Project health monitoring by Earned Value Analysis

Transcription

1 13th International Software Testing Conference (STC 2013) December 0-06, 2013 in Bangalore, India. Project health monitoring by Earned Value Analysis Gangadhar. B. Kallur Honeywell Technology Solutions, 19/2 Devarabisanahalli Village, KR Varturhobli, Bangalore 60037, 1

2 Abstract Execution of engineering projects are tracked against critical metrics such as safety, quality, delivery cost and inventory. Earned value is a key parameter that helps in assessing delivery (schedule) and cost. Static shows that 70% of projects are over budget behind schedule, 2% of all projects finish at 189% of their initial budget and some, after huge investments of time and money, are simply never completed. The rest of this paper gives a perspective on monitoring project health by Earned value analysis. I. INTRODUCTION Earned Value (EV) technique is an industry standard way to measure a project s progress, forecast its completion date and final cost, and to provide schedule and budget variances along the way. By integrating three measurements, it provides consistent, numerical indicators with which you can evaluate and compare projects EV. EV is primarily a schedule tracking tool, determining schedule from cost and earned value of individual Work Breakdown Structure (WBS) tasks performed in time. A) History Earned Value Management (EVM) emerged as a financial analysis specialty in United States Government programs in the 1960s [1], but it has since become a significant branch of project management and cost engineering. Contractors pushing high tech, newly developed weaponry. Military having critical schedule needs ( Arms Race ) 1990s Policy moved Earned Value into all Federal agencies. OMB Circular: A-11, NASA Policy Directive 901.3, DOD 000.2R, and DOE Order to name a few. B) Terminologies i. PV - Planned Value or Budgeted Cost of Work Scheduled (BCWS) It is the amount of work (person-hours) planned to accomplish at a given point of time (this is from the Work Breakdown Structure in plan) Project management research investigating the contribution of EVM to project success suggests a moderately strong positive relationship. Implementations of EVM can be scaled to fit projects of all sizes and complexities. - Earned value based performance management began in the 1960s, based initially on Department of Defense (DOD) Cost/Schedule Control Systems Criteria (C/SCSC). - Earned Value was used as an objective measure for progress, i.e., physical accomplishment. 1970s-80s The DOD continued the use of Earned Value in response to bearing cost and schedule risk in cost-plus contracting. Figure 1: Earned Value example ii. AC - Actual Cost or Actual Cost of Work Performed (ACWP) It is the amount of work (person-hours) which have been actually spent at a given point of time 2

3 iii. EV - Earned Value or Budgeted Cost of Work Performed (BCWP) It is the value (person-hours) in terms of base budget of what you have accomplished at a given point of time (or, % complete X Planned Value) iv. SV: Schedule Variance (EV-PV) Refer figure 1&4 It is the comparison between amount of work performed during a given period of time and what was scheduled to be performed. A negative variance means the project is behind schedule. Ex: Schedule Variance = EV-PV 49,000hr -,000hr SV = - 6,000hr v. CV: Cost Variance (EV-AC) A comparison of the budgeted cost of work performed with actual cost. A negative variance means the project is over budget. Ex: Cost Variance = EV-AC 49,000hr- 6,000hr CV = - 7,000hr C) More Derived Metrics i.spi: Schedule Performance Index SPI=EV/PV If SPI is lesser than 1, means project is behind schedule. ii.cpi: Cost Performance Index CPI= EV/AC If CPI is lesser than 1, means project is over budget. iii.csi: Cost Schedule Index (CSI=CPI x SPI) When CSI is greater than 1, the less likely project recovery becomes. Example of performance metrics SPI: EV/PV 49,000/,000 = CPI: EV/AC 49,000/6000 = 0.87 CSI: SPI x CPI.891 x.87 = Once a project is 10% complete, the Overrun at completion will not be less Than the current overrun. Once a project is 20% complete, the CPI does not vary from its Current Value by more than 10%. The CPI and SPI are statistically accurate indicators of final cost results J a n-03 F e b-0 3 M a r-0 3 A p r M ay -0 3 J u n J ul-03 A u g-0 3 S e p-0 3 O c t-0 3 N ov -0 3 D e c -0 3 BCWS BCWP ACWP BCWS BCWP ACWP Figure 2: Schedule and Cost Variance 0 J a n -0 3 F e b -0 3 M a r-0 3 A p r-0 3 M a y -0 3 J u n -0 3 J u l-0 3 A u g -0 3 S e p -0 3 O c t -0 3 N o v -0 3 D e c -0 3 Figure 3: Projection of PV EV and AC. 3

4 Remaining 30 hours are for review, moderation, re work +closer. If we break 100 hours work in smaller tasks, then 16 hour will be the baseline. So that each task activity will be traced and efforts can be updated and manger can do better delivery. Figure 4: Schedule and Cost Variances D) An Example to illustrate the EV D 1. D 2. By refereeing to Figure 6, all four projects of schedule and Budget variance says that Project A is in good shape, under budget and ahead of schedule. Project B is over budget, but if we realize it is ahead of schedule, we can slow the effort to save money. Figure : complete cycle of Test task By Referring the Figure, A set of module test is divided based on function calls. To test each function call complete cycle of Test task need to be done. For initial module it will take more than base line hours. For all models base line hour will be assumed as 16 hours. From Task Gather Data to Structural coverage, total 70% of test task completed. Once it is sent for review, then only Manger will notice the EV and can take credit for the module and till then based on Tester and manger trust, work will be done. If baseline hour is more i.e. 100hr. considering above example, 70% means 70 hours of credit is taken for completion of Test task. If work is not done properly again project manger may need to take alternate way to compensate these 70 hours. Figure 6: Budget Variance Vs Schedule Variance Project D s cost variance looks better than A s, but if we realize it is behind schedule, we can spend more to catch up. Project C looks just like B if we only look at the cost variance, but we can t slow down to save money. And it looks just like D if we only look at schedule, but we can t spend more to catch up. So, now this can be re-scoped. 4

