In the 21^st century, Wall Street went gaga for subscription revenue.

There are many good reasons for this. Software releases turn out to be a clean example of this, so I’m going to talk about that for simplicity, but it applies to a greater or lesser extent to many other business transactions as well.

Pay On Release vs. Subscription

Consider the differences:

Pay On Release	Subscription
Unpredictable costs; when will the new software come out? How important will it be to upgrade? How much will it cost, that is, it may be more or less than last time? How shall we upgrade? All at once, incrementally, etc?	Predictable costs; this month it will cost $X.
Unpredictable benefits relative to costs. Will the upgrade be amazing? Will it be a dud? Will it be a dud but we have to upgrade to work with everyone else? etc.	Relatively predictable benefits.
Lots of decisions to be made. Decisions cost money.	Relatively few decisions to be made.
Feedback loop between consumer and producer is weak; once paid, the producer has no direct reason to care about the consumer.	Constant need to keep consumers happy to keep the renewals rate up.
No “stickiness”; once I’ve purchase the product, I’m done. The next purchase all competition has an effectively equal footing again; I might as well do a “big bang” spend with one vendor as another.	Very sticky. For a business, even just the task of analyzing the options can be a source of stickiness, and many subscriptions build in even more, like hosting documents (“if we switch, how do we migrate our content”) or other harder-to-switch things.
Accounting is complicated because of the unpredictability of spend.	Accounting is relatively simple because it is predictable.
Generally get to “own” it forever.	Generally technically tied to the further existence of the subscription entity.

That isn’t a full analysis, of course, and a given subscription has more details a business must consider (that last cell, for instance), but overall, subscriptions provide significant business benefits for both the subscription consumer and producer.

I highlight business because some of the benefits are generally applicable, such as the incentives analysis, but many of them are very business-centric, such as the accounting point. A consumer household has immensely simpler accounting needs than a big business, so the fact that subscription services make projections more reliable is nearly irrelevant to them. You, the reader, may even think I’m making that up or something if you’ve never worked in a large business or never been exposed to the issues of accounting in such.

Substituted Approximations

Humans are humans, and it’s well known we need approximations to deal with the complexity of the world. But there is always a danger when you wrap those approximations around something, as the streaming world is now demonstrating.

Wall Street humans, being humans, have now decided that Subscriptions are GOOD. I spell GOOD that way to represent the fact that subscriptions are not good because of the detailed analysis I provided above; the detailed analysis may have been done in the past, but now it has been simplified into an atomic statement: Subscriptions are GOOD. And as investors invest, so the business world develops.

And lo, the media business has rushed to get in to the subscription world. Subscription music, subscription TV, subscription movies, heck, subscription movie tickets. Subscriptions GOOD.

First Principles Violated

But the subscription world is not having the anticipated effect on their revenues. Why? Look back at the subscription table above. Consider a standard media streaming service rather than a software service, and look at the streaming side:

• Predictable costs: Yes, but consumers don’t care. At ~$10/month, mostly what I’m worried about is the cumulative cost of having too many, or having them too long; per month is irrelevant to most budgets.

• Predictable benefits: Sort of. I’m in control as a consumer as to what subscription services I watch, but if I’ve got ten, which ones I will actually watch this month is unpredictable on a per-service basis.

• Few decisions to be made: Yes. Though there’s a risk of your customers being pissed you changed them for the last 15 months despite them never even logging in. But customers can generally accept some portion of the responsibility for that, and it’s not enough money to fuss about usually.

• Stickiness: NO.

  This one is changing in the public consciousness, but I expect it to be an ever-bigger problem for the streaming media industry. Four years ago I posted about rotating my subscriptions rather than leaving them on all the time. The only reason it even took me that long is that’s about the timeframe where everyone started disaggregating from Netflix and suddenly there’s a dozen services. It’s not a hard conclusion to come to, and I’m seeing more and more “normal people” talk about it.

