To add to Sinclair’s comment, one thing that you can (and to my opinion should) consider is that normally you can negotiate with the Telcos a hosting package where you should figure out what your average user cost is, this should be assuming average usage of your application. It will come a point in the lifecycle of the usage of a given customer that either because he has been a very long time client or because of heavy usage of the application where he is consuming considerable more resources than any average client. I would say that it is fair to charge that client an additional fee for the extra amount of resources that he consumes every month.

In essence your pricing would be something like $n a month per user for y number of resources and for every additional z number of resources used you would pay $m a month on top of the regular subscription fee. In a similar way you can be creative about this and say that you would price your application at $m a month (where $m = 10 x $n) for up to 10 users for x number of resources (where x = 10 x y), this helps you absorb costs a little better and you don’t have to bother the client with overage for each independent user. Further you don’t have to introduce the notion of resource utilization when exposing it to your customers to keep things simple but you would explain it as part of the contract that their usage is tied to the average number of resources consumed for the average client and that they should not have to incur additional expenses unless they go over the normal usage of the application.

The problem of not doing this is that as time goes on, depending on your acquisition rate of new customers the cost per client can increase when it should only increase for the certain number of customers that are using more resources than normal.

Hope this helps!
Abe

To be honest, I don’t think we’re bouond to any particular approach. The Internet, unlike the cell phone industry mentioned in the comments of Wainewright’s post, has created the expectation that you can get what you want, when you want, and how you want. That said, I think the industry will always be a pricing “grab bag” with options to satisfy everyone. PaaS will more than likely drive this since PaaS can consolidate pricing mechanics down at the platform layer, allowing SaaS providers to take advantage of the variety of pricing mechanics offered by the PaaS offering with no effort.

Essentially, PaaS will and should drive toward allowing for a variety of monetization schemes that SaaS providers can take advantage of. Prior to PaaS, it was the responsibility of each SaaS provider to deal with pricing intricacies, and since most SaaS providers are resource strapped, they would generally focus on a primary pricing model and wait until the “next version” before offering other options.

It seems to me that you’d want a marketing person and finance person that understand price buffering on the sale side and a finance person that can help trend cost changes. You would also need to realize that price stabilization is good, but reserve the right to change price at discrete times (such as yearly).

From a price buffering perspective, the person filling your marketing role would have to be skilled in being able to gauge what your market could bear, and pricing at a high enough price point to help buffer against cost fluctuations coming from the telcos. Here is a fine example of where not competing on price and being able to sell your product at a justifiable premium offers you margin protection.

Second, someone with trending and analytic experience can help project cost fluctuations and coordinate your price changes (if necessary) to your customer base.

If you can get individuals with telco experience, even better, but most reasonably smart individuals with the above skills should be able to help regardless. Hope that helps!

Phil Wainewright is talking some feedbacks on the poay per use approach on PaaS.

Here : http://blogs.zdnet.com/SAAS/?p=559

Conclusions from Coghead’s CEO are going in the same direction than the current discussion even if PaaS had until now more of a pay per use model.

So are we condemned to the packaged approach ?

Key roles are really dependent on your goals. Off the cuff, however, I’d say that any growing SaaS company should generally fill upper management positions dealing with SaaS/subscription based sales, a technologist role (CTO) with experience in hosted software, growth and risk management related to centralization.

Any details as to what your concerns are with the growth so we can give you better input?