Project Estimation: Two Conflicting Trends

In a number of organizations I have noticed two conflicting trends:

• Organizations want business cases produced earlier (to prevent wastage of funds)
• Organizations want increased estimation accuracy – typically ± 10%.

While the objectives are totally understandable, they are unrealistic.

If you insist on a business case too early in the project life-cycle you incur a high level of uncertainty — there are many things that you don’t know or don’t know well enough to accurately estimate. Often business cases are generated based on “high level requirements” – but, as we all know, the devil is in the detail and when the ‘detailed requirements’ are later defined they will often blow the budget.

We need to remember what we’re trying to achieve – ie not waste money, not allocate money to poor projects (however you want to define that) and not have big cost blow-outs. Bringing the business case forward and insisting on greater accuracy will not achieve this.
Instead, what you need to do is…

1. Recognize that there are ’sunk’ funds, ‘at risk’ funds and ‘total’ funds.
   • Sunk funds are the expenditure to date – this money has gone whatever happens in the future. It may be painful, but it is not recoverable.
   • At risk funds are the expenditure that you’ll lose if you invest in the next stage and then cancel the project. While you want to minimize ‘at risk’ funds, you also want to invest enough to see if the project will be valuable and viable on a reliable basis.
     
   • Total funds are the total expenditure to the end of the project and are only at risk when they’re allocated and then spent. However great the total funds are, they are not at risk until you put them at risk. Better to spend, say, $5m determining the true value and cost of a project than, say, $500K to generate a baseless business case that later explodes in cost or implodes in value.
2. Recognize that the level of uncertainty is high until the requirements are known and the design agreed. Up until then the level of uncertainty is certainly above 10%.
   The progressive questions a project needs to answer (as cheaply as practical) are:
   
   1. Is the idea/concept relevant?
   2. Is the project likely to be valuable and viable?
   3. Can we afford it?
   4. Can we deliver it?
   5. Do we have the capacity to deliver it?
   6. Is this the best use of our resources now?
   As soon as any question is failed the project should be stopped and any remaining funding reallocated to other initiatives. The early evaluation process should, therefore, be focused on answering these six questions at minimum cost rather than insisting on early but unrealistic certainty.
   
3. Ruthlessly review projects/ideas/initiatives to see how they are faring against the six questions – and stopping any that fail as quickly as possible.
   Along the way there will still be areas where you won’t know the answer yet – and this then needs to become the focus of the next stage of the project.

By the time you get to the final business case stage you should have no unknowns. Then you can have an estimate ± 10% and rely on it.
Accuracy demands in the face of reality is not a useful approach.

Article by Jed Simms for PM Hut.