Problem Statement In project management, accurate estimates are the basis of sound project planning. Many processes have been developed to aid engineers in making accurate estimates, such as
- Analogy based estimation
- Compartmentalization (i.e., breakdown of tasks)
- Cost estimate
- Delphi method
- Documenting estimation results
- Educated assumptions
- Estimating each task
- Examining historical data
- Identifying dependencies
- Parametric estimating
- Risk assessment
- Structured planning
We all know that cost and time overruns can lead to project cancellation or failure. Yet these overruns continue to happen. IT projects have a poor track record for meeting budget goals. What is surprising is that all project estimates go through a series of reviews for validation to ensure their correctness. Even with those reviews, we can’t seem to get it right for many — if not most — projects.
Despite the use of sophisticated tools and techniques many IT project cost estimates are still very inaccurate.
This inaccuracy is even more prevalent in IT projects involving new technology and software as there are no previous historical examples to compare with. Why that is and what can we do about it?
Why Do We Overrun?
Neuroscience and social science suggest that humans in general are more optimistic than realistic. So we tend to exaggerate our own ability and minimize potential difficulties. Project teams ignore lessons learned from previous projects, hoping that external factors influencing failure were one-time incidents. Illusion of control is the tendency to exaggerate one’s perception of control over external events, also possibly leading to problems when these events prove more difficult than expected. Short-term versus long-term perspective (i.e., lowering the estimate on purpose in order to “win the bid”)
There are typically four reasons attributed to inaccurate cost estimates.
-
Developing an estimate for a large software project is a complex task requiring a significant amount of effort.
-
The people who develop software and hardware cost estimates often do not have enough experience with cost estimation, especially for large projects.
-
Human beings have a bias towards underestimation.
- Management might ask for an estimate, but are really requesting a number to help them create a bid to win a major contract or get internal funding.
Minimizing the Gap
Although these factors cannot be completely avoided, organizations can take steps to limit their negative impact. Approaches include:
- Improving systems and procedures for historical data gathering and trend analysis
- Maintaining extensive checklists based on lessons learned from previous projects
- Collecting reference estimates for similar projects within the organization and/or industry to establish baselines
- Auditing forecasts to enforce transparency of information
Where possible, use multiple techniques for verification
- Analogous estimating – top down estimating (using actual costs of similar projects)
- Parametric modelling – using project characteristics in a mathematical model projecting costs
- Bottom – up
- Computerized tools
- Procurement estimating – Vendor bids
- Use a systematic approach
- Use multiple estimators
- Document assumptions and techniques used
- Opportunity evaluation and project initiation represent 5-10% of project effort
- Analysis and design represent 50 - 60 % of the total effort
- Construction represents 30 - 40% of the total effort
- Implementation and evaluation represent 5 - 15% of the total effort
- Retain estimates and assumptions for future use
- Use comparison techniques for explaining variances and arriving at consensus
- Top down approaches = underestimates
- Bottom up approaches = overestimates
