If you ask people what makes a project successful, "a realistic schedule" usually tops the list. But ask them to be more specific, and several characteristics of a realistic schedule emerge. A realistic schedule:

Finally, a realistic schedule takes into consideration all the objectives of the project. For example, a schedule may be just right for the project team, but if it misses the customer's completion date by a mile, then it's clear that the whole project will need reassessment. Building a project plan that includes all the necessary parts and achieves a realistic balance between cost, scheduling, and quality requires a careful, step-by-step process.

Chapter 6 dealt with the first step in planning the project: the work breakdown structure. This chapter explains planning steps two through five. Each planning step relies on the one preceding it, and each adds a new element to the plan. You can use the diagram in Figure 6.1 (see page 127) as a guide in this chapter.

Here is a quick recap of the steps involved in planning a project. The first two actions prepare the groundwork for planning and so can be considered preplanning activities. The remaining five steps develop the detailed plan.

These steps generate all the information required to understand how a project will be executed. The steps are systematic, but they don't necessarily come up with the "right answer." It may take several iterations of these steps to find this answer, which is the optimal balance between cost, schedule, and quality. Chapter 9 will address the challenge of finding the optimal cost-schedule-quality equilibrium for a project, but right now, we will look at the step-by-step instructions for creating a project plan. Step one is the creation of a work breakdown structure, which was covered in the previous chapter. We start here with step two, which involves identifying the relationships between the different tasks.

The sequence in which detailed tasks—work packages—are performed is determined by the relationship between the tasks. To illustrate this point, consider the following five tasks from the landscaping project described earlier. These tasks constitute a subset of that project:

As the homeowner and the teenage children who will be working on this project contemplate these tasks, the question arises: What is the proper sequence? Any time a series of tasks is performed, there will be sequence constraints— that is, certain tasks that must be performed before others. Sequence constraints are governed by the relationships of different tasks. For instance, rocks, weeds, and other debris must be removed before the lawn seed can be planted. Performing these tasks in the reverse order doesn't make sense because the seed would be lost when the weeds were removed. Figure 7.1 shows both a predecessor table and a network diagram, two different ways of recording sequence constraints. A predecessor table is a common way to display task relationships. (In fact, this is exactly the way most project management software records the relationships.)

Notice that tasks 1 and 2 have no predecessors. Either one can be done first, or, if there are enough people, they could be done at the same time. Tasks that can be performed at the same time are known as concurrent tasks.

There are just two basic rules when graphing task relationships with a network diagram:

Figure 7.1. Network diagram developed from a predecessor table.

Milestones Are Useful Markers

In setting up the sequence of events, many project managers find it useful to mark significant events in the life of a project. These markers—called milestones—are often used in work breakdown structures and network diagrams (see Figure 7.2). Milestones have zero duration, so adding them to a project doesn't affect the schedule at all. There are three great reasons to use milestones:

Milestones are useful to show major progress points, but the real progress indicators remain the detailed work packages. Every work package has specific completion criteria and a tangible result—which is the ultimate progress indicator.

Figure 7.2. Milestones can show external dependencies.

Finish-to-Start Relationships

The finish-to-start relationship states that one task must be completed before its successor task can begin. The network diagrams in this chapter all follow this simple assumption because it is the most common, but there are other types of relationships. Tasks with start-to-start (SS) relationships allow the successor task to begin when the predecessor begins. Finish-to-finish (FF) tasks can start independently of each other, but the successor cannot finish until the predecessor finishes. Figure 7.3 shows the value of using these other types of task relationships.

In order to determine the cost and duration of an entire project, it's necessary to build a cost and schedule estimate for each work package; this is called bottom-up estimating. A lot of information is generated in the estimating process, so it is critical to record it in a systematic manner. (Table 7.1 shows the work package estimates for the home landscape project. Tables 7.2 to 7.5 illustrate some of the variables that affect work package estimates.)

The schedule estimate for a task measures the time from initiation to completion. This estimate is usually referred to as duration of a task. When building a schedule estimate, it's important to include all the time the task will span. For instance, it may take only 1 day to order materials, but if it takes 10 days for delivery, the total duration of the task will be 11 days. Similarly, while a certain decision might take only two hours to make, it might be more realistic to estimate duration at 5 days if the decision maker is likely to be busy at that time.

