CHAPTER 4
Planning a Single Project
Patrick is responsible for publishing a monthly magazine for industry professionals. The magazine contains articles on new developments by industry companies. Freelance journalists write articles and supply photographs. In addition, the president writes a monthly column (which is always late), and the executive director writes another. Every issue requires a cover and cover story that focuses on one key person in the industry. The magazine also contains advertising from member companies and by organizations that supply products and services to the industry. The back of the magazine contains regular display ads and a classified section.
The deadline for each issue allows little flexibility. To reach members during the first full week of a month, the issue has to be mailed the last week of the previous month. Printing takes six days. Therefore, the May issue must be at the printer the last full week of April. All editorial, design, and production work must be completed, all approvals from management obtained, and the complete package turned over to the printer by that date.
That is one issue of one magazine.
First plan this single project, then learn how to combine it with others to make a multiple project schedule.
DEFINITIONS
The triple constraints are the three components of a project definition: how much time you have, how much money or resources you can use, and what the project must accomplish. Usually these three parts are called time constraint, budget constraint, and performance criteria.
The driver is the most important of the project’s triple constraints. The other two constraints are ranked as middle and weak.
The work breakdown structure (WBS) is a method for breaking your project into component tasks and organizing the management structure.
Network planning is the process of putting the project’s tasks in the order they need to be done. You may put your tasks in a dependent or parallel sequence. The network diagram is also called a PERT chart or CPM chart.
Task analysis is the process of gathering together all the information you need to manage each task in your project. It is a great aid in delegating.
A Gantt chart is a timeline chart that shows the sequence of tasks in your project over a calendar period.
THE TRIPLE CONSTRAINTS
Baseball great Yogi Berra once observed, “If you don’t know where you’re going, you might just end up somewhere else.”
The first stage of managing a project is determining the project goal. That tells you what you have to do, the conditions under which you have to do it, and (properly done) the priority relative to other projects and tasks in your schedule.
The first thing to remember about a project is that it is not an end in itself, but rather a means to an end. You do a project so that you can get the result. If you manage the wrong project—that is, if you get a different result—the project is a failure, even if you worked hard and followed good management procedures.
Sometimes it is challenging to pin down a project objective.
Every project begins with an underlying need. There is a problem that needs to be solved, or an opportunity that is available. You need to know not only what the project is, but also why you are doing it.
In Patrick’s case, he needs to understand why the association is publishing this magazine. It could be any combination of the following, to:
earn a profit for the association
provide a public relations vehicle for member companies
transmit industry information
serve as a technical journal to help raise quality standards and solve complex problems.
These objectives are not mutually exclusive, of course. Likely, these reasons can be ranked in some sort of priority order. Is it more important to show a profit, or is it more important to provide quality technical information needed for industry engineers to develop new processes? Is it more important to provide public relations for member companies or to tabulate good industry information?
Understanding the reasons for the magazine will help set editorial policy, determine how to spend money, and affect the design and production standards, even the choice of paper stock.
From the basic reasons, you must define three elements of a good project, that is, the triple constraints:
how much time is available (called the time constraint)
how much money or how many resources: people, tools, equipment you can use (called the budget constraint)
what the end result must achieve (called the performance criteria).
Put them all together, and you have the triple constraints (see Figure 4.1).
You cannot manage a project successfully unless you know the triple constraints. They shape the world of the project.
Take the case of Noah’s ark. When God commanded Noah to build the ark, he set performance criteria:
Make a boat from resinous wood, sealing it with tar, and construct decks and stalls throughout the ship. Make it 450 feet long, seventy-five feet wide, and forty-five feet high. Construct a skylight all the way around the ship, eighteen inches below the roof, and make three decks inside the boat—a bottom, middle, and upper deck—and put a door in the side.
All these performance criteria relate to the goal: to accommodate two of every animal and keep them safe from the flood. The deadline, of course, is that the ark had better be built before the flood. Budget? Not specified, but notice that the available resources are primarily the work effort of Noah and his sons.
One consequence of the triple constraints is that they are never ranked equally. The most important triple constraint is the driver. If you fail to achieve the driver, the project is always a failure. The next is the middle constraint. It is at least slightly less controlling than the driver; it may have more flexibility in its interpretation. The remaining one, the weak constraint, has the most flexibility.
