Wait, what’s CuriosityStream? How can something I’d never heard off pass 10 million subscribers? Well, that mystery, and the rising importance of subscribers for evaluating success or failure combine in my “story of the week”.
Most Important Story of the Week – The Subscriber Numbers of Smaller Streamers
Let’s get something out of the way. Subscriber counts matter. As we enter the streaming wars, how many folks will actually pay streams for their services is a direct sign of how well they’re doing as a business. The intrinsic logic of this makes sense and it’s why we celebrate Netflix for getting 60 million US subscribers.
The trouble is not all subscribers are created equal. Even with the data we do have, comparing things apples to apples is always the tricky part. So let’s see what we know.
The Data Set
Again, the big new subscriber numbers were really DAZN, Curiosity Stream, and Hallmark Movies Now. They each had leaks to the press about new subscriber milestones. Combining with some past data I’ve collected (and updated) yields this table:
The problem is that as far as we know CuriosityStream is doing as well as Disney+. Or as Bloomberg put in its headline, it’s beating HBO Now. Which doesn’t feel right. Does it? To explain the problem, we need a quick diversion to the NBA…
The NBA Analogy
Why the NBA? Because it’s such a great testing ground for statistics. If my first rule of data analysis is: “Always show the distribution” the second may be: “Always use two variables to judge performance”.
Why two? Because one is too easy to game. Which we will see with subscriber counts. Why not more variables then? Honestly it’s because I have trouble seeing in three dimensions. Quad charts are the easiest way to reconcile these two ideas.
The NBA provides a great example for this. Take a player like James Harden. He’s a great shooter, and for those who don’t follow the NBA, we usually use percentages to describe this. Harden is a 36% three point shooter, meaning he makes 36% of the 3 point shots he takes.
The problem is if I just used percentages, well, it would look like a lot of guys are a lot better shooters than James Harden. Take Philadelphia 76ers rookie Matisse Thybulle. He shoots 46.8% from 3. That’s 10% points better!
The problem is how many shots these guys take. (I use three point shots on purpose, since percentage is used much more than shots attempted.) Harden leads the league by taking a whopping 13 three point shots a game, whereas Thybulle shoots just 2.2. The narrative behind this data is that Thybulle can choose opportune times to shoot, whereas Harden takes and makes a lot of tough shots.
In the NBA, the data/experience shows, the more shots you take, the harder it is to keep a high percentage. We can make a quad chart out of this, with the goal to get to the upper right:
Using this example, the best three point shooter is actually Davis Bertans, who is splashing 46% of this threes while taking a whopping 8.6 a game. See, two variables tell a better story than either percentage or total on their own.
So which number should we pair with “subscribers”?
When it comes to streaming services, we need a similar quad chart. The question is which one?
The biggest problem we’re trying to solve for is the idea that a subscriber who pays $0 isn’t worth as much as one paying $15 a month. If we’re looking at subscribers as a proxy to customer demand, price seems super relevant; lots of folks will happily subscribe to something they pay $0 per month for.
At first, I thought this would push me towards using “Average Revenue Per User” (ARPU). This is reported, by Netflix for instance. It has a few flaws, though, especially when trying to use it to gauge customer demand. The biggest flaw is that some customers don’t actually provide the revenue. Instead, someone else pays the bills. For Netflix, this includes T-Mobile subscribers who get Netflix “on them”. In that case, T-Mobile pays Netflix for the subscribers, not really the subscribers themselves. Further, even if customers do pay, the monthly price is actually split between distributors (like cable, Amazon and Apple). Some companies can negotiate this better than others. (While still relevant for business performance, it is less relevant when gauging customer demand.)
The opposite end of the spectrum would be to just take the price a company charges. There are two problems with this method, though. First, a company could offer a wide variety of prices. Disney+ can be purchased as part of a bundle with Hulu and ESPN+, or on its own. DAZN can be purchased monthly ($20 a month) or annually ($100 a year), which is a huge gap. Netflix has multiple tiers too. Updating my table from above, you can see how complicated this is:
Worse, those prices don’t account for the folks getting promotional offers. So Apple TV+ folks who got it because they bought a phone. Or Amazon customers getting Prime Video for free. Or Verizon members who get Disney+ for free. That doesn’t say much about how willing they are to keep the service without those free offers.
