than to maximize viewing time.

As such, it seems to me that their viewer data goes beyond being a different generation of ratings, into a completely different thing. Maybe it's closer to box-office (seats, not revenue) statistics for movies.

Now, I know they're not actually ad-free, what with product placement, in which case traditional "ratings" can translate to real money, but I'm not sure if that's a large enough proportion of revenue (or reduction in content production cost) to change their behavior.

> Actually, I'd go so far as to argue that viewer time is worth, incrementally, a negative amount. I'd expect their goal is to have just enough content to keep subscribers interested, rather than to maximize viewing time.

That might be going a LITTLE far. Netflix subscribers who don't actually watch Netflix not only don't give Netflix enough data to make programming decisions from, but I'm willing to bet that an engaged Netflix subscriber is much, much, much less likely to cancel than someone who barely ever watches Netflix. If you think of viewer time as inversely correlated with probability of cancelling, the obvious metric to optimize for is viewer time. You're never going to watch so much Netflix that the marginal infrastructure cost is meaningful to them.

Keep in mind, though, that I mentioned incremental value, though maybe what I meant was marginal?

I do agree that NF's business model can't really be compared to that of a gym, with a goal of maximizing subscribers that (eventually) don't use the service at all.

One of the big mysteries is how good a deal did they got for that outsourced delivery infrastructure, as well as how much of it is metered (and how) versus flat-rate.

Because Netflix can't fund infinite shows, and needs to allocate their resources that will attract and retain customers. It's not as if the crappy show producers will let Netflix keep the shows for free.

But maybe Netflix's advantage over a traditional TV channel like HBO is that it can offer ten shows 1 million people will watch instead of one show 10 million will watch?

Traditional channels have limited slots. Obviously there's costs to produce shows, but making 10 small budget shows that attract 10 different types of viewer might be a better way to lock in subscriptions than a small number of "prestige" shows.

Possibly. But there's also a dynamic on Netflix (and HBO, etc.) of must-see-TV shows (to steal a term from an earlier era) that pull someone over the hump of signing up and then staying. You need the "filler" (which isn't really intended to be disparaging) too. But a large number of shows which are OK/fine doesn't get people to pull the trigger. Game of Thrones, House of Cards (originally), Man in the High Castle, Orange is the New Black, etc. do.

More to the point: Netflix can't fund infinite shows, and since it's relying on total subscribers vs. ad impressions, they need to make sure they're getting as little audience overlap in their programming as possible.

They can have big, expensive hits, or they can have niche programming. The model falls down at the decently-viewed, but broad appeal, shows because they won't attract a specific audience but also aren't popular enough to be destination viewing.

That's an interesting way to think about it - similar to β and α in the portfolio analysis sense.

I don't know how well that model holds up in practice though. To use HBO Now as an example: when Game of Thrones was the only show I was interested in, I would sub for 3 months to watch a new season live and then cancel. Once I caught the Westworld bug I was that much less likely to cancel my subscription during the down time.

Defaults are powerful too. I admit that I signed up for HBO for GoT though also watch other things (esp. Westworld). I could probably optimize month-on/month-off but I don't bother.

In general I agree (see: Linodes and D.O. droplets, domains for possible side projects) but I'm pretty quick to cancel extraneous streaming services now that every media company is trying to get their own $10/mo IP silo off the ground.

Yeah, I'm pretty conservative about signing up for new entertainment services--especially given how few hours per week I actually watch/listen.

HBO's model actually somewhat aligns with Netflix's, only it's far more focused (and tends to be less ambitious).

They want a single big hourlong prestige drama as a tentpole (Game of Thrones, Westworld; in the past, True Blood, The Sopranos, The Wire), and a pair of highly-regarded and experimental comedies as more niche programming (Veep, Insecure, Girls). Keep subscribing to watch the big tentpole, and maybe get hooked enough on one of the niche shows that follow it to keep your subscription running in the times when the big shows are off. A show like Brooklyn 99 wouldn't survive very long on HBO, only because its style of comedy isn't catering to a specific niche (compared to something like Silicon Valley or Veep) and at the same time doesn't have a broad enough appeal to act as a tentpole.

You only need 1 good show that would encourage people to keep their $10/month Netflix subscription. So I assume that extremely high ratings has no particular advantage, but they need to have very high breadth in terms of population coverage, so that almost everyone has a reason to get a Netflix account. At that point, you can calculate how profitable a TV show is, based on number of users vs cost.

If you have a small but very targeted show that costs $1,000,000 to make, but it attracts 10,000 users who would otherwise never get a Netflix account, then the show is profitable after 10 months. If you have a show whose demographics overlap preexisting shows, then its value decreases, unless you can have it premier at different times of the year.

