Section N
‘S’ Curves
Plotting the work done (man hours, orders placed, money spent, or percentage progress) against time produces the ‘S’ curve shape. Unfortunately the indications they provide are retrospective; nevertheless, they indicate the status of the project and help to forecast where the project is going.
As well as the overall ‘S’ curve, individual curves for design/engineering, the purchase of materials and equipment, and construction should be plotted to give an overall picture of the status of the project
A common flaw with any progress measurement system (that needs careful monitoring) is that people will perform the easy items of work first. This is because that is how they get measured/rewarded with the progress recorded on the ‘S’ curve.
In the early stages of the project (when the basic controls are still being established), it is difficult to obtain the realistic data needed to create meaningful curves. Nevertheless, it is important to analyse the health of the project as soon as possible in the launch phase. Some suggested diagnoses of typical project situations are shown below.
1 Interpreting the Curves
1.1
Case 1—see Figure V.N.1: In the diagram, the man hours used and the actual progress are coincidental, but progress is way behind schedule.
What's wrong?
1.1.1
If the man hours used and the progress are matching, then there is nothing wrong with the productivity—it's 100 per cent. So, the answer must be that not enough man hours are being expended. There are not enough people doing work in order to achieve the required progress. However, is there enough work available for more people at this early stage?
1.2
Case 2—see Figure V.N.2: In this diagram, the actual progress matches the schedule requirements, but the man hours used are excessive. Again, what's wrong?
1.2.1
There is nothing wrong with the motivation of this team if they are consistently meeting the schedule. So, why are excessive man hours being expended?
1.2.2
If the motivation is good and the hours are high, there could be two or three similar reasons. Firstly, the estimate was wrong. However, if the organization is sophisticated enough to measure progress in this manner, then they are unlikely to get the estimate consistently wrong over the whole range of the project work.
1.2.3
I believe that it is more likely that the team is doing additional work that has not been recorded in the system. Is this additional work internal to the company, or is a client interfering in a reimbursable project? If this pattern existed more in the middle of the project rather than the launch phase, then I would suggest that the extra work are client changes that have not been recorded.
1.3
Case 3—see Figure V.N.3: In this diagram, the actual progress is way behind the schedule requirements, and the man hours used are far too excessive. Once again, what's wrong?
1.3.1
This project team is a mess! The project will probably never recover unless the project is stopped and started again after some serious team building. Too many people have been brought onto the project too early, and there is not enough work available for them to do.
1.4
Case 4—see Figure V.N.4: In this diagram, the scheduled progress, actual progress, and the man hours used are all coincidental. This time there is nothing wrong. Nevertheless, the project is not perfectly under control since the cost of the man hours is over budget. The project manager has to decide whether to use experienced and more expensive people or younger people costing less than was budgeted for. However, one would require less supervision and thus save some money, and the younger group would require more supervision and cost more. If money is the only criteria, the skill is in getting the balance right. However, other factors come into play; see Section Q, subsection 1, Selecting the Team.
1.5
A Deception?
If you can justify taking a break, say over the Christmas holidays (and in the real situation we were moving offices as well, so it seemed justified), then the ‘S’ curve goes horizontal as in Figure V.N.5. If you didn't start well and, despite the circumstances, you manage to get some people to carry on working (as we did), then hey, presto, you are ahead of schedule.
1.6
A Self‐deception
See Part 1, Section E The Manager of Projects, paragraph 1.1, d, Figures I‐E‐1 and I‐E‐2.
1.7
Plotting the early‐start (ES) dates for all activities, from the analysis of the critical path network, also produces an ‘S’ curve. Another similar curve is also produced by plotting the late‐finish (LF) dates (see Figure V.N.6). The question now is: do you hide the LF dates before issuing a schedule of completion dates for the activities and putting the derived curve up on the conference room wall?
1.7.1
At some stage, an activity will be finished late. However, since you also have the late‐finish dates and associated ‘S’ curve, you may not be too concerned. You can see that the date achieved is within the envelope of the ES and LF curves. Consequently, you don't make too much of a fuss about it. What happens next is that more dates slip because the target provided doesn't seem to be too important. The project then slips into trouble.
