When designing the control system, there are certain helpful guidelines to keep in mind. For instance, the primary purpose of the control system is to correct error, not to identify and punish the guilty. Managers must realize that the past cannot be changed, no matter how loudly they yell. Moreover, investment in control is subject to sharply diminishing returns. The cost of control increases faster and faster while the degree of control—and its value—increases more and more slowly.
We could, if we wished, control the quality of “hardware store” type wooden yard- sticks, so that they were accurate to 1/10,000th of an inch. But who cares? Moreover, who would be willing to invest the increased cost in a product that is usually given away as a gift? Thus, there is some optimum amount of resources worth investing in the con- trol process. The control system should exert control only to the degree required to achieve its objectives; additional control will not help and may be cost-inefficient.
In the same vein, as the degree of control increases beyond some difficult-to-define point, innovation and creativity are discouraged until they are finally shut off com- pletely. In general, the control system should employ the lowest degree of hassle consis- tent with accomplishing its goals. It is best to avoid annoying those people whose cooperation is needed to reach project goals. To summarize, the control system should be cost-effective and should operate with the minimum force required to achieve the desired end results.
Project control, the final activity in the planning-monitoring-control cycle, in- volves taking action when reality deviates from plan. It includes both mecha- nistic and human elements, and because it is closely concerned with human behavior, is one of the most difficult tasks of the PM. It includes two seemingly antithetical purposes: (1) stewardship of the organization’s and the client’s, physical, human, and financial assets, and (2) the use of these assets to bring project actuality into conformance with the plan. Somehow, the PM must meld these two purposes into a uniform focus of activity.
7.5 DESIGNING THE CONTROL SYSTEM • 257 There are three primary mechanisms by which the PM exerts control: process re- views, personnel assignment, and resource allocation. The process review is directed to an analysis of the process of reaching the project objectives rather than on the results, per se. Because results are largely dependent on the process used to achieve them, the process can be subjected to control even if the results cannot. On R&D projects, for ex- ample, project team members usually cannot be held responsible for research or techno- logical outcomes, but they can certainly be held responsible for adherence to the proposal, the budget, and the schedule. Care must be taken, however, not to overstress method as opposed to results; although methods are controllable, it is still the results that count.
As an example of the effective use of process controls, Australia’s 7-year Parliament House construction project matched the inherent construction complexity with an equally sophisticated set of schedule, cost, and time controls. The time controls, for in- stance, included four levels of increasing detail that could be accessed during progress review and coordination meetings (Nixon, 1987).
Control can also be exercised through personnel assignments based on past project productivity. Although it is relatively easy to separate workers in the top and bottom quartiles by measuring performance or productivity, separating the middle two quartiles is much more difficult so this approach should be used carefully. Moreover, reassign- ment can have drawbacks in terms of creating elite groups of top producers but demoti- vating everyone else in the system.
Controlling resource allocation can be a powerful motivator—and demotivator.
Resources are usually allocated to the more productive or important tasks and this can significantly influence the attainment of project results. (Remember that all tasks in a project must be completed to complete the project.) As in the use of other control tech- niques, the PM needs to exercise care when making decisions about which tasks need the resources in the future, regardless of past efficiencies.
There are some common mistakes PMs and other organizational managers make when trying to control projects. For example, when controlling processes, there is the danger of emphasizing short-run results at the expense of long-run objectives. Excessive control directed to specific objectives can result in sacrificing other project objectives.
Across-the-board cuts in resource allocations tend to reward those who have already overspent or overhired while penalizing the frugal and efficient. Finally, focusing on certain items for control can distract the attention of team members from other, equally important items: “What isn’t counted, doesn’t count.” There is hardly a community in the United States that has not adopted a set of standardized tests to measure the learn- ing of school children. The salaries of teachers are affected by these test results. It should, therefore, hardly come as a shock to the public that teachers spend considerable time and effort “teaching to the test.”
Types of Control Systems
The process of controlling a project, or any other system, is more complex than might be expected. Decisions must be made concerning where in the project we will try to exert control, what is to be controlled, how it will be measured, and how much deviation from plan will be tolerated before we intercede. It is helpful in making these decisions first to understand thoroughly the primary types of control systems used by project managers:
go/no-go controlsand post-control.
Before discussing the nature and use of these control systems, it is important to note that every control system must contain certain elements if it is to be useful. Any project (or production system) can be described in terms of its inputs, the process by which it
works on the inputs, and the outputs that result. To control a project (or any produc- tion system) requires the following components:
1. Each control must have a sensor, the duty of which is to measure any aspect of the project’s output that one wishes to control. For instance, it must be able to measure the length of the yardstick we mentioned earlier.