In the case of Noah’s ark, budget would have to be the weak constraint, since it is clearly most flexible. The question of whether time or performance is the driver would depend on whether God has fixed an absolute date when the flood will start (that would make time the driver), or whether God will hold off the flood until Noah finishes the ark (performance would become the driver).
The choice of whether the driver is time or performance in this case would strongly affect how Noah approaches the work. If the deadline is absolute, and performance is slipping, perhaps he can trim some of the optional details, or finish up the stalls after the ark is floating. On the other hand, if performance is the driver, Noah can take some extra time, if needed, to make sure everything is shipshape. In both cases, the best solution would be to spend more money or more resources. Once the flood starts, leftover money will not be of much use.
The savvy project manager needs to know what the constraints are and understand their ranking. When things go well, these are not at issue. When things go wrong, you need to be able to use the flexibility of the weak and sometimes middle constraints to make sure you meet the driver, no matter what comes.
In the case of Patrick’s magazine, assume that it is a general-interest magazine published first as a profit center for the association and second as a public-relations vehicle for the industry. The monthly deadline forms the time constraint, which is also the driver, since missing the deadline destroys the value of having a monthly magazine. The middle constraint is budget since the magazine needs to make a profit. While good public relations is important, it is also flexible in its definition. How many pages does each issue need? How much text is in each issue? These are subjective criteria and are therefore the most flexible. Performance criteria in this example are the weak constraint.
WORK BREAKDOWN STRUCTURE
The second step in managing a single project is creating the WBS, which is the equivalent of creating an organization chart for your project.
Step 1: Brainstorm tasks. List all the tasks that need to be completed to make your project a success.
Step 2: Group tasks into logical categories. You may use a variety of approaches to identify these categories, including departmental, skill, or phase. The goal of grouping the tasks is to improve management control.
Step 3: Display the tasks in an organizational chart. Figure 4.2 shows an example of the final product.
Tip! Use sticky notes to brainstorm the tasks. You will find it easier to group them together and create your WBS chart. Sticky notes are also valuable to the next step of managing your project.
NETWORK PLANNING
Now that you have a WBS, you know the tasks in your project, and you know which part of the project team is responsible for each task. Next, you have to put the tasks into a project sequence. This stage is called network planning. You will hear the final result variously called a PERT chart or a CPM chart.
Program Evaluation and Review Technique and Critical Path Method
The program evaluation and review technique (PERT) and critical path method (CPM) are two powerful project management tools, both invented in the 1950s. PERT was developed by the United States Navy for managing the Polaris project; DuPont and Remington Rand developed CPM. While both were originally distinct approaches, they had strong similarities. In the real world of project management, people immediately began choosing those parts of each system they liked best, with the result that most people who use these systems today use a hybrid. For the purposes of everyday project management, PERT and CPM are effectively synonyms.
In this book, this kind of charting is called either PERT or network diagram.
Creating a Network Diagram
To create a network diagram (see Figure 4.3) or PERT chart of your project, take the sticky notes from the WBS and put them in the most logical and sensible order.
Tasks can be organized either in a dependent or parallel manner. A dependent task is one that must wait for a predecessor task to complete. For example, in the diagram, Storyboard issue is dependent on Select manuscripts and photographs. You cannot start storyboarding (determining what goes on each page) until you know what the contents of the issue will be.
A parallel task is a task that can be performed in the same time frame as other tasks. For example, while you are selecting manuscripts and photographs and storyboarding the issue, you can Obtain editorials in parallel to those tasks. (This means pestering the president and executive director to turn over their manuscript copies. You already know that the editorials will be one page each in the magazine—that is a standard—so you can storyboard them even though you do not have the articles.)
The advantage of parallel tasking is that it saves time. Sometimes, parallel tasking increases resources needed, and sometimes it increases risk. You have to make strategic choices because there is more than one acceptable way to lay out a project.
Something very interesting happens when you make tasks parallel. Look at a close-up section of our network diagram in Figure 4.4.
Figure 4.4 shows an estimated time for each task, which is the next step you would follow in preparing your own network diagram. Notice that Start issue has a time of zero days; this is called a milestone in project management terms. Milestones identify major turning points in projects; they do not represent actual work to be done.
