estructura desglosada del trabajo

n a previous chapter, I said that planning answers the ques-tions What must be done?, How long will it take?, andHow much will it cost? Planning the what is vital; proj-ects frequently fail because a significant part of the work isforgotten. In addition, once tasks have been identified, thetime and resource requirements must be determined. This

is called estimating.A major problem in project planning is determining how long

tasks will take and what it will cost to do them. Inaccurate esti-mates are a leading cause of project failures, and missed cost tar-gets are a common cause of stress and recrimination in projectmanagement.

The most useful tool for accomplishing all of these tasks is theWork Breakdown Structure (WBS). The idea behind the WBS issimple: You can subdivide a complicated task into smaller tasks,until you reach a level that cannot be further subdivided. At thatpoint, it is usually easier to estimate how long the small task will

56

Using the WorkBreakdown Structure to Plan a Project

CHAPTER 5CHAPTER 5

II

16093-FundProjectMgmt3rdEd 8/16/06 2:40 PM Page 56

Guillermo_PCSticky NoteCHAPTER 5: Using the Work Breakdown Structure to Plan a Project. p56-68.Lewis, James P. Fundamentals of Project Management 2007

take and how much it will cost to perform than it would havebeen to estimate these factors for the higher levels.

Nevertheless, it is still not easy to estimate task durations foractivities that have never been performed before. Because this isthe typical situation in engineering hardware and software devel-opment projects, we might expect many of these estimates to bein error, and this seems to be demonstrated by experience. Still,the Work Breakdown Structure makes it easier to estimateknowledge tasks than any other tool we have.

A Simple ExampleAs an example, if I want to clean a room (see Figure 5-1), I mightbegin by picking up clothes, toys, and other things that have beendropped on the floor. I could use a vacuum cleaner to get dirt outof the carpet. I might wash the windows and wipe down thewalls, then dust the furniture. All of these activities are subtasksperformed to clean the room.

As for vacuuming the room, I might have to get the vacuumcleaner out of the closet, connect the hose, plug it in, push thevacuum cleaner around the room, empty the bag, and put themachine back in the closet. These are still smaller tasks to be

Using the Work Breakdown Structure to Plan a Project 57

CleanRoom

Pick up toys& clothes

Vacuumcarpets

Wash walls Dustfurniture

Cleancurtains

Get vacuumout of closetConnect hoseand plugPush aroundroomEmpty bag

Figure 5-1. WBS diagram to clean a room.


performed in accomplishing the subtask called vacuuming. Thediagram in Figure 5-1 shows how this might be portrayed inWBS format.

Note that we do not worry about thesequence in which work is performedwhen we do a WBS. That will be workedout when we develop a schedule. How-ever, you will probably find yourself think-ing sequentially, as it seems to be humannature to do so. The main idea of doing aWBS is to capture all of the tasks. So ifyou find yourself and other members ofyour team thinking sequentially, dont betoo concerned, but dont get hung up ontrying to diagram the sequence or youwill slow down the process of task identi-fication.

The typical WBS has three to sixlevels, and these can be named as shown in Figure 5-2. It is, ofcourse, possible to have projects that require a lot more levels.Twenty levels is considered to be the upper limit, and that is ahuge project. Note that level 1 is called the program level. Thedifference between a program and a project is just one ofdegree.

An example of a program is the development of an airplane.For example, the WBS for Boeings 777 airplane program mighthave been drawn as shown in Figure 5-3. Notice that the engine,wing, and avionics are large enough jobs to be called projects intheir own right. In fact, the program managers job is to makesure that the projects are all properly integrated. The enginemounts on the wing, so, somewhere in the structure to developthe engine, there will be an activity called Design wingmounts. And for the wing, there will be an activity calledDesign engine mounts. If these are not coordinated properly,you will wind up with an engine that wont mount on the wing.The job of coordinating these is called system integration.

58 Fundamentals of Project Management

A Work BreakdownStructure does notshow the sequencein which work is performed! Suchsequencing isdetermined when a schedule is developed.



