Pricing is notoriously hard, specifically when treading new ground with a SaaS or PaaS offering for which no established practices exist. We experienced this first-hand while developing the business model and pricing for Metrics: it took us multiple iterations and several months of elapsed time. In the process a pricing philosophy evolved that may be of more general interest, so I’m sharing it in this post.
I believe that SaaS pricing should be easy to understand, predictable, fair, flexible and consistent with the intended use of the service.
Easy to understand and predictable go hand in hand and translate into pricing that is defined in terms of service units consumed under direct control of the user. For example I’d like my mobile service provider to charge me for talk time and even for the amount of time I spend watching videos, but not for the MB’s of data that somehow is transferred to and from my phone. Similarly I’d like my metrics service to charge me for the number of measurements I send and the number of measurements it retains for me, but not for the number of API calls I make.
Fairness should apply both to the pricing structure (pay-per-use vs. subscription, freemium pricing) and the amount that is actually charged for the service (including discounts).
Pay-per-use vs. subscription
In my belief fairness calls for a pay-per-use pricing structure. I don’t want to choose between a few different programs, each with rigid upper limits on the amount of service I can consume, as this forces me to buy more than I need and unfairly let’s some customers pay significantly more for consuming just a small amount of additional service than others. Examples of “unfair” subscription models are abundant, so I don’t think I need to give examples.
I use the term freemium pricing to denote pricing models that provide a certain amount of a service for free, but charge for use of the service beyond the free limit. For the same reasons subscription models are not fair, freemium pricing is not fair, as it gives the largest % discount to the smallest users. Take typical pricing for consumer cloud storage as an example: is it fair that I get 2GB of storage for free and have to pay $10/mo if I use 4GB, which is the same as what others pay for using 50GB? Why not give me a period of free use to try out the service and after that simply charge me $1/mo per 5GB, maybe with some discount for larger users?
Vendors sometimes argue that a fraction of freemium users will eventually become paying users and that they refer others to the service. Although it doesn’t make freemium pricing fair, this may be a valid business consideration for some consumer services. However our experience with Silverline’s freemium pricing leads me to believe that such benefit doesn’t exist or is insignificant for most business oriented services. We see two classes of Silverline users: those who always make sure they stay under our freemium limit, and those who see value in the service and are willing and able to pay. It appears from our analytics that freemium users rarely refer companies to our service that end up being paying customers.
It’s also fair that larger consumers of a service receive a discount. After all, they usually require less support, lower administrative overheads and lower customer acquisition costs per $ of revenue. I prefer discounts that are applied to the monthly charge over discounts that are embedded in the pricing model, since it’s more transparent and keeps the pricing model simple.
Fair (cost+) pricing
Even though it is difficult to verify, pricing should be fair to customers while allowing the provider to build and sustain a healthy business. For example, to establish Metrics prices we measured the resources required for ingesting and storing metrics and used this data to build a cost model for the service. We also estimated the support and operations staff required to maintain the service, and with these “Cost of Goods” estimates and assumptions about the gross margin we need to continue improving the platform and build a profitable business, arrived at our price/measurement. I believe that pricing that is established in this manner will inherently make adoption of the service financially more attractive for most companies than hosting a comparable service themselves.
Flexibility means that you should be able to use as much or little of a service as you want, which again argues for “pay-per-use” pricing with no minimum payments. It also means flexibility in payment schedules. The typical SaaS pricing model charges your credit card once a month, in advance in case you’re using a subscription service and in arrears if you’re using a pay-per-use service. While most organizations are fine with that, some prefer a yearly “in advance” payment schedule (at least that’s what some of us old enterprise software types think). That’s easy for services that use a subscription model, but not for services with a “fairer” pay-per-use model. We came up with a scheme that allows customers to open a prepayment account in which they can deposit as much as they like and whenever they like.
Price consistent with intended use
Finally, pricing models should promote the intended use of the service. For example, if the use of alerts on a metric is promoted as a means for early problem detection, charging per alert discourages the use of that feature and potentially prevents the user from getting the benefits intended by the designers of the service.
Summarizing, SaaS pricing requires some science and some art, but when it comes to enterprise SaaS should:
- be based on service units consumed at the user level
- use a strict pay-per-use model, with no minimum payments and no freemium
- include a free trial
- apply discounts to the service charges of volume users, as opposed to including these in the pricing schedule
- provide flexible payment schedules, that allow both post and pre-payment
- be fair to the customer and the provider
- be consistent with the intended use of the service