Earned Value Evaluation Curve
The biggest characteristic of a software project is "two uncertainties and two limits", that is, the requirements and design are uncertain; the budget and construction period are limited. There is little "living space" for earned value management.
To play the role of earned value management, you need to:
1. Truly understand the meaning and significance of earned value management.
Earned value management is difficult to understand at the beginning, even if you use it for a period of time, you can't understand the whole meaning.
There are fewer people who can understand the meaning of earned value management. Many people just want to play the digital game, and they don't know why they want earned value management.
2. The project plan and plan tracking must be carried out strictly and practically.
Project planning and plan tracking must be done well, which is the premise and foundation of earned value management.
The software project is different from other engineering projects, work can not be calculated by piece, usually can only use man-hour to calculate labor cost.
Man hour has two characteristics: 1. Task cost; 2. Task schedule
This is also an important feature of getting through the relationship between schedule and cost in software project management.
Three basic elements of Earned Value Management: PV (planned cost, planned schedule), AC (actual input cost), EV (actual completion schedule)
In the implementation of plan and plan tracking, we focus on three aspects:
1. Plan each task, the order of tasks, and the expected degree of task completion.
2. Actual investment, such as how much time and money are actually invested.
3. The actual completion of the task.
From the perspective of the company's operation, the most important concern is the actual cost (different from the time cost mentioned above, the cost here is money)
In the process of the software project, the difference of personnel and other expenses form the project cost, so how to unify the project cost and time cost
My own approach is:
A comprehensive (average) cost price is used to represent the standard personnel cost of all people (the personnel cost difference, non-personnel cost expenditure,and other factors are evenly allocated to the personnel cost).
This comprehensive cost price is based on the company's public relations expenses, site expenses, taxes, water and electricity expenses, equipment depreciation
The salary of support personnel (such as administrative and financial personnel), the salary that employees have to pay when they are not busy, the cost when the research fails, the salary of employees,and so on,
Assigned to each direct production software employee (project manager, designer, programmer, test, implementation, configuration management, QA, etc.).
Generally, an employee's salary only accounts for 30-40% of his total cost.
The standard of 100 yuan per hour is usually used. [person month = 100 yuan / hour × (22 × 8) hours = 17600 yuan]
The above is just a reference. As long as there is a way of conversion between working hours and money, for the sake of simplicity, only working hours can be used.
PV, AC,and EV used to be called BCWS, ACWP,and BCWP.
BCWS: budget cost for work scheduled, which is PV now.
ACWP: actual cost for work performed.
BCWP: budget cost for work performed, which is now EV.
Cost deviation (CV), schedule deviation (SV)
To track the progress of the plan, we will pay attention to the cost and progress. We can use the following indicators:
CV, cost variance, cost deviation, CV = EV-AC.
SV, schedule variance, schedule deviation, SV = EV-PV.
If CV and SV are equal to zero, it means that the cost and schedule are completely consistent with the plan.
If it is negative, it means that the cost is overspent and the progress is backward.
If it is a positive number, it means cost-saving and ahead of schedule.
The bigger the two numbers, the better! The larger the CV is, the lower the cost is. The larger the SV is, the more advanced the schedule is,
When CV and SV are zero, they are consistent with the plan; when CV and SV are negative, they are cost overrun and progress backward.
Cost index (CPI), schedule index (SPI)
CPI, cost performance indicator, CPI = EV / AC.
SPI, schedule performance indicator, SPI = EV / PV.
CV is the subtraction operation of EV and AC, while CPI is the division operation of these two; similarly, SV is the subtraction operation of EV and PV, and SPI is the division operation of these two.
If CPI and SPI are equal to 1, it means that the cost and schedule are consistent with the plan.
If it is greater than 1, it means that the cost is saved and the progress is ahead of schedule.
If it is less than 1, it means cost overrun and progress lag.
The bigger the two numbers, the better! The larger these two figures are, the more ideal the project is, the lower the cost and the more advanced the schedule is.
CPI and SPI are the performance indicators of a project. These two indicators actually reflect the current performance of the project's cost and schedule. The larger the number, the better.
Cost forecast (EAC)
EAC is the abbreviation of the estimate at completion, which means the cost estimation when the project is completed.
Project managers and project owners must be very concerned about the final cost of the project, so how can we predict the final possible cost in the process of the project?
There is such a prediction formula: EAC = AC + (BAC - EV) / CPI
BAC is the abbreviation of the budget at completion.
Assuming that the current CPI value of the project can be maintained consistently, then the final cost of the project must be the number predicted by this prediction formula.
If the CPI can be continuously increased in the future, the final cost of the project will be less. On the contrary, if the CPI is reduced in the future, the final cost of the project will be greater.
In other words, as a project manager, in order to reduce the final cost of the project, you need to control the CPI at any time, so that the CPI value is greater than or equal to 1, the larger the better.