Guidelines for Developing the WBSOne important question in constructing a WBS is When do youstop breaking down the work? The general guideline is thatyou stop when you reach a point whereeither you can estimate time and cost tothe desired degree of accuracy or thework will take an amount of time equalto the smallest units you want to sched-ule. If, for instance, you want to sched-ule to the nearest day, you break downthe work to the point where tasks takeabout a day to perform. If you are goingto schedule to the nearest hour, thenyou stop when task durations are in that range.

1. Program

2. Project

3. Task

4. Subtask

5. Work Package

6. Level of Ef fort

Figure 5-2. WBS level names.

Stop breaking downwork when youreach a low enoughlevel to do an esti-mate of the desiredaccuracy.



Remember the rule that the people who must do the workshould participate in planning it? That applies here. Usually a coregroup identifies top-level parts of the WBS; those parts are furtherrefined by other members of the team and then integrated toobtain the entire WBS.

One important point: the WBS shouldbe developed before the schedule. In fact,the WBS is the device that ties the entireproject together. It allows resources to beassigned and estimates of time and cost tobe made and shows the scope of the jobin graphic form. Later, as the project istracked, the work can be identified asfalling in a particular box in the WBS.

There is at least one software pack-age, SuperProject Expert, that prints aWBS after schedule data have beenentered. That is a nice feature, since itgives a graphically attractive WBS, but the rough drawing shouldbe made before you use the scheduling software. The reason isquite simple: until everyone has agreed that all tasks have beenidentified, it is misleading to develop a schedule. You cannot besure that the critical path identified by a partial schedule will bethe same for the full schedule.

DesignAirplane

DesignWing

DesignEngine

DesignAvionics

Figure 5-3. Partial WBS for the 777 development program.

The WBS shouldalways be devel-oped before theschedule is workedout, but withouttrying to identifythe sequence ofactivities.



There are a number of approaches to developing the WBS. Ide-ally, you proceed top-down, following development of a good prob-lem statement and mission statement. As I have mentioned,however, the mind does not always oper-ate in such nice, linear fashion; as youdevelop the WBS, you may sometimesfind that it helps you to understand thejob better. For that reason, I am not apurist about doing things in a specificorder. You do what works best for you.

The WBS does not have to be sym-metrical. That is, all paths need not bebroken down to level 6 (or whateverlevel you stop at). Since the rule is to break work down to alevel sufficient to achieve the estimating accuracy you desire,one path may take six levels, while another may need onlythree.

Uses of the WBSAs I have said, the WBS is a good way to show the scope of a job.If you have ever given someone an estimate for project cost ortime and seen the persons horrifiedlook, you know that they are seeing theproject in their mind as much simplerthan it is. When you show a project inWBS form, it is clear to most individualswhy the job costs so much. In fact, Ihave had the experience of the planning group members them-selves being overwhelmed by the complexity and magnitude of the WBS. If it impresses them, think of the impact on theoutsider.

Assigning responsibility for tasks is another important use ofthe WBS. Each task to be performed should be assigned to a par-ticular person who will be responsible for its completion. Theseassignments can then be listed on a separate form, often called aResponsibility Chart (see Figure 5-4).

The WBS is a goodway to portray thescope of a project.

A WBS does nothave to be symmet-rical. All paths donot have to go downto the same level.



Estimating Time, Costs, and ResourcesOnce the work is broken down, you can estimate how long it willtake. But how? Suppose I ask you how long it will take to sort astandard deck of playing cards that has been thoroughly shuffledinto numerical order by suit. How would you answer that question?

The most obvious way would be totry the task several times and get a feel-ing for it. But if you didnt have a deck ofcards handy, you would probably thinkabout it, imagine how long it wouldtake, and give me an answer. Peoplegenerally give me answers ranging fromtwo minutes to ten minutes. My testsindicate that about three minutes is aver-age for most adults.

