The Challenge

An estimate is a prediction based on probabilistic assessment of the total cost of the tasks, time, and resources required to deliver a project’s scope of work. Even people with little understanding of probability seem to accept the probabilistic nature of estimating, but there is no real consensus about which probabilistic assessment ought to govern an estimate.

PMI’s 2018 findings show that only 57% of projects are finished within their initial budgets. Although numerous tools and techniques can assist in creating project cost estimates, many IT project cost estimates are still inaccurate, especially those for software development and/or new technologies.

Defining the right software development process for your organization will have a profound impact on controlling schedules, costs, and quality of a project. Tom DeMarco, a well-known author of software development, suggests four reasons for these inaccuracies and some ways to defeat them.

1. Lack of Estimating Experience

The people who develop software cost estimates often do not have much experience with cost estimates, especially for large and/or new technology projects. They also do not have enough accurate, reliable project data on which to base their estimates. If you do not collect data to use when preparing estimates, you will always be guessing.

Other factors that can influence the project’s budget are scope, quality, schedule, human resources, risk, and procurement. But for an organization that uses good project management techniques and develops a history of keeping reliable project information like the lessons-learned database and/or cost database, the organization’s estimates should improve over time.

People typically learn cost estimating through a combination of:

• Formal education which could include online cost estimating courses
• On-the-job training
• Industry certifications
• Mentorship from experienced estimators
• Practical experience working on real projects

2. Biased Toward Underestimating

The bias towards underestimating is where people tend to be overly optimistic about time, cost, and resources required to complete a project. Often focusing on the best-case scenario and neglecting potential obstacles or complexities, this leads to inaccurate underestimations.

Estimators might also forget to allow for extra costs needed for integration, testing, and resource constraints on large and/or new technology IT projects. Also important is integrating risk analysis by adding contingencies if needed.

It is important for PMs and top management to review estimates and ask important questions to make sure the estimates are not biased—which can mean the difference between getting projects up and running quickly vs. encountering delays or cancellations.

3. Estimates Are Done Too Quickly

Many estimates must be made quickly and before clear systems requirements have been produced, which requires significant effort. For example, the Surveyor project was a NASA program that sent a series of robotic spacecraft to the moon’s surface with the primary goal of demonstrating the feasibility of soft landings—which obviously involved a lot of complex software development. Before fully understanding what information evaluators need in the system, somebody would have to create rough order of magnitude (ROM) estimates for this project.

The Estimation Process

ROM estimates are considered calculated guesses based on high-level knowledge of the target. The next level of estimating is the definitive or budgetary estimate. These are typically derived by reviewing each work unit in the WBS and adding the results, which are usually higher than the ROM estimates.

Estimates are done in various phases because it allows for:

• More accurate assessment of the project’s overall time and cost by breaking down complex work into manageable segments
• Better understanding of potential risks and uncertainties at each stage
• Opportunities to refine estimates as more information becomes available throughout the project lifecycle

Finally, use the previously mentioned lessons-learned database and/or cost database for assistance (i.e., capturing and leveraging historical information) in future projects.

4. Management Desires Accuracy

Management might ask for an estimate but really want a precise picture of the project’s cost to help them create a bid to win a major contract or get internal funding. This situation could turn into a project time management problem in which top managers or other stakeholders want project schedules to be shorter than the estimates.

The additional pressure to meet specific project deadlines can encourage individuals to underestimate effort to appear more capable. It is always important for PMs to help develop good cost and schedule estimates which minimizes the risk of cost overruns and ensures that funds are allocated appropriately—and to stand by those estimates.

Management Reserve Strategy

For large and/or long projects, a Management Reserve (MR) should be set up which is for unforeseen work (i.e., unknown unknowns) that is within the scope of the project. In this case:

Project Budget = Cost Baseline + Management Reserve

Typically the MR is 5-15% of the Cost Baseline, which is dependent on project complexity.

The Reality of Estimation

It is important to be cautious with initial estimates. Project budget accuracy is an often-misunderstood concept. Too often an early initiation phase rough estimate gets carried into an approved project budget before detailed planning takes place.

It’s extremely important to have a process in place for informing top management of any revised cost estimates for the duration of the project. Cost estimation helps you determine your project’s budget, scheduling the necessary work, and managing new resources. Likewise, reliable estimations are crucial when it comes to winning new business.

“You can’t control what you can’t measure.” – Tom DeMarco

Your feedback is always welcome here in the comments, and in the Community Discussion Forum!