The Burden of Proof in Project Estimating

Conrad Weisert, November, 1998
©Information Disciplines, Inc.

"I didn't call this meeting to discuss whether we can meet the deadline. We're here to decide how we're going to meet it."

"What we've got to do now is to roll up our sleeves and do whatever it takes to get the job done!"

"I agree with you in principle, but this project is so urgent that we just don't have the luxury of doing it right."

from the Management by Cliché Handbook.

"You can commit to accomplishing just about anything in 18 months."

from the Cynic's Guide to Practical Project Management

Major disappointments are still common

The passing decades bring disappointingly little improvement in the manageability of major software development projects. In one organization after another, we hear of some ambitious project that has experienced a huge schedule slippage and a massive cost overrun. Some of those projects may have been pioneering new methodology, such as object-oriented design and programming, and in the bitter aftermath controversy rages over the degree to which the methodology is to blame for the fiasco.

Focus on the estimates

In the great majority of those fiascos, however, we shouldn't condemn the project's performance as much as its original estimates of duration and cost. Even when the estimates are substantially raised during the course of the project, they still turn out far short of reality.

Why do supposedly competent software professionals year after year persist in seriously underestimating their projects?

The main fault most often lies less with the project team members than with management -- both in the development organization and in the sponsoring user organization -- and the main mistake they make is usually the same: They accept the first estimates they hear that fit their desires. (Or, they may dictate a "deadline" based on factors unrelated to the project itself. See "Deadlines and Target Dates: 5 Guidelines for Avoiding Disaster".)

Credibility and the project plan

What many managers fail to grasp is this simple principle of project management:

In assessing the credibility of a project estimate the burden of proof falls on those who claim it can be done.

In all too many organizations, management takes exactly the opposite view. Someone who questions a target date or project budget is castigated as a poor team player, while someone who tells management what they want to hear gets rewarded in the short term.

That's understandable, of course, when the organization stands to benefit greatly from the optimistic view or to suffer greatly otherwise. That approach, however, virtually guarantees eventual costly project failure.

What kind of proof?

Except in trivial small projects, estimates of duration and cost are derived from a detailed project plan. Credible estimates are not influenced by desirability or by the threat of dire penalty, although such factors may affect the priority of a project relative to other projects.

Whenever a project manager presents his or her estimates of the duration and cost of a project phase, management should expect and demand to see:

(Specific criteria for project plan credibility are presented in IDI's course PM-01 Project Management Concepts and Techniques. Also see some good books on project management)

In the absence of such clear and convincing evidence, management must avoid making any commitment based on the estimates. If a manager, after looking over the project plan and meeting with the project manager, still feels unsure, then he or she should solicit an unbiased evaluation from an uninvolved project-management specialist.

And never punish or intimidate anyone in your organization who calls attention to unrealistic estimates. The whistle-blower may save you from a costly and embarrassing fiasco.

Return to Management articles
Return to IDI home page

Last modified (minor format changes) November 21, 2003