Suppose, however, we gave the cardsto a child about four or five years old. It might take a lot longer,since the child would not be that familiar with the sequence in

Linear Responsibility ChartProject: Date Issued: Sheet Number: of

Manager: Date Revised: Revision No. File:

LRCFORM.61

Project Contributors

Task Descriptions

CODES: 1 = ACTUAL RESPONSIBILITY; 2 = SUPPORT; 3 = MUST BE NOTIFIED; BLANK = NOT INVOLVED

Figure 5-4. Responsibility chart.

An estimate can be made only bystarting with theassumption that acertain resourcewill be assigned.


which cards are ordered, and perhaps not yet even that comfort-able with counting. So we must reach a very important conclusion:You cannot do a time or cost estimatewithout considering who will actuallyperform the task. Second, you must basethe estimate on historical data or a men-tal model. Historical data are best.

Generally we use average times toplan projects. That is, if it takes three minutes on average for adultsto sort a deck of cards, I would use three minutes as my estimateof how long it will take during execution of my project. Naturally,when I use averages, in reality some tasks will take longer than thetime allowed and some should take less. Overall, however, theyshould average out.

That is the idea, anyway. ParkinsonsLaw discredits this notion, however.Parkinson said that work always expandsto fill the time allowed. That means thattasks may take longer than the estimatedtime, but they almost never take less.One reason is that when people findthemselves with some time left, theytend to refine what they have done.Another is that people fear that if theyturn work in early, they may be expectedto do the task faster the next time or thatthey may be given more work to do.

This is a very important point: if people are penalized for per-forming better than the target, they will quit doing so. We alsohave to understand variation. If thesame person sorts a deck of cards overand over, we know the sort times willvary. Sometimes it will take two min-utes, while other times it will take four.The average may be three, but we may expect that half the timeit will take three minutes or less and half the time it will take


We must be carefulnot to penalizeworkers who per-form better thanexpected by loadingthem down withexcessive work.

An exact estimateis an oxymoron!

Parkinsons Law:Work expands to fillthe time allowed.


three minutes or more. Very seldom will it take exactly threeminutes.

The same is true for all project tasks. The time it takes to per-form them will vary, because of forces outside the persons con-trol. The cards are shuffled differently every time. The personsattention is diverted by a loud noise outside. He drops a cardwhile sorting. He gets tired. And so on.

Can you get rid of the variation? No way.Can you reduce it? Yesthrough practice, by changing the

process by which the work is done, and so on. But it is importantto note that the variation will always be there, and we must rec-ognize and accept it.

The Hazards of EstimatingConsider the case of Karen. One day her boss stopped by herdesk at about one oclock. Need for you to do an estimate forme, he told her. Promised the Big Guy Id have it for him byfour oclock. You with me?

Karen nodded and gave him a thin smile. The boss describedthe job for her. Just need a ballpark number, he assured her anddrifted off.

Given so little time, Karen could compare the project herboss described only to one she had done about a year before. Sheadded a little for this and took a little off for that, put in somecontingency to cover her lack of information, and gave the esti-mate to the boss. After that, she forgot all about the job.

Two months passed. Then the bomb was dropped. Her bossappeared, all smiles. Remember that estimate you did for me onthe xyz job?

She had to think hard to remember, but, as her boss dronedon, it came back to her. He piled a big stack of specifications onher desk. Its your job now, he told her and drifted off againinto manager dreamland.

As she studied the pile of paper, Karen felt herself growingmore concerned. There were significant differences between thisset of specs and what her boss had told her when she did the esti-mate. Oh well, Im sure he knows that, she told herself.



Finally she managed to work up a new estimate for the job onthe basis of the real specs. It was almost 50 percent higher thanthe ballpark figure. She checked her figures carefully, assured her-self that they were correct, and went to see her boss.

He took one look at the numbers and went ballistic. Whatare you trying to do to me? he yelled. I already told the oldman we would do it for the original figure. I cant tell him its thismuch more. Hell kill me.

But you told me it was just a ballpark number you needed,Karen argued. Thats what I gave you.But this is nothing like the job I quoted.Its a lot bigger.

