I’ve been involved for over a decade in various shapes and guises selling software as a service to enterprise. But had yet to find an innovation in pricing models that truly reflected the value/volume conundrum that so often stifled adoption of solutions that only truly demonstrated their value at mass enterprise usage.
And then along came this post from Julien Le Nestour – pure brilliance. As he points out:
Every product bearing what is usually dubbed a “social component” has significant network effect and peer production dynamics. The more employees actively use the application, the more they — and so their organization — extract value out of its use. Marginal benefit per user, and hence total value, thus increases with the number of active users. Yet, most pricing structure are degressive, Volume-Discount schemes: price per user decreases with the number of users. Price and value varies in opposite ways.
Take more upfront risk for deeper customer deployment
Using Yammer as a case study he argues:
The more users will use Yammer, the more value the client organization will get out of it. In most organizations however, the value of a Twitter-like for corporate use will not be obvious, and will slowly build up with time, as it spreads internally.
Yet, pricing is desperately of a Volume-Discount type, making an after-pilot deployment with a small group of early-adopters look very expensive per user (large companies will compare it to the price per user for fully deployed applications like email or IM). Smart vendors will reverse the price structure, offer organizations the opportunity to try out the new technology, experience its value over time after a pilot, and scale up accordingly. They have to forgo immediate but short-term benefits, in order to get a chance to demonstrate their value added and reap the benefits as the client scales up its use.
I suspect there is a good deal of tweaking that will need to be done to make this model work in practice so as not to totally shift the burden of risk to the technology supplier, but inverting the model is a good start.