In this section of the project, you will see that all the tasks total sixteen days of work. Because of the parallel structure, this part of the project will take only twelve days to do—the longest path from start to copyedit. The longest path in a network diagram is very important; it is called the critical path, and a task on the critical path is called a critical task.
First, imagine what happens if either select manuscripts or storyboard goes over the time estimate. Immediately the bumper-car effect pushes out the start of copyedit. You just made the longest path longer, which makes the project longer.
Second, imagine what happens if the president is a day late with the editorial for this issue. That’s right: nothing. Since the critical path segment between Start Issue and Copyedit manuscripts totals seven days (5 + 2), and Obtain editorials takes four days, Obtain editorials can be up to three days late (7 − 4) without impacting the start of the next task! Because Obtain editorials is not on the critical path, it is called a noncritical task.
Noncritical tasks are those that can be late without instantly bumping the start of the next critical task. The extra time available, called slack or float, is the difference between the critical path total and the noncritical path total.
Slack is your friend. Slack is an important tool in planning your project. Even if time is the driver, tasks with slack can be slipped in the schedule without harm to the ultimate deadline. In addition, if the slack task is done by a resource (person, equipment, tool), then slack resource availability may occur, which means that you have some resources not scheduled in that period. You can hold them back for emergencies, use them to get a head start on other work, or assign them to your other projects. (We will look at resource scheduling in much more detail later.)
TASK ANALYSIS
You can use forms to record and track the various pieces of information you need to manage your projects. A good task analysis worksheet (see Figure 4.5) lets you keep in a single place all the information you need to manage it. (It is a great help in delegating, too.)
While the form in Figure 4.5 is similar to the project control worksheet in Figure 2.1, the one depicted in Figure 4.5 is more detailed. The task analysis worksheet works better on larger projects with more complex tasks. Feel free to edit the format, based on issues and concerns specific to your own projects.
Tip! In managing multiple projects, you may find that many involve similar subject matter. When possible, reuse your task analysis forms to simplify planning and organization.
Here is a breakdown of each component of the task analysis form.
Task Number: Number your tasks (1, 2, 3, or A, B, C)—a common shortcut. If you use a project management software package, your program will automatically number your tasks.
Task Name: Give each task a distinct name.
Predecessor/Successor Task(s): From your network diagram, write down the tasks immediately prior to the current task (predecessors) and the tasks immediately following the current task (successors).
Specifications/Deliverables: Define what the task must accomplish to be considered complete.
Resources: Include people as well as tools/equipment.
Time Estimate: For a simple project, fill in only the most likely item in this block. More complex projects may require more detailed estimating. For example, normal tasks are assumed to begin when their predecessors are complete. If the predecessor finishes early, the task starts early; if the predecessor is late, it starts late. However, some tasks have calendar date issues. Our magazine must be at the printer on a certain date, so the final galley proofs must be ready then—no matter what!
Milestones show the completion of parts of the task. In design/produce cover, one milestone might be cover photo selected, followed by pencil rough of cover completed, and finally cover master artwork produced. Use milestones for finer control of what is going on in the middle of a task.
Cost Estimate: Figure costs by category. Every organization has its own internal financial procedures. Your project may or may not be charged certain costs, from staff salaries to equipment and supplies, including percentages for overhead and general administrative expenses.
Tip! Filling out forms requires a lot of work. Here are two shortcuts. First, you may delegate completing the form to team members and subordinate project managers. Second, you may be able to reuse (with some editing) forms you have already designed.
Estimating Techniques
How long will each task take? On some projects, it is easy to predict how long a task will take. If you are going to send somebody to a three-day training course, you expect it is going to take three days, not two and not four. On the other hand, if the task is to create a new idea, it is hard to predict accurately how long it will take. Try some of these techniques to improve your estimating skills.
Six Steps to Exceptional Estimating
1. Use actual time estimates from similar tasks in other projects. Good estimates are based on historical data; use the best historical data you can find.
2. Involve the people most knowledgeable, and ask for their estimates. Be sure to learn their tendencies. Are they optimists or pessimists? (Some people claim to be realists. They are usually pessimists.)
