Streaming Revenue?

Uh so there’s this article. It’s ludicrously wrong yet kind of close.

Based on this article we know the Minimum Guarantee for a series is about 1000-2000 per episode streaming. But did a million people watch Sakura Trick on CR? To use rating as a proxy for video views†, There are 61 series rating. Episode 1 had about 750 ratings. In comparison Naruto Shippuden has 48000+ series ratings. So let’s say 0.1% of CR’s viewers watched Sakura Trick. We know CR has 20M users, so that is roughly 20000 viewers. If that GB article is right, then at 7 cents per episode Sakura Trick would have made just $1400 USD per episode. Sakura Trick is probably not a 2000-per-ep series, but it’s easy to see that CR probably is not making much money from this show if it’s gonna cost approximately the same as the MG.

In that GB article, the math to get to that point is protracted, and the logic is just wrong. It uses world-wide counts on US prices, it uses guestimate from Netflix’s 2013 PR, quoted from a Quora article you can now google, which are based on even more questionable numbers. Let’s forget the links I left up there as they’re not referenced, which is probably way more on target because we now respect the geography distinction (the Netflix number quoted by the GB article is for US/CAN really anyway).

The ballpark estimate, however, is done in a way that respect the rules of ballpark estimates. Like one of those order of magnitude interview questions (eg., “how many grains of sand is in a playground sandbox?”), you can’t know the specifics but you can make a good guess, by certain definitions of good. The problem in this case is that if you guess this way you’re bound to be wrong, when the wrong guess can actually skew people’s perception in a business of fairly tight margins.

Because, if you really need to understand the displacement aspect of illegal streaming, it’s worth asking the question:

1. Are any of the illegal streams based on views from regions not served by legal alternatives? In this case probably a solid yes for Sakura Trick.
2. Is there displacement if, say, a CR subscriber uses an illegal stream to watch something available on CR? The streaming version of the “download but buy” use case. It might be safe to say this isn’t likely the case.
3. For ad-driven revenue models, the revenue comes from ads served. Wouldn’t it make more sense to calculate displacement based on that (and not what Netflix pays to license stuff)?
4. For subscription-driven models, how can we calculate displacement?

We can probably ignore the marginal cost of the streaming service, although there is some kind of price tag attached to this in a way that satisfies demand (I’m looking at you, Funimation, will FunimationNow not suck so much…probably).

So like, if you estimate displacement based on #4 and, say, the average CR viewer (as model) watches 20 episodes a month, at $7. So that comes down to 35 cents an episode. Or using the same PR, 1.5B minutes per month is roughly 62.5 million 24-minute episodes, assuming sub and free users with the same viewing habits on average, we’re talking about 3.125 episodes a month. You can ratchet up a huge value for displacement with this math, up to $2.24 USD.

So yeah, that article manages to get the price to roughly where it really is, but doesn’t really do anything to calculate the cost of streaming piracy. So let’s do #3 real quick, using Daisuki as an example.

I don’t know what Daisuki uses for their video ads, but let’s say it’s something like Google and they take a roughly 50% cut. Then if the CPM is like $5, then Daisuki makes maybe $2.5 on a per X number of videos. I think they work with some small platform because the ads on there are location specific and not exactly very varied, so their rates may be good. If we use Google’s rates, we’re talking about anywhere between $1 to $20 per 1000 views, or let’s just say $10/1000, 3 times per ep, or 3 cents per episode. Which is $30000 per episode per million views of displacement.† Of course CPM is also region-sensitive so who knows?

I think that’s a much more reasonable number if you ask me. I mean LOL one million views of one episode of Sakura Trick LOL. Sure you can apply this to One Piece and get another zero attached, maybe two, but that’s getting into “reap where ye have not sowed” territory.

†Speaking of territory, because of the region nature, the ratio-to-Shippuden based on series rating is a safe bet to fold into a same-service comparison. Meaning that Sakura Trick on CR is for US/CAN only but that fact would fold into the small number of viewers as a result of more or less strict regional limitations. The same cannot be said of the CPM calculation, so it’s actually much harder to estimate displacement unless we get the country breakdown of all the pirate streams. All this is just to say calculating displacement based on a subscription model makes no sense, yet this is where everyone is going with streaming anyway.

PS. The most relevant way to look at piracy in terms of streaming is to look at the break even pageview. Meaning that if the MG is 1000 per ep, and you have a million total views for a 12-ep series, you’re paying $12000 and getting, say, $30000. And yeah if almost 100,000 people watched Sakura Trick that is a no brainer. Compared to actual TV ratings, for example, that’s quite a lot of people (like 0.1 of a rating)! But let’s say if 3333 people watched Sakura Trick all the way through, or one-thirtieth of people or so, that’s like where the break-even is. Which is not an unrealistic number for something streaming only in the USA, and obscure. Translated to DVD/BD sales, if something sold a 1000 units I think that thing is in good shape, so if 1 in 3 people watched the simulcast or whatever… something to think about, eh? And as usual region breakdowns will screw with anything you can think of concretely but if we think of the distribution globally as one single basket, and if the ad market can support that (big if), there is some money to be made.

Which is just to say that I keep on snickering on that 1M number because that’s ludicrous. We should do a pool and guess how many of that 1M is from Japan!