If IT never gets in your way on a purchase, that’s great. But, I’m seeing this more and more across all business units (not driven by IT) in companies making purchases that have lifetime value > $250k.

My simple point is that your religion should be around your deployment model, not your licensing or pricing model.

The subscription vs. perpetual analysis is sound, but missed out one key innovation of SaaS, which is the pay-as-you-use model.

As you point out, perpetual and fixed rate subscriptions are really two side of the same coin, although the ability to cancel a fixed price subscription makes it much more flexible.

At ProofHQ we have gone to market with a fixed rate month;y subscription model. We have different levels of pricing for different profiles of user based primarily on the number of documents that they send for "proofing" each month. All familiar stuff for users of Basecamp, Freshbooks, etc. Of course in a minority of cases we are asked for a more traditional perpetual license, even when we are hosting or managing the service and we don't turn those deals down.

However, what has surprised us has been the number of people who want to buy a "block" of proofs to consume as and when they want. To me this is the most interesting direction that the SaaS payment model is going in – "pay as you consume". I guess the mobile phone companies got there first, then Google with Adwords and Amazon with S3 etc.

They key to maintaining flexibility in pricing models is being able to define a product's "unit of value". In our case "proofs". If you get that right then you can apply it across a wide range of license models without too much conflict or inconsistency.

While the appliance is installed behind the firewall, as Brad points out, once you stop paying for 'maintenance', there goes support / updates.

SaaS subscription model, Behind the Firewall perpetual license.

The next one is harder but doable. I'll agree that you can go behind the firewall subscription. In fact, I think its better to start that way, but given you're in a perpetual model and you want to move to subscription ok, get ready to feel the pain from the finance types but doable.

I want to see how you deliver SaaS in a perpetual license. The definition of perpetual is continuing forever. So the organization can buy once refuse the 20% support and continue to use. Just like Windows XP. How does that work?

This stance has allowed us to deliver what our customers need in a manner that they require, not what we want to force on them. We have our ideal customers and our ideal deployments. We target them very aggressively, but should one come along that doesn't fit the mold perfectly, one that wants to buy ergonomic, left-handed knives, we'd be foolish not to have one or two available somewhere. Be easy to do business with.

To Phil Sugar's comments, I have an amazingly rich, well structured comp plan (and hiring too). I have no incentive to push a deal one way or the other, simply to do what is best for the customer. As a rep in a small start-up, it is also important to do what is best for the company, so I try to structure my deals in a happy medium.

Having all of the knives is a great position to be in. It allows me to deliver the right solution for my customers and allows them to pay for it on terms that work best for them. As a sales guy, it makes my life insanely easier as I don't need to jump through a bunch of hoops to make one off exceptions that ultimately become the rule.

Great post, Brad.

One other thing I learned for orgs that offer both is that it's extremely important to show the customer how the models are cash neutral after x years (x=3-5 depending upon your business). If you can't show this, some customers will spend an inordinate amount of time trying to figure out which is cheaper. They need to understand that this is all about being "easy to do business with", and that's it.

Comp models are solvable. Even big public companies have been doing this for years. I've worked for a few of them.

Re: paying commissions to get good experienced salespeople – I agree. However, I’ve never run into a great sales exec + great sales force that has struggled with multiple licensing approaches when the comp plan is clear – and I strongly believe you can create a clear comp plan to address multiple licensing issues.

2. I think Joel Spolsky is much more eloquent than me on unintended consequences of bonuses. You have to pay commissions to get good experienced salespeople, I think as soon as you add another order of complexity you can't make it work.

3. I think you take the tech view and they are both as difficult I agree with your SaaS and behind the firewall comment.

Regarding the public market stuff – I’ve completely given up on the validity of public market analysts. As an investor and entrepreneur, I just don’t care about them anymore. My goal is to build companies to maximize cash flow as that’s the ultimate driver of economic value.

2. Same as #1 – this is easy to deal with if you have strong sales management that understands incentives and how to structure them. You should be able to comp the sales organization in a way that that they are indifferent to licensing. If you want to incent a type of licensing, you can always lean the comp toward that one, but I think that undermines the idea of giving the sales people all the knives.

3. This is an assertion that obviously is different than mine. I think you have to build your deployment model one way or the other (it’s extremely difficult to have both an SaaS and a behind the firewall deployment model in the same company). But – I think you can mix licensing models easily.