1.7.2
It has happened because you showed a lack of trust in your team. Give people some flexibility and trust them; show both the ES and LF curves. Nevertheless, make it clear that if they complete after the LF date, they are in real trouble. Options are to show a target curve somewhere in between or, in order to motivate people, only issue the target curve as the visible one to aim for. Discuss these issues as part of team building.
2 Change Orders
2.1
Assuming a project of 250,000 hours and seventeen months duration, then the ‘S’ curve would be as shown in Figure V.N.7. This shows a straight line portion of 69.3 per cent to be performed in nine months, that is, 7.7 per cent per month
(a good rate of progress) equivalent to 19,250 man hours per month and, on the basis of a forty‐hour week (160 hours per four‐week month), 120 people.
2.2
The project has expended 75,000 man hours and is thus 30 per cent complete and on the schedule curve. However, the client has asked for a variation, a change of 50,000 man hours, making the total project hours 300,000. Consequently, the project is now only 25 per cent complete as shown by the vertical line drop in the progress curve; see Figure V.N.8. There is a problem now, since to maintain the same rate of progress at 7.7 per cent, additional people are required.
2.2.1
7.7 per cent of 300,000 is 23,100 man hours. Therefore, 23,100 – 19,250 = 3,850 additional man hours. On the same forty hours per week basis, this equates to twenty‐four additional people.
2.2.2
If additional people are not brought on board, the man hour expenditure will remain at 19,250, meaning that the rate of progress will drop to 6.4 per cent per month with a consequent two‐month delay to the schedule; see Figure V.N.9.
2.3
The client will insist, nevertheless, that the project is completed to the original schedule and to ‘man up’ accordingly. On the basis of the man hours required to complete the project and the time available, a progress rate of 8.2 per cent seems acceptable. However, this is now becoming a high number to maintain consistently.
2.3.1
8.2 per cent of 300,000 is 24,600 hours per month. 24,600 hours at 160 per month is equivalent to (153.75) 154 people. It may be tempting to ‘take the money and run’, namely, to man up and get paid for the additional man hours and say that the schedule can be met because that is what the calculations indicate.
2.4
However, we started with 120 people, and we are now proposing to man up to 154 people, an extra thirty‐four people. You cannot add this number of people and expect to make immediate progress at the desired rate. A good number of the existing project team will need to stop work and explain what needs doing. The office layout of the project will need to be adjusted, there may need to be some additional team building to bring everyone on board, and finally there will be a learning curve as there would be at the beginning of a project. Thus, despite manning up, the project is still going to be late as illustrated in Figure V.N.10.
2.5
If the project is going to be late in any case, it might be better to consider not adding the extra people but working overtime. On the basis of one hour an evening or five hours at the weekend, a forty‐five‐hour week, the number of personnel drops to 137, that is the additional personnel drops by half to 17. Working evenings and weekends (a fifty‐hour week) means that the total number required on the project is only 123 people. Three more than the original team, and theoretically, the project will finish on time. However, productivity drops off with regular scheduled overtime (see Part IV, Section Q Installation and Construction, paragraph 4.7).
2.5.1
The other option to consider is to take the extra work and to set up the thirty‐four additional people as a separate project team. You might even be able to generate some competitive spirit between the two teams, based on productivity levels.
2.6
When many individual changes have become significant, they should be summated and treated as a single change (as in paragraphs 2.2 to 2.4 above). Their cumulative effect should then be determined by adding a re‐launch learning curve (in 2.4 above). The re‐launch learning curve should be justified by being based on the historical data obtained in the launch phase. The re‐launch curve should be plotted in direct proportion to the hours used and time taken to achieve the initial straight line section of the ‘S’ curve. You may have difficulty persuading the client to accept this reality check! See Part III, Section F Contracts, paragraph 3.7 to 3.9 about restraining the client's changes.