Cost estimates come from four sources:

Is it really necessary to concentrate on costs when trying to build a realistic schedule? It is, because each cost represents a resource constraint. Costs such as hiring subcontractors and purchasing materials will constrain the schedule. Later on, the schedule will be adjusted to account for these resource constraints (this is the fifth step of planning), but before adjusting the schedule we need to identify all the resource requirements, one work package at a time.

How the Amount of Labor Relates to Duration

How long it takes to perform a task usually depends on how many people are assigned to do it. When you estimate the duration of a task, you will normally figure in the amount of available labor. In the example in Table 7.3, when a training manager added one more technician to a task involving a software upgrade, this addition, plus longer working days, cut the duration of the task from five days to two.

The work package estimates for the home landscape project in Table 7.1 include an example of using extra people to reduce the duration of a task. When the project team (the homeowner and the family's three teenagers) estimated task 10, "Remove debris," they agreed it would be a long, unpleasant task if they worked on it alone. So they recruited some of the members of their local youth group to help them for a few days. With eight people working on the task, it can be accomplished in four days.

How Productivity Relates to Duration

When estimating the number of people needed for a task, you will need to consider their productivity. Adding people to simple tasks always reduces the duration. Productivity is said to be constant when the total labor hours do not change as the number of people assigned to the task changes. (The productivity of the technicians in Table 7.3 was constant; so was the "Remove debris" task in Table 7.1.)

However, in the case of tasks involving knowledge workers, adding more workers does not always result in greater productivity and a shorter duration of a task. For example, if two engineers are working on a complex problem, adding three more may actually slow the task and produce no recognizable change in product quality. The result is a measurable decrease in productivity because the cost for labor increased while the product stayed the same. (The estimate in Table 7.5 is another example of this factor.)

Another point to consider when measuring productivity is that people who spend all their time on one project are usually more productive than people who spread their time across multiple projects (again, see Table 7.5). The analyst estimated that the most productive way for her to work—requiring the fewest hours and fewest days—was to be able to devote her attention 100 percent to one project.

When calculating hours for your part-time project workers, it is usually unnecessary to figure out exactly what hour of which day they'll be working on your project. By assigning an availability assumption, such as "two hours per day," to all their tasks, the estimates for the duration of the task are proportionally increased. This approach allows each individual team member to decide for him- or herself when to actually work on the task. It doesn't matter whether they get started on it as soon as it's scheduled or work on it the last day. All you need to do to prepare a detailed plan is to give them a start time and a completion time.

As you can see, there is much to be considered when estimating the costs and duration of work packages. Chapter 8 will discuss the subject further, with many more tips and techniques.

Calculating a schedule may be one of the most well known, but unappreciated, of all project management techniques. It can be particularly tedious and time-consuming when done by hand for large projects. Yet it is the key to establishing realistic schedules and meeting them. (The tedium involved is a compelling reason to use project management software.)

Figure 7.4. Calculating a schedule.

As discussed earlier, the initial schedule is calculated by using the network diagram and the duration of each work package to determine the start and finish dates for each task, and for the entire project. Figure 7.4 shows how the network and task duration can work together to produce an initial schedule. Schedule calculation provides a set of detailed schedule data for every work package, as shown in the following:

Calculating the schedule to determine these four dates is a three-step process. Referring to Figure 7.4 will help to clarify this process.

Step One: Forward Pass

The forward pass will help you determine the early start (ES) and early finish (EF) for each task. It is so named because it involves working through a network diagram from start to finish. (The next step will involve the reverse—a backward pass.) In Figure 7.4, we follow a forward pass through the diagram, step by step.

Figure 7.5 shows another way of displaying this information. It's called a time-scaled network because it uses a time scale across the top and each task is laid out across the calendar. Notice that all the early start dates are the same in Figures 7.4 and 7.5.

