## Estimating Project Budgets

In order to develop a budget, we must forecast what resources the project will require, the required quantity of each, when they will be needed, and how much they will cost—including the effects of potential price inflation. Uncertainty is involved in any forecast, though some forecasts have less uncertainty than others. An experienced cost estimator can forecast the number of bricks that will be used to con-

struct a brick wall of known dimensions within 1 to 2 percent. (The estimator knows almost exactly how many bricks are needed to build the wall and must simply add a small allowance for some faulty (broken or discolored) bricks and a few more being broken during the construction process.) On the other hand, the errors are apt to be much larger for an estimate of the number of programmer hours or lines of code that will be required to produce a specific piece of software (see Section 7.2). While the field of software science makes such estimations quite possible, the level of uncertainty is considerably higher and the typical error size is much larger.

In many fields, cost estimation methods are well codified. The office walls of organizational purchasing departments are lined with catalogues detailing what materials, services, and machines are available, from whom, and at what prices. Also on the book shelves are volumes devoted to the techniques of estimating the quantities of materials and labor required to accomplish specific jobs. Every business has its own rules of thumb for cost estimation. These usually distill the collective experience gained by many estimators over many years. An experienced producer of books, for example, can leaf through a manuscript and, after asking a few questions about the number and type of illustrations and the quality of paper to be used, can make a fairly accurate estimate of what it will cost to produce a book.

At times, the job of cost estimation for entire complex projects may be relatively simple because experience has shown that some formula gives a good first approximation of the project's cost. For example, the Goodyear Aircraft Company makes an initial estimate of the cost of building a blimp by multiplying the estimated weight of the blimp by a specific dollar factor. (The weight is estimated in pounds, presumably prior to the blimp's inflation with helium.) The cost of buildings is commonly estimated as dollars per square foot times the square feet of floor area. Obviously these approximations must be adjusted for any special characteristics associated with each individual project, but this adjustment is far easier than making an estimate from scratch.

We will have more to say about gathering data shortly. Before doing so, however, and before discussing budget construction and presentation, it is helpful to understand that developing project budgets is much more difficult than developing budgets for more permanent organizational activities. The influence of history is strong in the budget of an ongoing activity and many entries may ultimately become just "last year's figure plus X percent," where X is any number the budgeter feels "can be lived with," and is probably acceptable to the person or group who approves the budgets. No single item in the budget for an ongoing activity is apt to be crucial, because over the course of years the budget has gained sufficient slack that internal adjustments will probably take care of minor shortages in the key accounts.

But the project budgeter cannot depend on tradition. At project inception, there are no past budgets to use as a base. At times, the budgeter may have budgets and audit reports for similar projects to serve as guides, but these are rough guides at best. Tradition, however, has another impact on budgeting, this time a helpful one. In the special case of R & D projects, it has been found j 12| that project budgets are stable over time when measured as a percent of the total allocation to R&D from the parent firm, though within the project the budget may be reallocated among activities. There is no reason to believe that the situation is different for other kinds of projects, and we have some evidence that shows stability similar to R&D projects.

For multiyear projects, another problem is raised. The plans and schedules for such projects are set at the beginning of project life, but over the years, the forecast resource usage may be altered by the availability of alternate or new materials, machinery, or personnel—available at different costs than were estimated. The longer the project life, the less the PM can trust that traditional methods and costs will be relevant. As if that were not enough, the degree of executive oversight and review is usually much higher for projects than for ongoing operations, so the budgeter must expect to defend any and all budget entries.

Tradition has still another impact on project budgeting. Every organization has its idiosyncrasies. One firm charges the project's R&D budget with the cost of training sales representatives on the technical aspects of a new product. Another adopts special property accounting practices for contracts with the government. Unless the PM understands the organizational accounting system, there is no way to exercise budgetary control over the project. The methods for project budgeting described below are intended to avoid these problems as much as possible, but complete avoidance is out of the question. Further, it is not politically feasible for the PM to plead a special case with the accountants, who have their own problems. The PM simply must be familiar with the organization's accounting system!

One aspect of cost estimation and budgeting that is not often discussed has to do with the actual use of resources as opposed to the accounting department's assumptions about how and when the resources will be used. For instance, presume that you have estimated that \$5000 of a given resource will be used in accomplishing a task that is estimated to require five weeks. The actual use of the resource may be none in the first week, \$3000 worth in the second week, none in the third week, \$1500 in the fourth week, and the remaining \$500 in the last week. Unless this pattern of expenditure is detailed in the plan, the accounting department, which takes a linear view of the world, will spread the expenditure equally over the five-week period. This may not affect the project's budget, but it most certainly affects the project's cash flow. The PM must be aware of both the resource requirements and the specific time pattern of resource usage. The subject will be mentioned again in Chapter 9.

Another aspect of preparing budgets is especially important for project budgeting. Every expenditure (or receipt) must be identified with a specific project task (and with its associated milestone, as we will see in the next chapter). Referring back to Figure 5-5, we see that each element in the WBS has a unique account number to which charges are accrued as work is done. These identifiers are needed for the PM to exercise budgetary control.

With these things in mind, the issue of how to gather input data for the budget becomes a matter of some concern. There are two fundamentally different strategies for data gathering, top-down and bottom-up.