Much has been written lately about pay-for-use hardware and software licensing – including the many benefits and drawbacks compared to the traditional, perpetual approach. Pay-for-use licensing comes in many flavors—SaaS, IaaS, PaaS, cloud computing, utility computing, etc. In general, these license models allow customers to rent all or some of the software and/or hardware for which they would otherwise have to purchase perpetual or ownership rights.
One of the primary characteristics that differentiate the various types of software/hardware rental is time granularity – the duration of the rental period. Other time granularities include: multi-year, annual, quarterly, monthly, weekly and so on. Subscription and term are two popular rental models. Subscriptions generally provide for ongoing rental until cancelled with pricing based on specific configurations or products rented during one or more time intervals each. Term licensing is similar to Subscription but allows rental over a predefined period of time, not ongoing. I use the term "pay-for-use" to describe ongoing software and hardware rental types, in this case, all but Term.
Pay-for-use, in its most valuable form, is fairly dynamic – within reasonable limits, of course. Suppliers can use it to satisfy customer demand whenever and wherever it occurs, gaining market share and increasing revenue. Customers can use it to improve asset flexibility, access and efficiency. So why aren't more suppliers offering pay-for-use and why aren't customers adopting this model when it is offered? The lack of predictability seems to be one of the most often cited reasons. Ironically, pay-for-use is a very important model for suppliers and customers precisely because it addresses unpredictable asset demand. Let's face it, it's a dynamic world. Customers need the ability to quickly increase software and/or hardware resources to meet peak demands – without having to over-purchase for peak needs or risk onerous audit liabilities. Suppliers would much prefer allowing customers to rent assets during these peak periods vs. having to loan them free of charge or strain relationships with audits. The loan approach is much less feasible due to Sarbanes-Oxley financial controls.
So how do we crack this predictability "nut". Actually, with an adequate infrastructure and a supportive pricing model, it can be fairly straightforward. In fact, the common cell phone pricing model provides a useful basis. The cell phone model usually includes a committed "bucket-of-minutes" plan coupled with the customers' ability to exceed his/her committed volume for a peak per-minute fee. Now, there are two major changes needed to make the cell phone model feasible for business customers. First, the peak per-unit fee has to be much less than 5-10x the committed per-unit fee – 1.25x to 2x is probably more reasonable for the same time granularity. The price difference should be high enough that it fairly compensates the supplier while not being so low that it invites peak usage as the norm. Also, as with cell phone plans, peak usage is billed in arrears and customers should be provided the ability to increase their committed volume under suitable conditions. Second, business customers must be able to proactively control their peak use expenses, not experience the surprise of receiving a "budget-busting" bill for usage run amok, after the fact. This can be accomplished by providing customers the ability to control how much, when and to which employees the peak assets are allocated – preferably without having to contact the supplier beforehand. Supplier predictability is optimized because the relative committed vs. peak pricing difference minimizes peak usage. Customer predictability is no longer an issue because they control peak quantities and licensing durations so they can therefore fix maximum incremental cost exposure.
Great theory? Not at all. Flexera can help suppliers and customers realize these capabilities today. I've personally had the pleasure of working with two innovative software suppliers that have been providing these capabilities to customers using Flexera products since 1996. What is your experience with pay-per-use licensing models? Have you cracked the predictability "nut"?