The Subscription Model is the Future of Software—and for good reason

Microsoft recently announced the launch of Office for iPad. Many people took note about the pricing model behind Office for iPad: the apps are free to download, and the user can open and view documents, spreadsheets, and presentations for free, but must pay for an Office 365 subscription to edit these files. Many have taken notice.

At Microsoft, the move to a subscription-based model has been going on for some time. Office 365 was launched in 2011. With Office 2013, the move towards the subscription model has been consummated: a user with an Office 365 subscription can use click-to-run to stream any Office app on any machine, and has access to the latest bits of the software at all times.

Many in the tech community and elsewhere have met this move with caution or concern. At the heart of these concerns are two ideas:

  1. a subscription model to use the software, instead of being able to purchase it, means that we never really own the software we pay for, and
  2. a subscription model also seems like a way to bleed money out users, especially infrequent users.

These are legitimate concerns, but I will argue that out understanding of software ownership is outdated and harmful. Further, the financial concerns about software subscriptions can be mitigated.

It seems that it is fair to—and often, acceptable—to charge for a service on a continuous basis as it is being provided, but neither fair nor acceptable to charge a continuous fee for the use of goods, where one is expected to “own” such good. The renting of goods will come up as an obvious counterexample, but I suppose renting is an exception rather than the rule.

Software is seen as a good, and to pay a subscription for software exactly as an attempt to rent it out. Many will object to the business model of “renting out” software licenses because the use of software by one user does not exclude others. It therefore seems that renting out software is a greedy attempt at amassing profit, rather than a legitimate business practice.

But in this day and age, the use of software carries many hidden costs. A company must offer supports to users of software, including users who own older versions, or interoperate with other users who own older versions. A user could be complaining about a bug fixed several years ago, but if the user is having this problem, as a company, it is also your problem. A company could also choose to have strict terms of service and choose not to support some of these areas, but here we see another cost of having your old software out in the wild: its reputation.

Here, I hope an example from Microsoft software will be illuminating.

Case: Windows XP

Windows XP was released in 2001 and still commands a usage share of some 30% of traditional PCs today. Support for Windows XP has been extended multiple times since it was first slated to expire, but in April 2014, support for Windows XP will finally expire. This is good news for everyone in the tech world: a 13-year old operating system has no place in 2014.

30% of traditional PC users are currently using an operating system that is more susceptible to attack. Some team of programmers in Microsoft is writing software fixing bugs and closing holes in a piece of software written in 2001, when many of these bugs have been made irrelevant by new security models, access controls, and sandboxing techniques that their colleagues have developed years ago for Windows Vista, 7, and 8. Developers throughout the world need to make sure that their websites are compatible with IE6, 7, and 8, even though IE9 has been released in 2011 and support open web standards more properly than previous versions.

Windows XP costs everyone today. Microsoft especially sees issues: the reputation of IE has declined, many developers no longer support or attempt to use modern versions of IE because they have a negative impression of the browser. Customers compare a Windows XP netbook circa 2004 with a 2011 iPad and decide not to bother with Windows.

There is one “cost” that I have not mentioned yet, because I do not think it is a legitimate part of this debate. That “cost” is the cost of modern Windows licenses that could be sold to these PC owners. It is true that the biggest competition for a new version of Windows is the previous version of Windows. I will not concern myself with this idea as a justification for subscriptions; as users, it is not our job to accommodate our lifestyles to optimize a company’s profitability.

Therefore, I am building an argument that is quite straightforward: the continuous use of Windows XP costs Microsoft in real, meaningful ways, not only in opportunity costs of missed sales.

Many developers are horrified at the idea of anyone using Windows XP. Developers of course don’t care if those Windows XP users update their operating system to a newer version of Windows or a competing product.

These users, however, have purchased and therefore ‘own‘ the operating system. Imagine a washing machine maker going to your home and demand that you buy a new one: “It doesn’t need to be from us,” a representative would tell you, “you can buy any new washing machine, as long as it’s modern. This one is much too old and we can’t have you keep it.”

Enterprise users have more of a reason to continue using XP. The 30% usage figure, however, contains many home users of the operating system.

If we cling on to our conception of software ownership, then we really can’t demand anything out of these users. Maybe we should.

The Solution

The real solution to this problem is to rid ourselves of this conception of software ownership. It seems arbitrary at best to distinguish between software downloaded and running locally (“a good”) and software hosted on the web (“a service”), if the continuous use of both of them carries a similar cost.

A subscription means that you have access to the latest bits of the software you intend to use at all times. Third-party developers can count on the majority of their users using newer versions. The software maker will see a reduced need to support legacy formats, instead moving closer to (n) ↔ (n-1) compatibility guarantees. A subscription therefore means that development is more tractable, that users witness faster rates of improvement, and that software is generally more reliable.

Case: SimCity

Subscription-based models will also help speed up the advent of cloud-assisted and cloud-enabled software. When a user purchases a game or program, the user often feels entitled to the right of using this software offline. Developers who wish to be on the cutting edge are offering online-only options that are often being met with anger.

EA asked users to purchase SimCity, a game that required a continuous internet connection to be played. Users objected on for the following bases:

  1. the always-online requirement seems to be arbitrary and offer no real advantages that cannot be implemented in an offline-only version, and
  2. locking users into EA servers means that legitimate owners of the game will be left in the dark if EA decides to top supporting the game or close their servers.

I will not argue against the first point, other than to point out that other games could exist which legitimately take advantage of the cloud to offer superior online-only experiences. The real issue, however, is that users did not want to pay to own a game that they will not be able to use at some point.

The Big Picture: “Free” Software and Piracy

I am using “free” in quotation marks to refer to software that is “free (as in beer)” as opposed to “free (as in freedom)”.

My opinion on the EA and SimCity debacle is simple: EA should have offered SimCity for ‘free’, not purchase, and instead charged users a subscriptions to play the game. If EA ever shuts down its servers (as it often does), these subscriptions are terminated. The users got what they paid for (the use of the game for a given period) and we move on to the next innovation.

I will be the first to admit that this assessment is incomplete. It would be much preferable if EA opened up SimCity to allow other servers to spur up (at least if they choose to shut down the official servers). Whether EA owes this to the users, or whether it is simply a better alternative will be left as an exercise to the reader, and is definitely the next part of this debate on “free” software openness.

We are now looking at the big picture: in today’s world, buying, selling, and owning software are outdated concepts. Privacy advocates have long argued that the marginal cost of copying software is practically 0, and only potentially impacts software makers in terms of opportunity cost. Software is information and should be copied freely, they argue.

The subscription model ends the privacy debate. Companies should not be charging for copies of software, but the service of use and support.

What piracy advocates have noticed years ago is that software makers are charging for software illegitimately at the wrong stage of the development-distribution-use process. The subscription model admits this mistake and recognizes that the true cost of software—other than its development—is proportional not to the number of users using it, but instead the amount of user-hours the software is being used.

Does a future of subscription-based software mean it is more expensive for the user? I don’t think so. Companies should instead be expected to offer flexible options, such as day or week passes that should accommodate very infrequent users. We should demand that software makers provide affordable options: the cost of a one-year license to use a product should be significantly lower than the cost software makers currently charge to purchase a license indefinitely.

Simply put, charging users for a copy of an application is not the way forward for software. All (or perhaps most) software is a service that involves patches, updates, support, and more. A subscription-based model puts an end to many current-day arguments in favor of software piracy, allows software makers to innovate at a higher rate, and gives users access to the latest bits at all times.