One key to software success in DoD: Consumption-Based Solutions

Here’s a lengthy but important discussion on Consumption Based Solutions from Nick Tsiopanas at a nice NPS ARP webinar, Innovations in Software Acquisition at the Department of Defense.

The world has fundamentally changed in terms of technology since this law was written. What I mean by that is at the time the Clinger-Cohen Act was written (1996), IT was considered a capital investment and that’s why you have all that capital planning and investment control stuff in the Clinger-Cohen Act.

 

Now IT can be bought as a utility. IT no longer has to be a capital investment. If you think about it, companies are not out there building their own data centers now because they know that they can buy that capability as a utility and an on-demand basis. So why would they spend all of that capital investment?

 

Now a related advantage here in the government space is when you’re doing a capital investment you have a certain type of appropriation. You have to have R&D or you have to have procurement money. But if you’re doing it as a consumption based solution, then you can use operations and maintenance money which is a lot easier to come by — especially in a budget-constrained environment. So that’s a significant advantage of buying IT as a consumption-based solution as opposed to as a capital investment.

 

Now, I imagine that a lot of you are seeing more and more of this type of delivery model not just at work but in your personal lives as well. I think it’s fair to say that for the foreseeable future, this is the way that solutions are going to be bought and sold in the marketplace. So the takeaway is that the government needs to get on board with this and embrace it and not have it be another example where we create a bunch of unique rules and constraints and thereby surrender the advantages of buying a commercial solution to begin with. We just can’t keep doing that.

 

… The far categorizes all purchases as either supplies or services. Let’s talk about software — software as a service. Everyone’s probably heard of that. Well looking back in the not so distant past, software was pretty clearly a supply. It was something that you bought on a disk you installed it on your computer and then you had that software it was a one-time thing. You could maybe buy an upgrade later, but it was very clearly a supply and now with software delivered over the web in the software as a service model, we start to ask ourselves, ‘is software a supply or is it a service?’

 

The word ‘service’ is right there in ‘software-as-a-service,’ so maybe it’s a service. But what’s the problem with that? When we talk about services in the context of government acquisition, we’re really talking about all the rules around labor rates and how many people are performing the service and so forth. The reality is with software as a service, there’s no possible way for us to know how many people are behind the curtain delivering that software as a service solution to the end user. [emphasis added] They may have thousands of customers and their individual employees are supporting many customers in many different ways. So you really can’t parse it out.

 

The idea of service contracting rules in terms of consumption-based solutions is really quite a stumbling block now when we were working on this recommendation in the [Section 809] panel. There were some folks that had some ideas about existing special authorities in the FAR that we might be able to use rather than creating a new one to support acquisition of consumption based solutions.

 

One of those ideas was subscriptions. Could we use the special authority for subscriptions? It was an interesting idea we explored but as we look deeper into it what we concluded was it’s actually the opposite of what you want with a consumption-based solution and here’s the reason. Subscriptions — that special authority — is about paying for something in advance, before you receive it. Consumption-based solutions we can’t possibly pay for in advance — at least we shouldn’t because we don’t know how much to pay until we’ve actually consumed the service. Once we’ve consumed the service then we know how much to pay. So subscriptions pay in advance, consumption-based solutions pay in arrears. [Emphasis added]

Here’s some additional information on the timeline of events for consumption based solutions:

• Jan 2019 – Section 809 Panel recommends it in Vol 3.
• Oct 2019 – FY2020 NDAA asks for a feasibility study.
• Nov 2020 – DoD reports not-so-favorably on it
• Jan 2021 – FY2021 NDAA Sec 834 directs a pilot program.

Here’s a few considerations around consumption based solutions: (1) it increases private investment into product development and decreases government RDT&E funds; (2) it challenges usual methods for evaluating “price reasonableness”; (3) it affords DoD optionality from easier on/off ramps of suppliers; and (4) SaaS may not be a viable strategy for all applications, what are the rules of thumb?