Critical Path Analysis:
A Planning/Time Management Tool for Managing Research
By Roger Hiemstra
This paper urges use of Critical Path Analysis (CPA) in formulating, scheduling, and managing the various milestones or activities involved in a research or any other type of project that requires advanced and careful planning. CPA involves the careful delineation and sequencing of such activities, determining the time necessary to complete them, and computing a critical path of activities around which the project must be managed for success. As a researcher or planner you are shown how to carry out each step in the CPA process through a series of visuals and corresponding narration. The calendar dating of a CPA network, utilization of Gantt charts, and other time management ideas are included.
What is Critical Path Analysis?
Critical Path Analysis can be defined as the logical sequencing of a series of events necessary for a successful research project in such a manner that the most efficient route to some culmination point can be calculated. Consequently, the critical path technique has a multitude of uses: (a) As an aid in time management, (b) as a provider of ongoing data for assessing progress, and (c) to give the researcher or program planner considerable information for decision-making.
CPA is a useful planning and management tool for several reasons:
Carrying out the above steps provides you with an empirically constructed schedule of research or planning activities to meet some self-imposed or institutionally required calendar of events. CPA also provides you with a visual picture through the development of a model network displaying how all research or planning functions tie together. The availability of computer software related to CPA techniques even means that information regarding progress in the planning and implementation process is available as often as you input related data.
How CPA is Utilized
Learning how to utilize CPA means understanding a few basic principles, procedures, and definitions. The terminology and process to be described in the following narration and visuals are somewhat unique to the CPA technique. Once these procedures are mastered, you will find them applicable to any type of planning or time management need.
Step One: List the Major Activities
The procedure begins by listing all the known or expected major activities to be undertaken in planning and managing a research or other type of project. The level of detail or finiteness of the activities will depend on your determination of their importance or "criticalness" to completing your project planning and management according to various time constraints. Over time in using CPA, whether by hand or by computer, you will "learn" how detailed you need to be for success in using the technique as a time management tool.
Listing the major activities requires a thorough analysis of the steps involved in initiating, implementing, and bringing to culmination a successful research or planning endeavor. Several activities are described below. The examples used relate to planning a conference or meeting to provide a common reference point for those not yet familiar with necessary research steps. It should be obvious there could be many variations or each activity could be divided into several smaller activities for the examples shown:
Step Two: Sequence the Major Activities
The next step is to sequence each activity into a logical order based on when it must or is most likely to occur. The sequence you select will be peculiar to your research or planning project, but certain activities logically precede others (for example, in planning some type of meeting formulating objectives almost always takes place before you obtain the meeting speakers). This step usually requires some reshuffling of activities before the final sequence is determined. Experience with planning and the CPA technique will facilitate this sequencing step.
Step Three: Construct an Activity Flow Diagram
In the third step you construct a schematic or flow diagram detailing the sequence of activities determined for an overall planning effort or research project. This step requires an understanding of three basic definitions:
Figure 1 illustrates a simulated CPA network. It is important to reemphasize that events (typically represented by circles or squares in a network) do not consume time or resources, but merely are beginning or ending points of actual planning and management activities. They are usually identified by words within the circles or the circles are numbered and defined elsewhere. Activities are the arrows connecting various events. They are typically represented by capital letters shown on or near the arrows. Each activity can be defined in a footnote to the network or on a separate page if necessary for clarity's sake or if the network is to be part of a formal meeting or research plan (e.g., activity "A" could be defined as formulating the meeting objectives). Arrows made up of dashes or arrows identified with the word "Dummy" represent dummy activities, those activities requiring no time or resources.
Arrowheads indicate the flow direction (e.g., "finalizing the meeting site arrangements" must take place before "advertizing and marketing the meeting"). More than one arrow leading to or from an event indicates that more than one activity is occurring at the same time. In addition, it should be noted that the length of an arrow has no relationship to the amount of time; arrows merely serve as connectors between events and show the direction of the flow. See the following footnote for explanatory information on a short cut method of displaying activity paths by "assuming away" dummy activity arrows.
Step Four: Estimating the Time for an Activity
The next step involves estimating the time required to perform each activity. Time is usually represented on most networks in days or in weeks. In the CPA process it is recommended you estimate the time required for each activity utilizing as guides such criteria as experience, suggested time requirements made by colleagues or planning authorities, and known limitations of staff, resources, and money. A thoroughly constructed network might even have defined how many people and who will be working on a given calendar day, although there will be occasions when approximations will be just as useful.