I cant help that, her boss argued.I already gave him the figures. Youllhave to find a way to do it for the origi-nal bid.

Naturally, you know the rest of thestory. The job cost even more thanKarens new estimate. There was a lot ofmoaning and groaning, but, in the end, Karen survived. Oh, theydid send her off to a course on project managementhoping, nodoubt, that she would learn how to estimate better in the future.

Guidelines for documenting estimates:

Show the percent tolerance that is likely to apply.

Tell how the estimate was made and what assumptionswere used.

Specify any factors that might affect the validity of the estimate(such as timewill the estimate still be valid in six months?).

Could you fault Karen for anything? Well, perhaps. If shefailed to tell the boss that a ballpark estimate may have a toler-ance of perhaps 25 percent but that the margin of error canrange from 10 percent to +100 percent, then she allowed himto think that the estimate was better than it was. Also, she shouldhave documented all working assumptions, explaining how she


One of the primarycauses of projectfailures is that ballpark estimatesbecome targets.


did the estimate, what project she had used for comparison, andso on. Then, if management still pulled a whammy on her, atleast she would have had some protection. In fact, it is impossibleto make sense of any estimate unless these steps are taken, sothey should be standard practice.

Consensual EstimatingIn recent years, a new method of estimating knowledge work hasbeen developed that seems to work better than older techniques.Rather than have individuals estimate task durations, the newmethod asks at least three people to estimate each activity in theproject that they know something about. They do this withoutdiscussing their ideas with one another. They then meet to findout what they have put on paper. In a typical situation, there maybe a range of times, such as, for example, ten days, twelve days,and thirty days, in which two of the estimates are close together,but one is very different. How do you handle the discrepancy?

The best approach is to discuss what each person was con-sidering when he or she made the estimate. It may be that theperson who put down thirty days was thinking about somethingthat the other two had overlooked. Or, conversely, they may con-vince the thirty-day person that his number is way too high andget him to come down to a figure nearer their estimates. In anycase, they try to arrive at a number that they all can support. Thisis called consensus.

There are three advantages to this approach. First, no oneperson is on the hook for the final number. Second, inexperi-enced people learn to estimate from those more experienced.Third, they are likely to collectively consider more issues thanany one person would do working alone. For that reason, you aremore likely to get an accurate estimate, although it is importantto remember that it is still not exact!

Improving Estimating AbilityPeople cannot learn unless they receive feedback on their perfor-mance. If you went out every day and ran 100 yards, trying to



improve your speed, but you never timed yourself, you wouldhave no idea whether you were getting better or worse. You couldbe doing something that slowed you down, but you wouldntknow it. In the same way, if you estimate task durations butnever record the actual time it takes to do the task, you are nevergoing to get better at estimating. Furthermore, you have to trackprogress by recording times daily. If you record times once aweek, I can promise you that you will be just guessing, and thatwont be helpful.

Key Points to Remember Do not try to work out sequencing of activities when

you develop a WBS. You will do that when you develop aschedule.

A WBS ties the entire project together. It portrays scopegraphically, allows resources to be assigned, permits esti-mates of time and costs to be developed, and thus providesthe basis for the schedule and the budget.

An estimate is a guess, and an exact estimate is an oxy-moron!

Be careful that ballpark estimates dont become targets.

Consensual estimating is a good way to deal with activitiesfor which no history exists.

No learning takes place without feedback. Estimate; thentrack your actual time if you want to improve your estimatingability.

Exercise . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Following is a list of tasks to be performed in preparation for acamping trip. Draw a WBS that places the tasks in their proper rela-



tionship to one another. The solution is contained in the Answerssection.

Arrange for supplies and equipment.

Select campsite.

Make site preparations.

Make site reservation.

Arrange time off from work.

Select route to site.

Prepare menu for meals.

Identify source of supplies and equipment.

Load car.

Pack suitcases.

Purchase supplies.

Arrange camping trip (project).



estructura desglosada del trabajo

Documents