QUESTION

Dear Mr Risk,

I have populated my schedule with optimistic and pessimistic durations that sit around the durations assigned in our proposed baseline. We have run our analysis package and produced a very nice ‘normal curve’ with various percentages showing at the 5% and 95 percentile and various completion dates.

Should we be doing something with the costs assigned to the schedule as it seems sensible that cost, or budget, would be proportional to the various resultant schedules; the short ones, the on-time one and the very delayed ones?

I really wanted to understand what we should do with this information other than advise management we have a 20% chance of completing on time and a much larger chance of being late or very late.

Is there a process or protocol that we should follow to try to improve the schedule before we submit our tender?

Thanks in advance and please keep me anonymous as this is relating to a major tender bid in central London.

Thank you.

Regards…

SUGGESTED ANSWER

Dear Sir,

Thanks for the question. Actually there are two questions. First, you ask about cost. In my opinion it is less important to do risk analysis on cost than on schedule, because costs are merely additive whereas task durations interact in complex ways in a typical project network. Having said that, most risk analysis systems treat costs as well as schedule, so if your costs are loaded into the schedule (as opposed to being held in a completely separate system) the results should include distributions of cost as well as schedule.

Your second question is more problematic, namely what do you do with the results? This is a little bit out of my area of competence, which is in producing the results rather than acting on them, but I will give it a try. Knowing that you have only a 20% chance of meeting the required date is certainly useful information, and should for example inform the negotiation of any penalty clause. It may even mean that you should not bid. (I am assuming here that the date is dictated by the client.)

More usefully, you should try to re-arrange the project to increase the chance of meeting the date. Perhaps there are tasks which can be done more quickly with additional resources, or perhaps tasks which can be done in parallel. Perhaps you can reduce the uncertainty by acquiring additional expertise, for example by subcontracting. In short, you should consider all the things you would when trying to shorten a deterministic CPM. Resist the temptation to change your assumptions about the duration distributions unless you have really good reasons to do so, because otherwise it is probably just wishful thinking.

If none of that produces a schedule with a better chance of success, ask yourself this: are the reasons for this specific to my company or are they likely to be common to all bidders? If they are specific to your company, then it might be best not to bid. If the latter, maybe you should submit a “non-compliant” bid with an honest estimate of the completion date.

I hope this helps.