Accuracy errors in project estimates can send IT budgets into disarray, and possibly threaten your own performance and position. And while project management (PM) estimates are just that--estimates--there are several approaches to getting a clear idea whether they're on target.
There are many different models and approaches CIOs and tech leaders can use to effectively gauge the accuracy of project management estimates. There are also best practices that can provide a good comfort level as to how well the project team is providing cost estimates and forecasts.
The value of historical project data
A key factor in good estimate accuracy is updating estimates through the entire system development life cycle. As you get closer to project completion, most unknowns are better understood, and estimates become more likely predictors of actual results. CIOs also should obtain several estimates--optimistic, most likely, and pessimistic-to gain a sense of the potential project time and cost range.
If you have empirical data on similar completed projects, you already have some initial information to gauge the accuracy of PM estimates.
A project manager needs to use a good tool that will allow him or her to create an initial estimate of the project's physical and human asset costs and schedule, and this should be placed into a baseline in the project plan.
A baseline is a project's initial budget. End results of all previous projects can then be compared to project baselines to ascertain the accuracy of PM cost estimates and forecasts.
This metric can be a very good gauge of how accurate your current and future PM estimates will be.
Best practices for PM estimating
If you don't have previous data on project performance, there are some best practices that can lead to good cost estimates. Keep in mind that even if you do have historical data, it's not a good idea to use that data alone. As I have found in the past, empirical data is not always a good predictor, but it can be useful for initial estimates.
Employ modeling techniques at project start
Best practices in PM estimation call for accurate predictions of a project's time and costs. To begin this process, you must know what tasks or deliverables are expected in the project. After you know the tasks, then estimates can be made as to timeframes and associated costs.
The use of modeling techniques in the requirements-gathering phase of a project is the best method I've seen to drive out tasks. If your project team is creating context models, event models, and information models-such as entity relationship diagrams (ERDs)-your project team will likely uncover all end user requirements and be able to develop a viable task list.
Context model: A good context model defines all of the project's activities and processes. It involves data flow diagrams, as well as internal and external inputs and outputs. It will show the entire scope of the project system and will greatly reduce the potential for scope creep. Minimising scope creep will allow your initial estimates to better equal final project results.
Event model: The event model details the requirements of a project from an end user's perspective, rather than internal system events. An event list is created describing all customer or end user interactions with the new project or system. Event modeling is quite simple to perform, and will further ensure that your project team leaves no stone unturned. This also will provide a more thorough project task list or work breakdown structure (WBS).
Information model: The information model, which is graphically displayed by ERDs, shows all of the data that needs to be captured by the project system. This is also known as a data model. Capturing all of your data requirements will enable you to double-check the event model. The information model is probably the most important of the three because it ensures that all project tasks become known.
Utilising these three simple modeling techniques results in a thorough listing of project tasks, which in turn provides a complete WBS. When the WBS is placed into a good PM tool, accurate estimates and forecasts for any given project can begin.
Insight on modeling techniques
David A. Ruble's book, "Practical Analysis and Design for Client/Server and GUI Systems," provides excellent detail on these modeling techniques. If you need to gauge how well your project team is providing cost estimates, ask if it is performing the three simple but effective modeling techniques that I outlined above.
In my experience as a project manager, modeling is not often done. The estimation process is usually haphazard and not well thought out. Actual project results almost always deviate from estimates unless a good modeling process is employed. A good modeling process creates an accurate WBS to base estimates on.
Using these modeling techniques will provide your project team with a good start in the planning process, and will ultimately create viable project plans. A good project WBS will provide the basics for viable project estimates and forecasts.
6%
2%
