In his excellent article 10 Contracts for your next Agile Software Project, Peter Stevens described 10 different contracts that are often used in software development projects. At the end of his article he singles out one contract-type:
“Fixed price projects and agile development are considered not to mix. I know from experience that this is not strictly true. I am also convinced that other forms are better.”
I completely agree with him on this but since fixed price contracts are so common in software development it seems like a good idea to look at them a bit more carefully. In this article I want to look the problems with fixed price and why traditional software engineering methodologies are unable to solve these problems. In my next article I want to continue on a more positive note and show how agile practices will help you make the most of fixed price.
What’s so bad?
The most obvious downside is of course lack of flexibility. Fixing project scope in the contract often makes it non negotiable during the project often preventing change altogether. In my experience stakeholders will change their minds during any non-trivial project and fixing scope at the start of a project will limit the quality of the end-product because it will not fit the client’s wishes one-hundred percent.
Not only scope is fixed at the start of a project but price is also fixed, hence the name. Making the supplier to commit to price estimates even before the project has started forcing critical decisions at the worst time possible. The later you make a decision the more information you have to base that decision on. Agile tells us to make them at “the last responsible moment”. Making them sooner is risky.
Fixed price projects are often more expensive than they should be. The rigidity of these projects often causes scope to be too large. At the start stakeholders in the client-organization will specify many requirements and features they probably won’t need ‘just in case’. They know it will be hard to get them in later. Fixed price projects also force the supplier to make very accurate estimates at the start of a project to reduce risk, the time spent creating these estimates are later often added to the price of the project. Suppliers often also add a fee for the extra risk they have to carry, a former employer of mine added 20% to fixed price contracts and seeing how much trouble fixed price was for them this was actually low.
Instead of aligning interests of the client and the supplier fixed price projects artificially opposes them. The client will want to stretch the scope agreed in the contract as wide as possible while the supplier will want to do as little work as possible. This often prevents good communication while increasing risk. I have yet to see a fixed price project where the client delivered an on-site-customer and I don’t think this is a coincidence. Fixed price pushes stakeholders into roles that are not beneficial to the success of a project.
Fixed price and BDUF
I’ve heard more than once that these problems make fixed price problems especially to do in an agile way. A potential employer once told me when I asked him if they used any agile practices “We do fixed price so we can’t do agile”. When you look at the downsides of fixed price projects this seems logical. Fixed price has problems with some of the core agile values like flexibility and communication. The focus on up-front estimation also seems to point to big design up front or BDUF.
Unfortunately more traditional approaches won’t help us either. Big design up front will not help with estimations. We need accurate estimations before the project has even started. There hasn’t been time to gather anything resembling a complete set of requirements. Traditional estimation techniques like function point analysis are only as good as the data you put into them and unfortunately the data isn’t very good at the start of a project. Problems with flexibility and communication are only hidden, not solved.
Fixed price has some big problems. Not only for the supplier as many clients seem to believe but also for clients. These problems don’t go away by giving in to the traditional big design up front methodologies although some of the problems with flexibility and communication will be less visible. In the next article I want to look at how you’d do fixed price in a more agile way. This won’t solve the problems but it will give some tools to see them coming and to make it easier to switch contracts during a project.