Cone of Uncertainty

---

Accounting web has a good report on the The House of Commons Public Accounts Committee report entitled "Delivering successful IT–enabled business change". (Bear with me, I'll get to the cone in a minute.)

Amongst other refreshing expressions of common sense Sir John Bourn (Comptroller and Auditor General) is quoted as testifying that: "In our experience, many of the difficulties have arisen when a project with many elements is sometimes taken forward with a large number of partners, and the client does not have a full understanding of what he is trying to do and alters it as he goes along."

Never mind government IT projects, in my experience this is pretty much true of life. If you want to get the right stuff you've got to know what you want before people start delivering it to you.

Coincidentally I was introduced to Steve McConnell's Cone of Uncertainty yesterday, it certainly seems to imply, according to bar-room geometry, that a programme with many smaller projects is more likely to succeed (or at least behave reasonably predictably) than a more traditional Big Project.

That adds more weight to the notion that while the traditional waterfall method of managing projects may work for small projects it is fatally flawed when projects are large. In other words if they are large enough that phases cannot be completed reasonably quickly and/or that delivery dates are so far in the future that clients needs are bound to change in the meantime, and that's notwithstanding the I-remembered-another-thing-we-have-to-do scope creep syndrome.

However an experienced project manager I was talking to yesterday made the point that agile methodologies can be difficult to plan and manage, and even harder to estimate, without significant experience.

Perversely the risk of not having concrete dates and costs available at the start often seems to outweigh the risks associated with life intervening in between plan and execution.

---