2. The control system must have a standardfor each thing measured. For the yardstick, the standard is 36 inches.
3. Next, the control system needs a comparator, a mechanism that compares the output of the sensor with the standard.
4. Given the results of the comparison, the control system needs a decision makerto decide if the difference between what the sensor measured and the standard is large enough to warrant attention. For the yardstick, a group of 23 engineers attending a project man- agement seminar did not expect (or require) accuracy greater than plus or minus 1⁄4 inch. For more precise measurement they would insist on a steel tape measure.
5. The final piece required in a control system is an effector. If the decision maker de- cides that some action is required to reduce the difference between what the sensor measures and the standard requires, the effector must take some action. It may oper- ate on the input or the process (or both), to fix the problem.
Some control systems use all five of these elements automatically, for example, the mechanisms that control the ph (level of acidity) of the blood, the temperature of a home, the speed of an automobile operating on cruise control, or the speed with which a sheet of steel moves through a rolling mill. Such systems are called “cybernetic con- trol systems” (kybernet is Greek for helmsman or steersman) or “negative feedback loops” or “steering controls.” All of the parts listed above must be present and operative to control any process. It is the PM’s responsibility to ensure that such control elements are available and operating in project control. The process will not be automatic during project control. It must be operated by the PM or the PM’s deputy.
The go/no-go control takes the form of tests (sensors) to determine if some spe- cific precondition (standard and comparator) has been met before permission is granted to continue (decision maker and effector). This type of control can be used on almost every aspect of a project. The project plan, budget, schedule, earned value charts, and other such information can all operate as control documents so the PM has prespecified milestones (standards) as control checkpoints. The PM can intercede at any level of the project tasks and subtasks for which detail is available in these control documents.
It is worth repeating that the primary aim of the PM is to intercept problems before they arise—at least before they get serious—so it is worthwhile for the PM to include an early warning system with the control system. In this way, potential problems can be ex- posed and dealt with before they turn into full-blown disasters. Because any project early warning system will include people, it is important for the PM to make it known that the messenger who brings bad news will not be shot, but anyone who sweeps bad news or problems under the rug will be!
An example of a status report used by the agricultural products division of a large chemical company for go/no-go control is illustrated in Figure 7-7. As can be seen, some of the tasks are completed, some are in progress, and some have not yet started. Details from the PM or additional reports about those tasks completed and in progress are used to make the go/no-go decisions.
Post-controls, also known as postperformance reviews, are applied after the project has been completed. Although it might appear that this is the legendary situation of
7.5 DESIGNING THE CONTROL SYSTEM • 259
locking the barn door after the horse has been stolen, it is not. The purpose here is not to control the already completed project, but to allow future projects to learn and profit from past project experience. Such lessons might include information about certain sup- pliers, cost-estimating procedures, or even ways of improving the process of managing projects. Certain managerial methods, organizational procedures, or processes might be altered for future projects, resulting in greater predictability and control and, one hopes, better performance, cost, or schedule results.
The earlier the PM can intercede in a problem, the more likely the project team will be able to correct its activities, but humans respond to these controls in different ways. Response to go/no-go controls tends to be neutral. Because there is no gradation between excellent and barely acceptable, or between terrible and just unacceptable, the fine line of acceptability thus becomes a very sharp knife, subject to complaint and irritation. In a project, however, criticism tends to be leveled toward the team instead of the individual so the response is often less severe than it might have been otherwise.
Tools for Control
We have already described some of the tools that can help the PM in designing and ap- plying the control system: variance analysis, trend projections, and earned value analy- sis. With trend projection, for instance, the PM can plot a budget, plan, or growth curve as shown in Figure 7-8 and then, as actual values come in with project progress, plot these as a dashed line on the same chart. Using the dashed line, the PM can forecast on a continuing basis what the projected completion will be. Based on the projection, the PM can decide if there is a problem, what alternatives exist for control, what they will cost and require, and what they will achieve.
Another useful tool for a PM is the critical ratio. A critical ratio indicates to a man- ager when a task or process is becoming unacceptable, typically when the ratio drops below 1. By tracking the ratio, the manager can anticipate when a problem may be brewing. The calculation of the critical ratio for project tasks is the product of a schedule
Figure 7-7 Sample project milestone status report.
ratioand a cost ratio. The schedule ratio is actual progress divided by scheduled progress, as measured by some common standard such as earned value: EV/PV. Clearly, ratios greater than one are desirable. The cost ratio is budgeted cost divided by actual cost, or if earned value data are available, EV/AC. Again, values greater than one are most de- sirable. Taking the product of these two ratios thus gives us an overall measure that in- cludes performance, cost, and schedule.