Don't think if the show as "cancelled". It's the fact that Netflix won't invest money in paying for another season of the show to be made. They want to invest money in shows people watch the most and will keep them on the platform. So yes, ratings do matter. But in sort of a different way.

That sounds like exactly the same sort of way to me.

That's the longest possible way of saying the following two words, "ratings matter".

The marginal value of a season with good ratings matters. The content is there forever and two seasons of it vs three might not impact subscriber numbers as much as two seasons of it and one season of a brand new show.

This.

Kids shows are a great example - my kids don't (when pressed) care that they don't have season 4 of Octonauts, they do like seeing entirely new stuff.

For the price, I'm happy with what we get from Netflix. I much prefer it to Amazon Video where I'm not sure if I'm going to have to pay extra because my kids throw a fit (because it's visible but not on Prime/free).

It's a more subtle point, I think. It's not correct to say "ratings matter", because in this context "ratings" means something very specific in what's being measured and how. That methodology and measurement are largely irrelevant here.

It's also not correct to say "ratings don't matter", because rather implies that popularity and customer opinions are also irrelevant. Which is clearly false.

To my knowledge, no one has ever said ratings do not count.

In the article, there was no evidence to indicate that anyone at Netflix had made the claim that ratings do not count.

All they said was that they didn't release ratings, but everyone seems to know that ratings count.

So the article's title is misleading. It should say "As we all know, ratings determine whether a show is renewed."

I'd enjoy a news source that phrases things like that all the time.

"Ever since the American-Norwegian crime drama “Lilyhammer” became Netflix’s NFLX, +2.32% first original series to premiere in 2012, the company has been committed to voraciously generating new content in addition to its huge library of existing TV shows and movies."

"huge", my posterior.

U.S. context: In the last couple of years, aside from a very few new headliners (e.g. "Rogue One"), and some lingering headliners from older contracts (e.g. "West Wing", Star Trek), and ever-fewer good independent productions, I find Netflix's streaming library to be the dreggs of barely acceptable quality. (People will watch it when there's nothing better.)

Netflix streaming seems to be more popular than ever, but, "Jessica Jones", "The Crown", and a few other Netflix productions aside, Amazon Prime is bringing more engaging content into their lineup than Netflix is. (And that isn't saying too much. Most recently, on Prime, I'm enjoying re-watching the old "Kung Fu" show from the 70's -- in good if not spotless, widescreen prints -- apparently, original filming was with cinema-format cameras.)

There are still some worthwhile titles on Netflix streaming (U.S.). But after a few months, you've seen them all.

For those of you outside of the U.S., for all that your catalogs are smaller -- so I read -- you should know that, in the U.S., almost no new major films end up on Netflix. And almost all of the older ones are now gone -- contracts expired. Per what I read and hear, some do still end up or remain in other countries' NF catalogs.

Absolutely. They're down on anime and older TV shows as well. Most of their new "Netflix original" content have been Chinese and Korean movies with rather limited appeal.

The language used, strategy and even just context of TV Broadcast/Cable programming is this whole art/science/data mess that they are constantly tweaking to try and get an advantage.

Ratings "don't matter" to Netflix in the sense of "our single camera family based Thursday sitcom is 5 points higher than theirs but since they attracted more males age 20-35 they're pulling in more advertising".

Netflix is playing a different game entirely.

Because salience still matters. A show with few viewers who love it enough to want Netflix might get renewed. A show with few viewers all of whom think it's mediocre has no future in any commissioning system.

Also, I'd imagine it makes a difference how _replacable_ shows are. A show which is the only show watched by a user in a given month is probably more valuable to Netflix than a show that is one of seventeen shows that user watches.

I'm sure Netflix knows how much salience matters, but I'm guessing most people continue with their subscription as long as there is something good enough to watch.

I'm assuming Netflix's core KPI would be net subscriber count, and everything that Netflix does, in some form, would be driven by that.

Going further, I'm guessing that there are (or will be) some shows that show highly rated but producing more seasons would not draw in new subscribers or decrease retention rate. It might be that there are shows where their data shows that a large percentage of the fans of that show are unlikely to cancel their subscription, in the case that new season is not produced, and that a new season will not be the kind-of thing that will create new subscribers.

If wonder if Science Fiction might be a genre where the fans are unlikely to cancel their Netflix subscription, but a multi-season run in also unlikely to appeal to people that don't already subscribe to Netflix.

Yea, my guess is that the primary divergence is that Netflix probably "weights" ratings of shows that are correlated with new subscribers and subscriber retention. So a low-rated show that on boards a lot of new subscribers or seems to change the behavior of a cohort who'd otherwise be expected to churn would get comparatively more credit that a show that is slightly higher watched in a vacuum, but driven by subscribers unlikely to cancel their subscriptions in the absence of the show.