If I were king of the world, I’d try to marry these two numbers. It would be “Average Price Paid per Subscriber”. This gets to how many customers are actually paying full freight. (Heck, I’d actually love a distribution by dollar amount, but that’s just dreaming.) If you could combine this with total subscribers in a quad chart, THAT would tell you who is winning the streaming wars.
Implications for Streamers
Listen, if I were amazing, I’d give you that quad chart I just described. But I don’t have that data. While I may try to calculate it in the future, it would require too many assumptions for today. Just look at that pricing table above to see how much of a mess it is.
Still, we can learn some things from this thought exercise. Especially about how little we know, but how much we can suspect that subscriber numbers a misleading us.
To start, with CuriosityStream, they went from 1 million subscribers to 10 million in just a year. Likely, that’s a combination of the low price (only $3 per month), which very few folks actually pay. CuriosityStream has a deliberate strategy to be sold in cable internet bundles, meaning they give cable companies another discount just to boost subscriber totals.
I’d argue before CuriosityStream brags via Bloomberg that it’s beating HBO Now, or that it’s a winner in the streaming wars, it should tell us how much money it’s charging per subscriber.
Hulu is probably next. They just can’t quit offering promotional discounts. Even having lowered their price to $6 last January, they offered a Black Friday deal of just $2 a month. Really, all of Disney is hooked on promotional pricing as Disney offered lots of folks Disney+ at $4 a month for three years and there is that Disney+, ESPN+ and Hulu bundle I went nuts for back in August.
However if any wildly inflated number deserves the most scrutiny, it’s Prime Video. Since Amazon doesn’t even release Prime subscriber numbers regularly–good job SEC!–we only rely on selective leaks and estimates, with the current number at 100 million. But according to some surveys, only 11% of customers use Prime for the video service. So it’s really hard to say Prime Video is a 100 million subscriber powerhouse if 89 million people would drop Prime if the shipping went away…
Data of the Week – That Disney+ Survey and a Disney+ Check In
Lots of the Twitter world was very upset with a Cowen and Co survey-turned-analysis that predicted taht Disney+ has a likely 24 million subscribers globally. I have my own concerns about surveys too. First, asking consumers about behaviors is rarely as good as seeing what they actually do. Second, many surveys generate statistics that are quoted out of context, which definitely happened here. Third, surveys are hard to do right! It’s why FiveThirtyEight and Nate Silver spend so much time getting it right.
I’m willing to make a slight exception to the Disney+ subscribers question, though. First, when it comes to behavioral issues, this question if fairly straight-forward: do you have a subscription to Disney+? Most consumers know what they pay for. Second, unlike the worst surveys that are one off stories or trying to get obscure results, this question is being asked fairly regularly by investment analysts. So if you take what we know/have had surveyed–including some estimates based on surveys and Apptopia download data–I get this chart, sorted by time:
(By the way, if I missed any sub estimates based on data, send them my way.)
That’s not enough surveys to do a FiveThirtyEight-esque analysis, but I’d say there is a trendline here. Throw on top that Disney+ was the most searched for term by Google in 2019, and this launch is likely well above even Disney’s estimates. Factoring in that December/January are the largest months for sign-ups, and I think Disney+ could top 30 million global subs by the next earnings report. (Remember, Disney+ is available in Canada, Australia, the Netherlands and New Zealand too.)
These surveys, though, have a tougher time answering the question on the other side of this coin: how many of these folks cancelled Netflix? We don’t know. So we still have something to wait for in January.
Other Contenders for Most Important Story
A Peacock Check-In
At first, we had a report that Peacock may be ad-supported only, but then that was contradicted. Either way, it looks like Comcast plans to spend quite a bit on their streaming service. The biggest worry, though is that NBCU is changing up leadership teams again.
At some point, live sports as a category will migrate to digital. But not right now, as the PGA extended their deals–at a 60% value increase–on traditional formats. Oh, and remember that even though the 60% number is a big jump, the deal price hadn’t increased for 11 years, so amortize it out.
This is an old story I’ve been holding on to. Still, as long as we’re talking smaller streaming channels, they should get a mention. The biggest point I’d look at is the price point. NBA TV is $7…which is the same price as Disney+? This is NBA TV, not NBA League Pass. One of the streamers is either vastly under charging or vastly over charging. Time will tell.