  The result is that streaming services are increasingly less sticky. I don’t need a seven person team for three months to determine how to do this migration; I push a button to subscribe or unsubscribe effectively on a whim, or just to make a rhetorical point.

• Accounting: Consumers don’t care.

• Tied to the further existence of the entity: At first, consumers don’t notice this, but the more you’ve been with a streaming service the more you will start to notice things discontinuing, either removed from the service (common today) or the service collapsing and the content disappearing (common tomorrow?).

  Many consumers, like me, are leaning the other way; purchasing physical media is certainly an expense, but physical media is the “predictable” option. My family just watched Stargate SG-1 together. I purchased those DVDs before I had the young teens who just enjoyed them. I wouldn’t dream of predicting my streaming access to any media I don’t own 15 years from now.

What does all this mean? It means that subscriptions are not GOOD. They are contingently good. They are good, because of certain first principles. When you throw away the first principles and try to take your technical analysis into a domain they don’t apply in anymore, you may make big mistakes.

The Inevitable Consequences

And so, stories like this: Disney Flinches as 2.4 Million Subscribers Abandon Disney+.

I said ten years ago this wasn’t going to work:

As a consumer, I am not going to subscribe to a dozen different streaming video services, one per distributor who thinks they’re too good to be aggregated. I don’t care how good you think you are or how good you’re going to be, and frankly I don’t even care if you were each trying to charge me a dollar a month, because it’s just plain too much hassle to be subscribed to one service per distributor.

And I was nervous about this prediction for a bit, while customers did briefly seem OK with it, but more and more I’m hearing about streaming rotation from my peer group, rather than highly selected internet groups. This is disastrous for the streaming industry, because it effectively reduces the TAM (total addressable market) from (Number of Households * $10/month) to that value, multiplied by the ratio of the number of services to the number of services consumers will simultaneously subscribe to, a multiplicand whose value is certainly debatable but is distinctly smaller than 1. This annihilates business analyses of feasibility and profitability based on the TAM without that correction factor.

I understand the business analysis the individual companies made. I understand them staring at the Netflix revenue stream and salivating. I understand how easy it is living in your own world to imagine that the demand for Disney’s content or WB’s content or Paramount’s content was just so amazingly robust that we should totally rip it away from the generic aggregator and starting changing just as much as the aggregator does for a fraction of the content¹.

But in the end it was a variant of the Tragedy of the Commons, at least in terms of the game theory payoff matrix. Yes, HBO probably could go its own way, or if nobody else did, probably Disney could have gone its own way. But when everybody yanks their content and goes their own way, it turns out none of them quite independently have enough content to stand on their own.

Not even a behemoth like Disney.

Predictions For The Next Ten Years

• In the short term, businesses will conduct something like the analysis above, and identify as the “most actionable” the stickiness problem. Expect to start seeing “yearly discount subscriptions” pop up (the accountants in the companies can still book that revenue monthly, as subscription revenue), and expect to see some companies try to push hard on keeping you in through various dark patterns.
• In the longer term, expect to see some reconsolidation. Disney already has a variant of this with their Disney+/Hulu/ESPN subscription. I expect to see more. This effectively re-inflates the TAM by reducing the divisor in that ratio I discussed above.

But that latter has a lot of interesting caveats, and some ways the industry has sort of screwed up with their customer base. Consumers expect these streaming services to cost about $10/month. If one hypothetically tried to consolidate all of them into one service today and then charge $100/month for it, even if that is a “huge discount” relative to their prices today, consumers are going to balk at that.

The problem is that consumers are still getting only that $10/month of value out of your package. Jamming more content they don’t watch in a month doesn’t raise the value to me as a consumer much; only the content I watch matters to me, and to a lesser extent, content I want to or expect to watch. Content I have no personal interest in is worthless to me.

The rush into streaming and the greedy desire to have all the profits has created a serious problem in the vast money fields of the subscription streaming world. All I can really say is, stay tuned to watch what happens. It’s going to be an interesting lesson in business no matter what happens.