A project has many tasks, and each task has its PV, AC,and EV. At the beginning of the project, PV is planned. There are many tasks on the timeline. The later the project, the greater the cumulative PV value of each task. In the beginning, AC and EV of each task are zero, but with the development of the project, the cumulative value of AC and EV will continue to increase.
1. If the project needs to work overtime all the time, but can be completed on time, how will the relationship among PV, AC, and EV change?
AC > PV, EV and PV tend to be the same.
2. If the project doesn't need to work overtime and can basically complete the task on time, how will the relationship among PV, AC, and ev change?
3. How does the relationship among PV, AC and ev change if the project does not need to work overtime and can complete the task ahead of time?
Cost saving: EV / AC > 1; progress ahead: EV / PV > 1
Conclusion: earned value management is not a digital game, we should make it practical!
Earned value management makes project management digital and visual, but it is often used in actual project management, but the actual effect is not ideal.
Here are some common cases:
1. Project plan and project plan tracking are not well done.
Many software projects, there is no documented plan, no detailed schedule, a lot of work through meetings, oral arrangements and other ways to implement. In this way, the three basic values of PV, AC and ev can not be obtained, so the earned value management can not be implemented.
2. There is no effective and simple way to collect the actual data.
The implementation of plan and plan tracking is the basis of earned value management, but if there is no effective and simple way of actual data collection, earned value management can not be fully implemented.
Generally, everyone uses Project to make progress plans. If the tasks can be detailed, they can get the PV; if we can update the completion of the tasks in time, we can get the EV; but the actual working hours (AC) are not easy to count Up. Some companies collect actual working hours once a week or every month, and some companies also specially arrange project assistants to be responsible for data collection. But these practices will make the project team members very uncomfortable at the beginning, and there are very few projects that can persist.
3. The "two uncertainties and two limitations" feature of the software project itself.
Even if you can solve the above two problems, you cannot solve the "two uncertainties and two limitations" characteristics of software projects. This feature prevents you from planning all the work of the project at the beginning of the project, and the plan often does not match the actual situation, and you need to adjust it frequently. In other words, you cannot set all the PVs, and the PVs will be adjusted frequently. PV cannot be fully determined and will be adjusted frequently. This means that the baseline of the project plan is often changing. You will find that it is not just PV. The values of AC and EV will change very frequently. The changes of CV, SV, CPI, and SPI will vary greatly. Big. You will feel that after spending so much thought and work on earned value management, the benefits are not great.
The project plan includes various plans such as development plan, test plan, procurement plan, training plan, configuration management plan, etc. Generally, different people are responsible for these plans. These plans are generally written separately, not in the same document, and Their manifestations are also different. To systematically track the PV, AC, and EV of these plans, it is necessary to solve the problem of the unity of these plans, and the person in charge of these plans should not be responsible for a lot of extra work just because of the need to measure PV, AC, and EV.
Back to the most fundamental question, why do we use earned value management? Is it to pursue quantification?
If the CPI and SPI values of a certain project are ideal and are always near 100%, would you think the situation of the project is ideal?
Numbers cannot explain all the problems, numbers often deceive us! The value of CPI and SPI is 100%, and the project may still have serious problems. If there is a problem with the project plan, the key work is omitted and there is no arrangement, so the PV is actually problematic. After implementing such a plan, the AC and EV obtained and the CPI and SPI calculated are actually meaningless. Earned value management is actually a comparison between the plan and the execution of the plan. It is necessary to ensure that there is no problem with the plan. This comparison is meaningful.
The CPI and SPI of a certain project are both low, indicating that the project is overspending and behind schedule. How would you think of improvement measures?
The most common one is to work overtime to solve the problem of behind schedule;
In general, project overruns can only save a little as much as possible, and there is no better solution.
Generally, the way we think of is to find a way to make the AC less and make the EV more, but rarely find a way to reduce the PV!
Software projects are human high-level intellectual activities, and a very intelligent idea can solve many problems.
The software project requirements are uncertain and the design is not specific. We need to use more wisdom to actively meet these two challenges and find ways to reduce the workload!
Regarding earned value management PV, AC, EV, my opinion is:
PV: You have to do everything possible to reduce PV. This is the most effective way to make the project successful, and it is the most important thing you need to do!
AC: The level of the project team must be continuously improved, so that everyone can learn efficient working methods, and let everyone use the right methods to do the right things, so as to reduce the AC.
EV: The completion criteria for each task must be clear, the tasks must be sufficiently subdivided, and long-term tasks should not be arranged. There are only two states for a task: completed and uncompleted. It is not reliable to say that the task is 90% complete. As long as it is not completed, the state is incomplete, and the EV of the task is zero.
I think the greatest essence of earned value management should be your understanding of PV, AC, and EV! PV, AC, and EV do not need to be quantified. The key to project success lies in the overall ability and knowledge of the project team, and the organization and rigor of work. You think there is a problem with project management. In fact, the problem may not be because earned value management is not applied, but because the level of the project team is not enough and the work is unorganized.