3. Research the history of vendors and subcontractors you plan to use. How experienced are they? How good have their estimates been in the past? Get references, and check them out.
4. Consult standard reference books containing time estimates. From the standard garage manual that estimates the time a mechanic should take to fix your car to software industry norms for the development of code, many references exist for specific work. Check your trade or professional organization for details.
5. Watch for tasks with outcome and timing that directly affect the project final outcome. When possible, build in extra time and resources to allow slack for those tasks.
6. Train yourself to be an exceptional estimator. Whenever you need an estimate, make a guess and write it down. Then make an estimate. When you finish the actual work, compare the result to your guess. By giving your brain regular feedback, you will find yourself getting progressively accurate.
PERT Time Estimating
One of the more challenging parts of estimating involves uncertainty. How long a job will take can depend on unpredictable factors. For example, if a construction company cannot build in the rain, the number of rainy days in the construction season affects how long the job will take.
The creators of the PERT project management system used a statistical approach to uncertainty. For each task, they created three estimates:
the optimistic (best case) situation, called T(o)
the pessimistic (worst case) situation, called T(p)
the most likely situation, called T(m).
Even though it is impossible to predict what a given task will take, if you have a large enough project, you can reasonably expect some good luck, some bad luck, and a great deal of average luck. Therefore, you can create a weighted average (called T(e), short for time estimate) of the three times by using the formula:
T(e) = T(o) + (4 * T(m)) + T(p)
Note: For an explanation of the full statistical process, consult Practical Project Management by Michael Dobson.
If you calculate T(e) for all tasks in your project, and use those tasks in making up your schedule, you will find that although individual tasks may run over or under the estimated time, your actual schedule should come out nearly on time.
How close? There is a statistical answer to that, and it has to do with a concept called standard deviation, abbreviated with the Greek letter σ or sigma. You determine the sigma of a task with the following formula:
According to standard probability theory:
68.26 percent of the time the work will be completed ± 1 s of T(e)
95.44% of the time the work will be completed ± 2 s of T(e)
99.73% of the time the work will be completed ± 3 s of T(e).
That is the process for one task. There is one catch. If you are trying to calculate sigma for a sequence of tasks, take each task sigma, square it, add the squares together, and then take the square root of the sum.
You use the standard deviation concept in project management to determine the risk level and additional time you require. The ideal project has a safety margin of two sigmas of the critical (longest) path, because it then has a 95 percent probability of being finished on time, even if things go wrong. If you cannot add that much safety margin (and often you cannot), you can at least add as much as is available. With a standard statistical table, you can figure the percentage chance you have of finishing the project on time. This is called the confidence level.
Not all project managers use this technique. Utilize it only if your projects have high uncertainty factors and you need extra time to meet your deadline. It is simpler to use than it may appear. Once you put the formulas into a spreadsheet program, the process of calculating becomes quite speedy.
Critical Path Method Time Estimating
Another approach to time estimating evolves from the idea that, quite often, if one person can do the job in three weeks, two people may be able to do the job faster.
Speeding up a task by giving it extra resources—people, tools, equipment, or money—is called crashing the task. Not all tasks can be crashed. If you are trying to get the editorial copy from the president, two people will likely not have any better luck than one person.
The process of crash analysis is figuring your crash options and determining whether or not they are worthwhile (see Figure 4.6).
Imagine the following situation:
One person can do the job in three weeks at a salary of $1,000 per week.
Normal Time = 3 weeks Normal Cost = $3,000
Two people can do the job in two weeks at a salary of $1,000 per week each.
Crash Time = 2 weeks Crash Cost = $4,000
The crash slope is the difference between the normal and crash situations.
Crash Slope = (–1 week, +$1,000)
In other words, you save a week and spend an extra $1,000. Now the question becomes: “Is it worth it?”
To determine whether a task is worth crashing, first see if the task is on the critical path. If it is not on the critical path, saving time on the task will not save time on the project. It will just create more slack. (On the other hand, if the noncritical task is very risky, you might put the extra resources there to make the project safer.)
If the task is on the critical path, now you have to determine the value to the project being a week early. Is finishing early worth $1,000? Consider not only the value of turning over the project results but also the fact that you will no longer be spending money, people, and resources in the week you saved.