Basically, any show that can show that its continued existence will result in subscribers that would not exist otherwise would likely see less ratings scrutiny than shows that cannot show this.

Ratings don't matter to Netflix - subscriptions matter. The article hints at this but doesn't draw it out.

Netflix in general doesn't make its money from eyes on advertising - so ratings are not a good KPI for them.

Netflix in general does make its money from subscriptions - so shows that bring in subscriptions and those that are good enough to retain subscriptions are the KPI they will be looking at. I find it likely that they have metrics along the lines of what shows keep people clinging on to their subscription.

It seems likely to me that these two KPIs lead to very different conclusions about what shows to commission and cancel.

My guess is that they are related enough that most cancelled shows are on the low end of ratings. But I'd also guess that the relationship is nowhere near as linear as a traditional network. The key to renewal at Netflix seems like it should be proving that its continued existence results in subscriptions that would otherwise not exist.

Going wide with many shows instead of tall with a few shows with lots of seasons is probably way more optimal.

Netflix tries to produce shows for every audience segment instead of big mainstream hits. So the overall ratings don't matter as much.

Actually I suspect that this is what it's changed. It's likely that Netflix has abandoned the long tail. They are now focusing on shows that have immediate, short term mass appeal.

Previously Netflix was able to invest in "quirky" shows that had strong appeal to a small percentage of its user base. The classic examples here are indeed Sense8 and Marco Polo. These shows were not blockbusters but they had a strong and loyal following and that should be highly valued by Netflix. Particularly when it comes to original content and as a subscription service, Netflix should be primarily focused on loyalty above all. These are the shows that keep viewers subscribed and make unsubscribing painful.

But the economics here quickly become complicated as the culture evolves. In a world where (a) many (most?) people are already have a Netflix account (b) there are now many competing streaming services each of which have their own exclusives and (c) the concept of "television" itself is gone, now you just have "screens" all over the place and (d) there is an explosion (probably bubble[1]) of original content and amateur content and (d) there is an arms race for talent driving up salaries across the board... the end result is that Netflix is now engaged in a month-by-month war not just to defend the many subscribers it has accumulated but also to take (and retake) subscribers who may have been drawn away to other competing steaming services that have access to historic brands. Netflix no longer has the luxury of gradually building a large, deep catalog that appeals to small segments of the market. It must offer everything, or at least something extremely compelling, to all users, every month, or those users will churn and be drawn away by competitors.

Netflix is likely moving towards the blockbuster model. In this light it starts to make sense why Netflix is dumping hundreds of millions on feature films and also focusing on big shows that can command huge audiences. The blockbuster model is driven by a series of big events carefully scheduled throughout the year to get users to show up in large droves when you need them for something they don't really need. Competitors are doing the same. The streaming business meets the blockbuster model and now you've got to produce a blockbuster every month to keep people coming back.

This is crazy, of course shows get canceled. Ratings are a huge portion of why shows are canceled on cable because ratings are directly tied to revenue. I guarantee Netflix is looking at revenue when deciding how to spend their Original Content budget. Ratings would be a good indicator, but that's not the stat to chase.

It's not about ratings, it's about subscribers. New shows generate subscribers. Many subscribers stay on board with Netflix even after the show that got them to sign on has been canceled. Ergo the more new shows you have, the more subscribers you have.

It seems Netflix uses an undisclosed season-by-season ranking to determine which shows are continued and which canceled.

In many cases, I only watch shows that have positive reviews (e.g. IMDB) after 2 seasons or more. I don't know what percentage of Netflix subscribers are like me, but I do wonder if the percentage is high enough that many good shows will get canceled not because they will not be watched in the future but because they hadn't yet gained a following.

"The second season of “Flaked” was released earlier this month, but industry observers anticipate it will become Netflix’s 20th canceled original show."

When was this written? That aired last year. Also IIRC Flaked was one of those pre-canceled shows where it was announced in advance that the second season would be the last.

They clearly matter not only in viewership but also in the whole stars vs thumbs up system.

i think its interesting that they can tell if someone has finished all episodes of a series and would be waiting for a next one.

Yes. And the switch from "Will they sit through all the ads?" to "Will they watch the whole story arc and give a thumbs up?" is the best thing that could happen to the medium.

I mean, it would be a kind of crazy business model not to take into account what the ratings of shows were...

Then again, Netflix is bathing in an ocean of cash, so they just might be crazy enough to be throwing out anything and everything to see if it sticks.

Because subscriptions matter.

The benefit of Netflix is that they can leave filmed shows on their platform and cancel future production if interest wanes.

Then they can see if interest comes back and can build future shows based on this feedback.

because they are bleeding cash behind the scenes

Reply on news.ycombinator.com