Insurances use the concept of aggregation to reduce the relative risk and thus are able to charge us low premiums for all types of insurance (it is the pooling effect). Why don’t project plans and project managers take advantage of this simple and well-known concept?
8. "Project Insurance"
Life insurance, car insurance, in fact every insurance uses the aggregation principle to allow then to charge premiums much smaller than they would have to if an insurance was for a single event. If the average number of accidents is 1 in a 1000, then premiums can reflect that every individual has only 1 chance in 1000 to suffer and accident (of course insurances charge according to the likelihood of partial or complete loss, and they charge for the administration of the insurances they manage. We, the consumers of insurance, benefit because our cost of coverage is significantly lower – we suffer a small/(ish( insurance cost to avoid the catastrophic cost of a total loss.
Why don’t project plans and project managers take advantage of the same aggregation effect?
We estimate the time for every defined task and include a certain amount of safety buffer in each estimate. In Part II of this series I claim a reasonable resource must give himself a 80-90% chance of completing his tasks on time. Less could endanger his career and remuneration. In fact he and his colleagues are in a reliability competition (and of course in a promised due date competition). Which is in fact more important reliability or an early due date promise? Whatever the truth is experience tells us (if you measure this) that the majority of tasks are on time, a few are early and a few are late to very late.
It must be the late or very late tasks that cause the majority of projects to miss their due dates. (Budgets probably show a similar logic.)
We should expect the vast majority of actual project completions to be completed early or on time – if it is true that project resources and their time estimates are reasonable – 80 to 90 out of 100 tasks should finish on time. Reality seems to be the reverse. Why?
The explanation I have learned (From Eli Goldratt) goes something like the following (I am explaining in my own words, I hope he is OK with my description of reality.
Imagine an individual resource. He has several tasks that he must work on; tasks that are necessary for a number of different projects. He has estimated all his tasks to give himself 80-90% certainty of finishing them on time. In his reflections he has considered the weather, things others must complete before he can start, surprise requests from management to do some urgent thing for them, and he has considered that he must work on several things at the same time (for several project managers). He is a reasonable resource with lots of experience. Lets say he has estimated 20 days for a particular task.
Now it is time to implement. Our resource knows that if he were not disturbed at all he would easily finish complete the task in 5 – 10 days. So, he exhibits ‘student syndrome’ and works on something more urgent instead (not necessarily important and urgent, but some manager is putting pressure on him. So, he works on other things and time passes – wasting the particular tasks planned time and consuming the safety buffer. Eventually our task becomes urgent – the time remaining is now so little, that our resource has at best a 50:50 chance of finishing on time. What was 80-90% certain is now just 50:50.
Alternatively our resource has no other pressures – so he does start the task immediately and, because he is not disturbed much (or can resist the pressure to do something else) he finishes in 12 days – half the estimated time.
Now our resource has a problem. If he actually delivers his task, he will be scrutinized the next time. There is a fair chance that his time estimates will be cut more than usually. What does the resource do? He either delivers the task early (assuming his boss will understand) or he will deliver on the due date – which will show great reliability and confirm his time estimate. If he takes the second option (the normal option?) he is wasting a lot of time that could be used by a task that is in real trouble.
The two phenomena can explain why most tasks are on time, very few are early and some are late to very late (the very late ones are those that have something truly bad happen). Multi-tasking (too many active projects) aggravates the situation. Project managers put pressure on task owners to work on the ‘right’ task – the others suffer student syndrome.
What we have seen is that every task owner includes his own insurance (80-90% certainty of an on time finish) and then wastes it in one way or another. Why not make a simple change in the structure of project plans – aggregate insurance in one place in order to protect your project. Whether or not any individual task is late is immaterial. It is the project that needs to be on time.
So cut every task’s time to just a 50:50 chance of finishing on time and aggregate what you have cut at the project’s end. Remember also that with this aggregation you have the same situation as all insurances. You do not need all that safety – some tasks will still finish early (a 50:50 chance means 50 out of 100 will be early and 50 late). The amount of safety can be cut.
Assuming this sounds reasonable; there is a major obstacle to overcome – common practice. How do I make the change and what other actions are essential for success?
<
Technorati Tags: 6-Sigma, Critical Chain, Execution Management, Goldratt, Project, Project Management, Project Manager, Projects, Risk Management, Theory of Constraints, TOC, Insurance, New Tag
/p>
No comments:
Post a Comment