The first question to agree on is whether to have any contingency in a project estimate. Some organisations refuse to countenance contingency and just want an ‘accurate estimate’ on the project cost and timeline. This really is a silly notion as you cannot possibly estimate correctly at the start of any sizeable project what the final delivery date and cost is going to be. In an agile project you may ‘timebox’ the project but what exactly is delivered may change from the view at the outset of the project.My strong view is that contingency should be applied to estimates throughout the project lifecycle. At the beginning of the project the contingency may be relatively large, say +or – 50%, but as the project progresses, this contingency should be reduced as there are less unknowns later on in the project and you are closer to delivery. I have seen this contingency model work very well at some of the clients where we have worked. In some clients where contingency is ‘not allowed’ it can be the Finance department that makes this stipulation. If this is the case then Finance need to be educated as to why contingency is a valid concept.
If we agree that contingency should be applied to project estimates (time and money), the next question is how should it be applied?
One client we worked at wanted to hide contingency in all project estimates. For example, if the project manager estimates that a project will cost £100k then he or she produces a revised estimate of £120k. The theory here is that the project team does not go back and ask Finance or the Sponsor for more money as the budget has been awarded. My issue with this is that, human nature being what it is, it is very likely that most projects will use all or most of their contingency using this model. There is no check by anyone as to why the contingency is being used.
The second option is to openly declare a contingency of £20k and this is put in a separate bucket not accessible to the project manager unless he or she seeks approval from the Sponsor or Finance to use £X contingency budget for whatever reason. In this scenario there is a check and balance on the contingency. My assertion is that it is highly likely that the overall spend on projects in an organisation will be lower if contingency has to be approved before it is used.
Whichever option of handling contingency is used, it is patently not sensible to allow for no contingency in projects.