Estimates are a means by which the project manager can get a notion of the scale and scope of the project and make important decisions on how it will be tackled.
Projects can range from a feasibility study, through modification of existing facilities, to complete design, procurement, and construction of a large complex. Too many projects have gone badly over time and exceeded their budgets and the blame has often been put on the original estimates. Whatever the project may be, whether large or small, the estimate and type of information desired may differ radically.
All estimating methods are prone to error and all depend to some degree on subjective views of the size and complexity of the task ahead. No single method is going to give a ‘right’ result and, in fact, a ‘right’ result is probably unachievable.
To save time, companies usually start with a parametric estimate, an analogy estimate, or expert judgment. A parametric estimate is based upon statistical data. An analogy estimate compares the features of a completed project to the features of the new project and adjusts the costs based upon the change in the degree of difficulty. Expert judgment comes from individuals or groups with specialized knowledge because of previous work on these types of projects.
In In the following section, we examine a number of the most commonly used methods for preparing estimates.
Types of Estimates
Order-of-Magnitude
The first type of estimate is an order-of-magnitude analysis, which is made without any detailed engineering data. The order-of-magnitude analysis may have an accuracy of ±35 percent within the scope of the project. This type of estimate may use past experience (not necessarily similar), scale factors, parametric curves, or capacity estimates (i.e., $/# of product or $/kW electricity). Order-of-magnitude estimates are top-down estimates usually applied to level 1 of the WBS, and in some industries, use of parametric estimates is included.
Approximate Estimate (Analogous or Top-Down estimating)
Next, there is the approximate estimate which is also made without detailed engineering data, and may be accurate to ±15 percent. This type of estimate is prorated from previous projects that are similar in scope and capacity, and may be titled as estimating by analogy, parametric curves, rule of thumb, and indexed cost of similar activities adjusted for capacity and technology. In such a case, the estimator may say that this activity is 50 percent more difficult than a previous (i.e., reference) activity and requires 50 percent more time, man-hours, dollars, materials, and so on.
The analogous estimating is one of the oldest, but one of the most reliable, of methods and depends on finding a project similar to the current one which has already been undertaken in the organization. The similarity should ideally extend to:
- The type of business involved.
- The overall size of the applications.
- The general scope of the systems – for example, the ratio of online to batch functions, whether there is a major communications component and so on.
- The technical methods, standards and languages used.
Where there are differences in any of these areas, suitable adjustments must be made. In addition, some judgement needs to be made if there are likely to be other significant differences between the historical project and the new one, for example: ( The customer’s company culture, The customer’s level of computer literacy, The degree of management support for the project.)
The major advantage of the analogy method is that it enables a broad-brush estimate for a whole project to be developed fairly quickly, perhaps during the preparation of a bid or a proposal.
Analogous estimating is generally less costly and time consuming than other project estimation techniques, but it is also less accurate
Analogous estimating is used to estimate project cost early in the project, when detail project information is scarce.
Parametric Estimate
Parametric estimating involves using project characteristics (parameters) in a mathematical model to predict project costs.
A parametric estimate is based upon statistical data. For example, assume that you live in a Ottawa suburb and wish to build the home of your dreams. You contact a construction contractor who informs you that the parametric or statistical cost for a home in this suburb is $150 per square foot. In Toronto, the cost may be $250 per square foot.
There are Two types of parametric modeling:
Regression analysis: scatter diagram
Tracks two variables to see if they are related and creates a mathematical formula to use in future parametric estimating
Learning Curve:
Graphical presentations or repetitive activities in which continuous operations will lead in reduction in time, cost and resources : “The 100th room painted will cost less than the first room because of improved efficiencies”.
Definitive Estimate
The definitive estimate, or grassroots buildup estimate, or bottom-up estimate is prepared from well-defined engineering data including (as a minimum) vendor quotes, fairly complete plans, specifications, unit prices, and estimate to complete. The definitive estimate, also referred to as detailed estimating, has an accuracy of ±5 percent.
Accuracy Ranges:
Best Management Practices:
- Estimating should be based on the WBS to improve accuracy
- Estimating should be done by the person doing the work
- Historical information is the key to improving estimates
- Costs (and time, scope and resources) should be managed to estimates
- A cost (and time, scope and resources) baseline should be kept and not changed except for approved project changes
- Corrective actions should be taken when cost problems (and time, scope and resources) occur
- A project manager should never just accept time or cost requirements from management: The PM and the project team should analyze the needs of the project, come up with their own estimate and reconcile any differences.