Figure 2 displays the previously constructed network with time estimates included. They typically are displayed as numbers on an activity arrow. The time estimates are simulated for this example and they will always be specific to any single planning project, although similarities will occur from project to project. They are used here primarily to facilitate an explanation of the CPA technique.
The simulated network utilizes 8-hour working days and a seven day week, assuming that weekends and overtime costs need to be computed if a dollar figure is attached to the network and any corresponding time requirements of salaried personnel are considered in overall management of whatever project being considered. In addition, concurrent activities will require consideration of time needs in relation to the availability of total staff and resources.
Step Five: Computing the Expected Times to Complete All Activities
Following an estimation or determination of the time required for each activity, the planner needs to calculate totals or the earliest possible time that each event can be reached. Symbolized on Figure 3 by the letter "T," this total is derived by summing the estimated times of all activity paths leading to any particular event.
As Figure 3 shows, times have been calculated for all events in the example. They are shown at the top of the corresponding event.
Step Six: Determining the Critical Path
A critical path, or that path of activities that must be completed within the times shown if the entire plan is to stay on schedule, can now be determined. Symbolized on the network by small perpendicular lines or hash marks, the critical path is found by selecting those activity paths requiring the most time to arrive at any one event. In other words, to arrive at the "Conference Room Needs Determined" event from the "Conference Site Finalized" event, it requires more time to complete activity D (see Figure 1), then activity G, and then activity I, than to complete either of the other two paths. More discussion of the critical path and its usefulness is in a later section.
Two concepts must be understood before this step can be achieved successfully. First dummy activities (represented on the network simulated in this paper as lines that contain dashes) to not consume time. They merely are logical connectors between two events. In other words, the meeting planner cannot declare the overall meeting completed until the final report is written, but immediately upon that task the meeting can be declared as completed. As a further illustration, a dummy arrow might have been drawn from the "Meeting Site Finalized" event to the "Meeting Program Developed" event. In this case, that event would not take place until all activities related to both formulating objectives and finalizing the meeting site had been completed. Thus, when summing up times to a certain event, the dummy activity times are assumed to be zero.
Secondly, time totals up to any event are based on the earliest possible time that event can be reached. Thus, as all activities leading to an event (more than one arrow) must be completed before it is considered reached, the activity path actually taking the most amount of time to an event is determined by adding all activities on the critical path. For example, as shown in Figure 3 finalizing the meeting site, developing the meeting program, arranging for food and hotel, determining the meeting room needs, finalizing the meeting booklet, managing the logistics, tabulating the evaluation results, and writing the final report represent the longest path, the earliest possible time, and the critical path. No other activity paths would take longer; thus the "conference completed" event is reached in the earliest possible time of 60 weeks.
Step Seven: Constructing a Gantt Chart
Once all of the totals have been reached, the planner can make the first application of the CPA process. Simply by starting at the first event with some selected or mandatory calendar date, completion dates at each even based on the corresponding totals can be established. Figure 4 shows the simulated network with appropriate calendar dates superimposed based on a July 1 starting date. A critical path also is superimposed (represented on the network by double arrows) to provide you with a visible reminder of what is required to keep on schedule (as will Gantt charts--see Figure 5).
It also has been the experience of many planners that a corresponding "Gantt" chart (see Figure 5) can be useful in quickly and constantly assessing a project's progress. (Henry Gantt, a management pioneer, introduced the bar graph concept many years ago. It is simply another way of displaying graphically the activities that comprise a planning or research project.) This bar graph approach reveals how many activities are taking place at once, when one activity must begin in relation to another, and how long various activities are in progress. In essence, you take each activity on the network, stretch the length to the appropriate calendar length, and place it on the graph. You can even color code the critical path as a quick visual reminder of what is crucial for staying on time. Appropriate decisions to add resources or to have staff members work on other activities can be made from these types of information if certain activities crucial to keeping on schedule begin taking longer than expected. Most computerized programs are capable of making Gantt charts.
CPA and Decision-Making
The ability to make appropriate planning and managing decisions based on the CPA process, in addition to its scheduling usefulness, can be enhanced by adding two final steps to the process. These steps are crucial if you find that the initially constructed network requires more days or weeks than you had originally estimated or if the final completion date must be moved up for some reason. Information is available that allows you to make the most sound decisions regarding where to apply additional resources in order to meet such needs.