Step Two: Backward Pass

The backward pass determines the late start and late finish dates. All of us have made this calculation hundreds of times—whenever we set an alarm clock. The goal of the backward pass is literally to work backward from the project finish date to determine how late any task can begin or end. The late start (LS) and late finish (LF) are calculated in Figure 7.4.

Step Three: Calculate Float

Some tasks have flexibility in when they can be performed in the schedule and others have no flexibility. The term for this schedule flexibility is float. (Another common term is slack.) Float is calculated by subtracting early start from late start. (How to calculate float is demonstrated graphically in Figures 7.4, 7.6 and 7.7.)

Figure 7.5. Time-scaled network. This contains the same network information as Figure 7.4, but the format is different.

When the initial schedule has been calculated, the project schedule begins to take shape. One of the key features of the initial schedule is the critical path. The term critical path is one of the most widely used—and most widely misunderstood—of all project management terms. But the concept is simple: The critical path is defined as all of the tasks with zero or negative float. When outlined on a network diagram, the critical path is the longest path through the network. (The critical path is boldly outlined in Figures 7.4 and 7.4.)

The tasks that have zero float must be completed at their early finish date or the project finish will be delayed. Making sure that all critical path tasks begin and end on time is the surest way of making the project end on time. That's why you'll hear a project manager motivating someone to complete a task by telling the person, "It's on the critical path!"

Since it is the longest path through the network (the longest path means the longest duration, not necessarily the most tasks), the critical path is one measure of schedule viability. This is because it demonstrates the minimum time the project will take. Sometimes it takes a network diagram with the critical path outlined to show stakeholders that their optimistic schedule estimate is unrealistic.

These are just two of the ways that the project manager might use the critical path. Several more are described later in this chapter and in Chapters 9 and 13.

Figure 7.6. Home landscape project network with initial schedule data.

Figure 7.7. Managing float.

When all the float has been used up, a new term emerges to describe the situation: negative float. Negative float results when externally imposed finish dates are impossible to meet (such as 10 presents to buy at 6 P.M. on December 24). Figure 7.8, for example, shows a network diagram with a critical path that is longer than the allotted schedule. When there is negative float, it means that adjustments will have to be made to bring the schedule in line with the critical path. This is the kind of information you will need when you renegotiate the cost-schedule-quality equilibrium.

A picture is worth a thousand words. The network diagram is essential in calculating the schedule, but it can be terribly difficult to decipher on a large project. Thankfully, there are two very good alternatives, which display both the schedule information and the task relationships.

Gantt charts, named after Henry Gantt, who developed them in the early 1900s, have become the most common method for displaying a project schedule. Figure 7.9 is a Gantt chart for the home landscape project. It has the same schedule dates as the network in Figure 7.6. Notice that all the tasks are currently scheduled at their early start date—you can tell that because all noncritical tasks display float. The great advantage of the Gantt chart is its clarity: The horizontal axis shows the schedule and the vertical axis lists the work breakdown structure.

Another excellent graphic for displaying a schedule is the time-scaled network (as shown in Figures 7.5 and 7.5). One advantage that this diagram has, compared to the Gantt chart, is the ability to condense the network onto less paper. On large projects, the Gantt charts can grow too large to print, whereas the time-scaled network, because it combines many tasks on one line, can be made one-half to one-tenth the height of the Gantt.

The initial schedule represents the combination of task sequence and task duration, but it's called an initial schedule because it hasn't taken into account people and equipment limitations. The next planning step uses the initial schedule as a starting point and balances it against the resources available to the project.

Figure 7.8. Negative float. When imposed deadlines result in negative float, that is a warning the project is out of equilibrium the cost, schedule, or quality objectives must be revised.

Figure 7.9. Gantt chart for home landscape project.

Figure 7.10. Time-scaled network for home landscape project. This time-scaled network has the same initial schedule information as Figure 7.9.

The goal of resource leveling is to optimize the use of people and equipment assigned to the project. It begins with the assumption that, whenever possible, it is most productive to have consistent, continuous use of the fewest resources possible. In other words, it seeks to avoid repeatedly adding and removing resources time and again throughout the project. Resource leveling is the last step in creating a realistic schedule. It confronts the reality of limited people and equipment and adjusts the schedule to compensate.

