The ability to accept online payments has been transformed in recent years. Payment providers such as PayPal, Stripe, Zuora have made it relatively simple to accept online payments for one-off items or re-occurring services. The license model applied to a re-occurring service is, of course, a subscription. Whether it's a monthly subscription to a music streaming service like Spotify, or a monthly subscription to digital news content provided by a company like the Economist, the subscription model can work well. Among content providers, it is popular because it is simple to understand for the end customer, relatively simple to price and, if priced correctly, can also be lucrative for the content provider (Note that here I am using 'content provider' to be any provider of digital content from applications to video, to music to text, etc.). The subscription model is also popular because it is the main license model supported by payment providers such as Stripe and Zuora. They make it easy for content providers to deploy the model, offer discounts, track usage, process payments on a global basis, reconcile local taxes (in some cases) and easily integrate the payment engine into their own websites.
But not all types of content products are suitable for being shoehorned into a subscription model. This is particularly the case in B2B models. Take a simulation application for example. It doesn't really matter what is being simulated, but such an application is likely to be relatively sophisticated in composition, requires a decent amount of processing power to run and is likely to be more, rather than less expensive to purchase. It may also not be used that frequently. In this case, a subscription model is not suitable because the end customer will be paying for something they are not actually using on a regular basis. In contrast, an 'aggregate use time' license model may be more appropriate as it would charge the customer based on a pre-booked amount of usage. Similarly, a 'consumption time' model may be appropriate if the end customer would prefer to pay on an ad hoc basis.
Customers want and expect flexibility in paying for online services, particularly in B2B cases.
Another example of where a subscription model is not suitable is where an online application/service is accessed by multiple people but each requires it on an infrequent basis. One subscription per person would likely cost the end customer more than they would likely be happy to pay, so being able to offer them a floating pool of licenses, which can be drawn down by different people on an 'as needed' basis would offer much greater flexibility than a simple subscription. A floating model, tied to an aggregate use time model, could offer the best of both worlds, by offering multiple employees access to the online service but at the same time also 'capping' the total amount of time the application can be used for, thereby limiting the cost.
It is worth keeping in mind that customers want and expect flexibility in paying for online services, particularly in B2B cases. The success of services such as Amazon's AWS for example, has fundamentally shifted the B2B buyer's expectation to one where he or she will likely prefer to pay only for what is used - whether measured by time, storage, CPU cycles, feature set or similar.
The license models listed above need not be difficult to deploy. The 10Duke Entitlements engine offers simple web-based configuration to these and many more licensing models, suitable for both B2C and B2B businesses. Rather than trying to shoehorn your online service into a subscription model, you can still leverage the processing capabilities of online payment providers such as PayPal, Stripe, Recurly, Zuroa, Braintree, FastSpring, and Ayden but apply a license model or models to your products which is tailored to them specifically and more appropriately suits the nature of the underlying application or service you are offering.