The first of these additional steps begins by establishing what is that final finishing date or allowable completion time. In Figure 6 it has been assumed that 60 weeks is an acceptable time frame and that it will allow you to run the meeting as scheduled. Then "L" (latest allowable time) is set equal to 60 weeks at the "meeting completed" event. On the network all "L" times are shown at the bottom of the corresponding event (Figure 7 converts these to slack times)
To calculate the "L" times for all events on the network, you must work backwards starting from the "meeting completed" event, subtracting the appropriate activity times. For example, as the meeting completion activity (closing accounts, storing evaluation data, writing the final report, etc.) was approximated to require three weeks, the latest allowable time at the "evaluation results tabulated" activity is 57 weeks. This process is repeated throughout the network.
It needs to be emphasized that when more than one activity path leads backward from an event, the activity path requiring the most time to complete is the one whose time is subtracted from the "L" time up to that point. This is demonstrated on the network, for example, in calculating the latest allowable time for the "food and hotel arrangements made" event. Although the smallest "L" times are the dummy activity from "meeting program booklets finalized" back to "displays and exhibits arranged" and activity H, it will require considerably more time to complete activities K and I, the activities leading backwards from "meeting program booklets finalized" through "conference room needs determined" to "food and hotel arrangements made." The corresponding subtractions (54 - 0 - 8 = 46 -vs- 54 - 4 - 10 = 40) show that the smallest "L" time is equal to 40. Following this procedure the appropriate "L" times at each even can be calculated.
The next and final step begins by determining a "slack" time at each event. Slack time is defined as the amount of excess time available to reach any particular event. Another way of looking at this is that slack is the number of weeks the latest allowable time (L) exceeds the earliest possible time (T) an event can be reached. It is determined by subtracting the "T" time from the "L" time at each event. Figure 7 details the network with all the slack (S) times shown above the corresponding event. This should not be confused with the various "T" times previously displayed above the corresponding events in Figure 3. If you were to have all the various times calculated via a computerized program, the results typically are displayed in columns under appropriate headings rather than superimposing them all on a single, and therefore most likely, congested looking network.
In explanation of the slack times, a value greater than zero indicates that there are that many weeks of spare time in terms of completing that event. The "slack" time concept also can be used advantageously when the required completion or contract time is equivalent to that determined in the original calculations. Those events with the least amount of slack (usually zero if the "L" times and "T" times are set as equivalent to each other--there are times when the number could be minus if the "L" time was required by law or some other reason to be set at a certain date and that date was prior to the calculated "T" time) will always lie on the critical path, thus facilitating the path's determination.
In addition, by setting the final event's "L" time equal to the final "T" time and computing all the slack times, the planner can identify those events where excess amounts of slack time exist. Thus, if at any point the overall plan falls behind schedule, the planner can divert staff time from activities not on the critical path to those more critical activities and often thereby restore the plan to its original schedule. Thus, in the simulated example if the overall project fell behind schedule before the meeting room needs had been determined, those people working on such activities as marketing and advertising or obtaining the meeting speakers could be asked to pitch in for awhile on something like developing the meeting program or making food and hotel arrangements.
Of course, once any event either is reached early or late, the remaining network must be recalculated with new calendar dates. For example completing any one of the activities on the critical path might be shortened by one or more weeks by such actions as asking people to work overtime, jobbing something out to another organization, or hiring temporary people. Note, too, that the critical path always passes through those events showing the least amount of slack, although there sometimes can be more than one critical path when the slack times are equivalent on two different paths leading to a certain event or the final completion event.
This type of information should aid you in making a variety of decisions. If the research or planning project is faced in the beginning with negative slack on the critical path because of some forced contractual obligation (such as a college requiring the completion of a thesis by a certain date in order to graduate), then you will know immediately that some changes will be required in order to meet the required schedule. You also will know where the most critical of these changes are needed. Available staff or resources from those events with surplus slack time might be shifted to more critical activities, new resources or staff assistance may be required to speed up certain activities, or you may need to decide to eliminate or greatly reduce certain activities.
The CPA procedure should facilitate most planning and research projects. It forces you as planner to organize and sequence the required activities carefully. Information also is provided that will enable a calendar of meeting events to be constructed. The Gantt chart, especially if reproduced in a very large format or displayed electronically for those used to computer networks, can serve as a major discussion point at staff meeting for determining progress or staff assignments. Finally, for the individual research or meeting planner the CPA network or Gantt chart can serve as a daily planner or self-disciplining tool to facilitate day to day meeting management actions.