Using the home landscape project as an example, we can see how resource leveling makes a project schedule more realistic. The network (Figure 7.6) shows, in terms of task scheduling, that it's possible to put in the lawn and build the fence at the same time. But when we consider that the family has only the three teenagers available to work on the project, that means they have just a total of 24 labor hours available per day (3 teens × 8 hours a day). Trying to put in the lawn and build the fence concurrently is unrealistic because it would require each teen to work far more than eight hours a day for more than half the project. (The resource spreadsheet on the Gantt chart in Figure 7.11 indicates clearly how unrealistic the schedule is.) Resource leveling will adjust the schedule to keep the teens busy at a consistent, reasonable rate. (Figure 7.12 on page 158 shows the same project as Figure 7.11, but with the resources leveled.) Not only does resource leveling take unreasonable overtime out of their project, but it keeps the teens employed for a longer time at a steady rate. That's usually an advantage for any project team.

Let's consider a few of the problems faced by project managers in this process of leveling resources.

Every project faces the reality of limited people and equipment. The idea is to avoid both over- and underallocation. As the home landscape project demonstrates, too many concurrent tasks can call for more resources than are available. For example, as discussed, the initial schedule had the teens working on the fence and the lawn during the same period, and this resulted in the teens being overallocated during the first half of the project (they would have had to work more than eight hours a day to meet this schedule).

Project managers need to remember that whether it's teenagers planting the lawn, bulldozers, or programmers, there are rarely a bunch of spares sitting on the shelf. This overallocation problem can become especially acute if project managers imagine that they have a large supply of a rare resource, such as the unlimited time of the only subject expert in the company. In this case, not only has the schedule become unrealistic, but the manager may have overloaded a key resource.

Figure 7.11. Gantt chart with resource spreadsheet for home landscape project.

Figure 7.12. Gantt chart with resource-leveled schedule for home landscape project.

The other side of the problem is underallocation. If the project team isn't busy on your project, it will likely be reassigned to other projects and be unavailable when the next peak comes. In the worst case, during lulls in the project some of the unassigned people may get laid off, becoming permanently unavailable and taking valuable knowledge about your project with them.

A further problem arises if people working on this project are also working on several others at the same time. If every project in the firm has wild swings in its resource requirements, it is almost impossible to move people smoothly between projects. Instead, people are yanked off one project to help another catch up, only to be thrown at yet another that is even further behind.

The Process of Resource Leveling

It is important to remember how we are defining the term resources. Resources are the people, equipment, and raw materials that go into the project. Resource leveling focuses only on the people and equipment; the materials needed for the project are dictated by the specifications.

Resource leveling begins with the initial schedule and work package resource requirements (see Table 7.1). The leveling follows a four-step process:

Reestimating work packages and delaying tasks within their float can remove the worst resource peaks and valleys, but the plan might still contain unrealistic or uneconomic resource peaks. At that point, your next option is accepting a later project completion date.

On the home landscape project, the schedule had to be extended by two weeks to balance the available labor with the amount of work. This can be a difficult decision, but that's what it means to create a realistic cost-schedule-quality equilibrium. In Chapter 9, there will be further discussion of the factors that balance the schedule against cost and quality.

The detailed planning techniques presented in this chapter are often downplayed as mere scheduling techniques. But not only do they build a realistic schedule, these techniques form the foundation of the science of project management.

The planning model in this chapter portrays the step-by-step method of building a detailed plan. Laying out a detailed plan may not solve all the problems you'll face as a project manager, but it does provide a tool set for solving many of them. Every one of these planning techniques will be referred to again in later chapters. They are the entry point to the science of project management; without them it's impossible to practice the art.

Large or complex projects can follow the advice in Part 3 together with project management software to build detailed action plans. Simpler projects may call for a simpler format.

The action plan template is simply a spreadsheet formatted with standard headings for recording the basic information of a plan and tracking actual performance. Fill in the calendar portion by shading the cells to create a Gantt chart. This template is available for download at www.versatilecompany.com/FFMBAinPM.

Answers to these questions are found at www.versatilecompany.com/FFMBAinPM.