Sometimes it is clear that crashing is worthwhile; if this is the case, do it. Sometimes it is clear that crashing is a very bad idea; if this is the case, do not do it. Sometimes the situation is difficult to call. If so, try postponing the decision until later in your project. You might discover that changing situations in the project may tip the crash decision one way or another. If you are running late and racking up overtime, the value of crashing may be positive. But if everything is going great, the cost of crashing may not be worthwhile.
The only other tip on crash time is to remember that if you promise your boss you will save a week by spending the $1,000, save the week!
Task Table
The final element you need to put together a project schedule is called a task table (see Figure 4.7). Fortunately, everything you need to put a table together follows the previous process. You simply organize a table that contains:
a list of tasks (from your sticky-note PERT chart)
an estimate of the time each task will take to complete (regular or PERT estimates, whichever you have decided to use)
a list of predecessor tasks (also from your sticky-note PERT chart).
You may number your tasks and use these numbers to list your predecessors and speed up the process.
You have already made estimates for a few of the tasks in publishing the magazine issue. Now it is time to fill in the rest and build the task table.
GANTT CHARTING
The final stage of developing your project plan is to make a Gantt chart. You will draw the chart in this chapter, then learn how to use it effectively in Chapter 5.
The Gantt chart is one of the most important tools of effective project management. It was developed by Henry L. Gantt, an ordnance engineer at Aberdeen Proving Grounds in Maryland during World War I. Gantt charts are relatively easy to prepare and enormously useful.
A Gantt chart in its simplest form is a timeline that shows the tasks in your project in terms of the calendar time they occupy. You may prepare them by hand or you may use one of a number of computer software programs that prepares them automatically. These programs may do PERT charts and perform other technical aspects of project management for you. Software options will be discussed in Chapter 5. Before you consider buying project management software, be sure you can do charting for simple projects by hand; this will help you learn what the software is trying to accomplish.
You can create a Gantt chart with a sheet of graph paper, or you may use the Gantt chart grid in Figure 4.8.
EXERCISE
Create Your Own Gantt Chart
You will need graph paper and a straight edge for drawing.
You have a project and a task table. Your assignment is to draw a timeline and determine the critical path for the project.
PROJECT: Buy a Project Management Software Package
The accompanying Gantt chart grid provides the task numbers and task names for this exercise.
Start with the first task, Needs analysis. Since the task has no dependencies, it starts at the beginning of the project. Draw a line from the beginning of the first day to the end of the fifth day because the task takes five days.
So far, so good. The next two tasks are both dependent on Task 1. If two tasks are dependent on the same predecessor, they are parallel to each other. That means that the Gantt chart bars for tasks 2 and 3 should start after the end of the Gantt chart bar for Task 1.
Now, draw tasks 2 and 3 on your Gantt chart.
Task 4 has two predecessors; it is dependent on both tasks 2 and 3. Therefore, it cannot start until both predecessors are complete; Task 4 starts after all its predecessors. Now, draw Task 4 on your Gantt chart.
Answer Key
You have worked through the various types of Gantt-chart relationships. Now, finish your Gantt chart to see how long the project will take.
Looking at the timeline, you now know that this project should take seventeen days to complete. This is the first useful thing a Gantt chart can tell you.
The next step is to determine which are the critical and which are the noncritical tasks.
Earlier in this chapter, you learned the concept of the critical path. A task is critical if delaying it instantly delays the project; it is noncritical if it has slack, or extra time, before it bumps the start of the next critical task.
Note on your finished Gantt chart the tasks that are critical and noncritical. On noncritical tasks, note how much slack is available.
There are two noncritical tasks in this project: Task 3, Choose software, and Task 6, Train user.
Task 3 has one day of slack, since Task 4, Purchase, is waiting for Task 2, Choose computer, to finish.
Task 6, Train user, has three days of slack—not two! Notice that according to the task table, Task 7, Install system, is only dependent on Task 5, Delivery time. If Task 6 is late, Task 7 can still go ahead. Task 8, Acceptance, is the only task waiting for Train user to finish. Task 6 can therefore be up to three days late without jeopardizing the start of a critical task.
Congratulations! You have successfully completed a Gantt chart. In Chapter 5, you will learn how to use it.