Estimating is one of the most controversial subjects in project management. There are some people who have turned the subject into a cryptic science involving esoteric techniques bearing a close resemblance to "voodoo."

Estimating is simply the process used to determine the amount of effort and cost required to implement a project, in part or in full. It is important to acknowledge that estimating is fundamentally an effort at projecting the future. Like all projections, the more facts and information available, the better the estimate. There is a natural human tendency to avoid making estimates because estimates are commitments, and people tend to shy away from commitments, particularly when they are not sure of the facts. Nevertheless, little progress would be made if we never attempted to plan for the future.

Most estimating errors are errors of omission, not commission. It is what we forget to estimate that often leads to problems. Methodologies, with their defined structure, materially assists with eliminating some of the unknowns when estimating. They provide frameworks and structures that act as checklists for estimating. Methodologies isolate the activities that have to be performed into small enough increments, thereby minimizing the margin of error.

An estimate improves in accuracy in direct relation to the level of detail considered. A methodology defines the sequence by which parts are assembled. For example, a construction methodology identifies all of the resources of a product, such as lumber, steel, glass, etc. and how they are assembled. An IRM related methodology specifies the sequence by which data elements, records, files, inputs, outputs, processes, etc. are assembled. This provides the ability to use a "bill of materials" technique to count all of the resources in the product and develop an estimate for the project, in part or in full, based on the standards developed for completing and/or installing a resource. This is why the "PRIDE" methodologies put in emphasis on "rough designs" in the early phases of work. At this time, the Project Manager is asked to consider:

• The number and types of NEW resources to be created.

• The number and type of existing resources requiring MODIFICATION.

• The number and type of existing resources that can be RE-USED as is (no modification required).

To illustrate, in a "PRIDE"-ISEM Project (Phase 1), a complete "rough design" of the envisioned system is produced in Activity F. In Activity G, the Project Manager takes the rough design and makes the following type of assessment:

IRM RESOURCE	NEW	MODIFY	RE-USE
SYSTEM	1
SUB-SYSTEMS	14
ADMIN PROC	23
COMP PROC	13
PROGRAMS	28
MODULES	33	10	112
INPUTS	17		5
OUTPUTS	37	13
FILES	56	5	43
RECORDS	250	50	306
DATA ELEMENTS	60		257

This analysis of the rough design is used as the roadmap for the project (in the above example, there will have to be 14 Phases 3-7 because there are 14 sub-systems and 13 Phase 4-II & 6 for the 13 computer procedures, and at least 28 Phase 5's for the programs). It is also the basis for the project estimate. Such estimating is greatly facilitated through the use of an IRM Repository.

The concept of direct and indirect hours permits the establishment and application of estimating guidelines. To illustrate; how much direct time does it take to weld a six inch pipe? Define a data element? Define an information requirement? It would be almost impossible to establish guidelines if indirect hours were included in the guidelines. Indirect time is a part of the work environment and can vary from company to company, even from group to group. Estimating, therefore, must be accomplished using direct hours only.  

TYPES OF ESTIMATES

There are fundamentally two types of estimates: Order-Of-Magnitude and detail. Both are expressed in direct hours and both are based on the resources of the system as known at that point in the methodology.

The Order-Of-Magnitude (O-O-M) is an estimate of the effort remaining to complete a project and is typically prepared by the project manager or project team. It is expressed in terms of remaining phases, not activities. The O-O-M estimate is principally used to develop a cost evaluation summary which contains such things as "return on investment" and "break even points." These figures are used by management for decisions relative to the project's economic feasibility and priority. Based on this information, management may determine to continue, discontinue or revise the project within a reasonable level of accuracy.

The Detail Estimate is an estimate of the effort required to complete the activities of an individual phase and is prepared prior to the start of the phase by the individual or individuals assigned to the phase. It is based on detail specifications defined in the preceding phase. The important difference here is that the detail estimate is an expression of someone's personal commitment to a project and is based on detail knowledge and study. Since it is prepared and expressed in direct hours by the individual assigned to it, that person becomes responsible for meeting the estimate. In this way, the staff is treated as professionals and held accountable for their actions. By working with direct hours, more accurate estimates are made. If the analyst has to take into account scheduling and indirect time, the estimates become meaningless.

TYPES OF ESTIMATES

	ORDER-OF-MAGNITUDE	DETAIL
PURPOSE	BASIS FOR PROJECT DECISIONS	EXPRESSION OF PERSONAL COMMITMENT
PREPARED BY	PROJECT MANAGEMENT OR PROJECT TEAM	INDIVIDUAL(S) ASSIGNED TO PERFORM THE WORK
WHEN PREPARED	END OF PHASE	BEGINNING OF PHASE
SCOPE	THROUGH THE END OF THE PROJECT	THROUGH THE END OF THE PHASE
LEVEL OF DETAIL	FOR ALL PHASES WITHIN A PROJECT	FOR ALL ACTIVITIES WITHIN A PHASE
SKILL LEVEL CONSIDERED	AVERAGE	UNIQUE TO WORKER

ALL ESTIMATES ARE EXPRESSED IN DIRECT TIME

Is it possible to develop an accurate estimate for an entire project from the outset? Yes and no. If the nature of the project is to modify or maintain an existing product or system, the answer is 'yes.' In this situation, the parts of the product affected are known and the level of complexity deduced. The answer is 'no' when the variables of the product are unknown. For example, it would be impossible to estimate the cost for constructing a building without first detailing requirements and preparing a rough design. The same is true in the systems field. It is a relatively simple task to estimate a systems project where the variables are known, such as the processes, inputs, outputs, files, records and data elements. However, if the resources are unknown, a preliminary phase involving requirements definition and rough design is needed. In this regard, there is virtually no difference between a systems project and a construction or manufacturing project. To try to estimate the time and cost of an entire project without first knowing all of the variables is like asking a physician to diagnose a problem without a preliminary examination. It is sheer nonsense.

In terms of estimating guidelines, there is very little difference between the two estimates. Both are based on the resources and the steps involved in the project. The fundamental difference is the O-O-M estimate is based on an average skill level and the Detail estimate is based on the specific skill level of the assigned Resource. Whereas the O-O-M estimate can use historical data of an entire group of people, the detail estimate would use only historical data of a particular person.

No matter how accurate or inaccurate the generated estimates are, the individual makes the final decision on whether to accept, revise, or reject them. This is because the estimates are intended to be an expression of personal commitment. If an estimate is missed, it is because it was prepared incorrectly, not because the guidelines are wrong. The guidelines can only advise, they are not a panacea. The person preparing the estimate must make the ultimate decision as to the realism of the figures. It should always be remembered that nothing will ever be accomplished if the individual does not assume responsibility for the estimates made. Management can make the time available, but it takes a motivated employee to take advantage of the direct time made available.

One last note: the project estimate, combined with effectiveness rate and the corporate calendar, is the principal basis for plotting project schedules (not the other way around). In other words, we formulate the estimate first, then compute the schedule second.  

EXAMPLES

Samples of both Detail and O-O-M estimates are included throughout the "PRIDE" methodologies. Consult the Examples sub-sections under the various activities. For forms, use the "Project Control Sheet" to prepare Detail estimates (also see alternate form), and the "Project Estimate/Schedule Recap" worksheet for O-O-M estimates; both can be obtained in the "PRIDE" Forms section.