5 E) Best practices followed to get EV E1. One of our Honeywell Operating system elements is called Visual Management (VM). Following are the list of VM. -VM will be able to understand the project status within 3 Seconds and from 3 meters by anybody. -To clearly and concisely identify the current status of a process management (VM process). -Brings problems to the surface on an everyday basis. -Easy maintenance. We are following person vice daily updates in VM board with user itself. Lead will collect the updates every week end and communicates the same to team (if it is good or bad). E3. Minority Report Tool/JIRA Tool For one TEST SLTP/CTP, total estimation will be pre-defined based on scope. Each person has to fill completion of his task and actual hours taken to complete the task in the tool. In Test Development, User will update the status states defined with % in the MR tool. Status will be given for the following states. Not started, Gather Data, Analysis, Test Analysis, Test Development, Hand Code Test development, Trace Development, Test Execution, Structural coverage, Test Review/Moderation, Test Re work, Set for Oversight review. Once updating the required fields in the tool, Planned EV and actual EV will be calculated and auto populated. Figure 7: VM Board E2. Activity based management (ABM) It is a method of identifying and evaluating activities that a business performs. In XLS sheet it will have Tabs called Life cycle, Data, PCT Chart, EV chart, RFS state. In Data tab, Task is assigned with responsible person, base line hours, actual start, Percentage complete these data s are filled daily/weekly/every 1days by Leads. This ABM xls will be maintained in secured share drive. Figure 8: Minority Report tool E4. With SAP login Data, we can fetch all logged efforts and calculate the EVs per person. User has to fill the timesheet with proper subphases (i.e check, Do, Improve, Prepare, Rework). Weekly timesheet will be submitted for approval.

6 User (Lead/Manager) who has admin rights will fetch the SAP data and sort according to the subphases. Calculates the EV for each person. F) What to Do With EVM Data EV indicators provides an early warning about the project performance and also gives an unambiguous lead to the project manager on what is the most appropriate action. Because of that the uantitative value of EV indicators is very important. Analogy: If your cholesterol is high, it is an early warning that you have to do something. However, what you have to do depends principally on the cholesterol level. If it s not too high, you can consider changing your lifestyle, doing more exercise, etc. On the other hand, if the cholesterol level is too high, your doctor will probably recommend more drastic measures, like medication, etc It s the same with a project. But, as your doctor, before prescribing a treatment, will probably perform additional exams, so the project manager has to analyze other factors to decide what the most appropriate action is. One of the important factors he has to take into account is schedule compression. If the schedule is already compressed, it could be very dangerous to try to recover the lost time by adding new resources. The best way may be to reschedule the project and to establish a new baseline. On the contrary, if the project schedule is comfortable (planned duration is much longer than the critical path), adding people could solve the problem. II. MAIN RESULTS By applying EVA technique, Project will be delivered at the maximum percentage on time. Each day s project status, how it is preceding and Project analysis will also be done. (i.e., to know if it is over budget, behind schedule). 6 Considering the example below: Suppose we are calculating EV for 6 employees for a week. As shown in the table1 below, the total EV required is But, actual EV is Which means the Cost Variance is negative and the project is over budget. Table 1: EV per person Name EV(hr EV EV EV EV Tot ) A B C D E F Tot Req per day Pros: i. EVA is a flexible and easy to learn tool adaptable to projects of any size, able to control them all with minimal effort. ii. Since all the process activities (artifacts, license, ctp/hilite queues, CM21, analysis time, etc) required for project execution are in place, the outcome of deliverables will be on-time. iii. EV analysis/calculation process can asses in real time true project schedule and cost performance. iv. With EV analysis, a project manager will be allowed to only study and manage cost in time, leaving the performance metric of technical variable to technical discipline leads. This burden release emphasizes the benefit of 20% effort for 80% results. v. EVA is very objective and robust when many owners contribute its data in calculating.

7 Pre measures to be considered for EVA: i. While doing earned value analysis, we don't take quality into consideration. It may be possible that our project is scoring high on earned value performance scale, but the quality of work is below par. Quality is an important criteria in any project, and unfortunately it is not considered in EVA. ii. If the Project schedule is very large (around x000 hours), 70% of the work schedule is based on the trust on the team members. If there is any deviation in the plan, the effort considered will not holds good. To recover the same, a project manager has to think on the alternative way to achieve. iii. If updating the EVA tool with cost and percentage of completed work is not done on regular basis, hides early problem identification. iv. When the percentage of work assessments is not discussed with team members, the EVA report is often slanted, leading to increased subjectivity. The one doing the work has the best assessment of its status. [2] Quentin W. Fleming & Joel M. Koppleman, Handbook of Earned Value. [3]Quentin W. Fleming Handbook of Cost/Schedule Control Systems Criteria Robert R. Kemps Handbook of Project Performance Measurement [4] Kevin Forsberg, Ph.D., Hal Mooz and Howard Cotterman Handbook of Visualizing Project Management [] Kevin Bury, President & CEO, Quick Arrow Paper on 'Earned Value Management' Websites III. CONCLUSIONS After analyzing through all the scenarios and data, we can come to conclusion: Earned Value is needed because different measures of progress for different type of tasks need to roll up progress of many tasks into an overall project status. Need for a uniform unit of measure (dollars or work-hours). ACKNOWLEDGMENT I would like to acknowledge my gratitude for presenting this paper and thank all my colleagues: Laveen, Pavan, Rafeek, Arunkumar and other managers for their support. REFERENCES [1] Booz, Allen and Hamilton Handbook Tutorial Earned Value Management Tutorial 7

8 8