Note that the two ratios are equally important in the calculation of the critical ratio. If one ratio is bad, it can be offset by the other ratio if it is equally good. For exam- ple, if the actual progress is 2 and the scheduled progress is 3, resulting in a schedule ratio of 2/3, and the budgeted cost is 6 and the actual cost is 4, resulting in a cost ratio of 3/2, their product, 2/3 3/2 1. Thus, although the project is behind schedule, the cost is correspondingly below budget so everything is fine if lateness is no problem for this activity. This may not really be acceptable to the PM so a wise PM will evaluate the components of a critical ratio as well as the overall value. Remember that the mea- surement of progress is subject to the same warnings we noted in the discussion of earned value.
Similar calculations are given in Table 7-3 where the above example is an illustra- tion of Task Number 1. Task 2 is on budget but progress is lacking so too much money was spent than should have been spent; when the project is completed, it will probably go over budget. Similarly, Task 3 is on schedule but is over budget, creating another probable cost overrun. The critical ratios for Tasks 4 and 5 both exceed 1.0, and none of the subratios is a problem. Task 4 is ahead of schedule and on budget, and Task 5 is on schedule and below budget. The PM may want to look further into these happy events to see what is going on, or if the monitoring system is reporting accurately.
CR (actual progress/scheduled progress) (budgeted cost/actual cost) Figure 7-8 Trend projection.
Table 7-3 Critical Ratio (CR) (actual progress/scheduled progress) (budgeted cost/actual cost)
Task Actual Scheduled Budgeted Actual Critical
Number Progress Progress Cost Cost Ratio
1 (2 / 3) (6 / 4) 1.0
2 (2 / 3) (6 / 6) .67
3 (3 / 3) (4 / 6) .67
4 (3 / 2) (6 / 6) 1.5
5 (3 / 3) (6 / 4) 1.5
7.5 DESIGNING THE CONTROL SYSTEM • 261
Beyond evaluating each of the activities of a project, a critical ratio for the project as a whole can be calculated as well, as illustrated in Table 7-4. Here, the project is as- sumed to consist of the five tasks in Table 7-3 and the values of each element of the critical ratio are summed on a daily basis with the critical ratio calculated for each day.
This ratio can then be itself tracked in a table (see Table 7-4 where the first day is from Table 7-3) and plotted on a control chart. The PM can also set some “control limits” for the critical ratio so that if they are exceeded on the upside or downside, an investiga- tion is in order, as shown in Figure 7-9. Different tasks, of course, may warrant different control limits. Further, the upside limits may be different, probably larger, than the downside (problem) limits.
Dealing with such differences is the purpose of another tool, the control chart. Any measure—the volume of raw material being used, the cost of contract labor in the proj- ect, the hours of computer time—can be plotted and tracked on a control chart such as that shown in Figure 7-10. As illustrated, control limits for intervention can be set by the PM and shown on the chart so that when a measure exceeds one of these limits, action is instigated.
Table 7-4 Monitoring the Critical Ratio
Actual Project Scheduled Budgeted Actual Critical
Day Progress Project Progress Project Cost Project Cost Ratio
July 17 (13 / 14) (28 / 26) 1.00
18 (18 / 17) (34 / 26) 1.15
Figure 7-9 Critical ratios with control limits.
Another recent development, benchmarking, can be a useful tool for a PM when designing a monitoring and control system (Christensen, 1994; Gupta and Graham, 1997; Ibbs and Kwak, 1998; Thamhain, 1996; Toney, 1997). The process here is to make comparisons to “best in class” practices across organizations, or divisions, or even departments within an organization. An example is a recent benchmarking study (Ibbs and Kandt, 1998) to generate data for a Project Management Maturity Model. The model measures project processes, tools, techniques, and practices across a range of industries, the various life-cycle phases, and the nine knowledge areas of PMBOK.
An example of an internal benchmarking study was Johnson Control’s bench- marking the project management procedures of their own highly successful product development project managers, to be used by less successful project managers. They identified four common sets of procedures these successful PMs all used. These proce- dures are now used to train new employees, standardize practices, create a common language, tie together company functions, and create a positive project management culture (Reith and Kandt, 1991). Benchmarking and the control chart are tools commonly used in quality management, a subject discussed in more detail in Chapter 8 of PMBOK.
Designing the project control system entails many issues but the major guiding objective should be to create a balanced system where the benefits obtained exceed the cost of control. The primary means to active control by the PM are process reviews, personnel reassignment, and resource allocation. Two types of control systems are useful for projects: go/no-go controls and post-controls.
Tools to aid the PM in control are variance analysis, trend projections, earned value analysis, critical ratios, control charts, and benchmarking.
Figure 7-10 Cost control chart.