+44 (0) The first annual earned value conference for continental Europe

Transcription

1 Measuring Integral Earned Schedule (IES) and Predicting the Project's Final Completion Duration: The Application of Kinematics Approach to Earned Value Management Metrics Mojtaba Zarei-Kesheh +44 (0) Presentation for the EVA Europe Geneva 2009 CERN, Geneva - 25Nov2009 The first annual earned value conference for continental Europe

2 Objective Traditional and Current EVM methodologies Introduce the Integral Earned Schedule (IES) Concept Develop the Schedule Performance Indicators and Predictors Comparison of Different Methods 2

3 TIME IS NOT MONEY On the whole complex construction projects are likely to be finished more than six months late, due to poor time control. ( Managing the Risk of Delayed Completion in the 21st Century, CIOB research 2008) Money: if you leave it, it might actually accumulate. Time: expires at a regular and consistent rate whether you use it or not. (Keith Pikavance, President of CIOB, conference 2008). 3

4 CPM or EVA? CPM: The most reliable method to predict the project s time dimension (Fleming & Koppelman) CP analysis as a predictor: too time-consuming & incapable of providing early warning signal due to its retrospective nature. C/SCSC & ANSI/EIA 748: These standards do not define the analysis to be performed on EVM other than to state cost and schedule variances need to be computed. 4

5 Traditional EVM Data Relationship Physical % Complete In current project Activity Duration In project baseline (PMB) Data Date (DD) Performance % Complete BAC Schedule % Complete AC EV PV BAC = Budget at Completion PV = Planned Value EV = Earned Value AC = Actual Cost CV CPI SV SPI CV = Cost Variance CPI = Cost Performance Index SV = Schedule Variance SPI = Schedule Performance Index BASIC INPUT DATA DERIVED DATA 5

6 SV & SPI in Primavera P6! SV = EV - PV 6 SPI = EV / PV

7 Two Different EV Camps THE BUDGET-BASED CAMP (Traditional EV Schedule Metrics: Budget-based) THE TIME-BASED CAMP (21 st Century EV Schedule Metrics: Time-based) SPI( ) = EV / PV SV( ) = EV PV =? At completion (For late finish project): SPI( ) ends up at 1! SV( ) ends up at 0! Schedule Variance in units of budget make no sense! They behave erratically for projects behind schedule. SPI(t) = ES / AT SV(t) = ES AT = Time delay At completion (For late finish project): SPI(t) does not creep toward 1 SV(t) does not creep toward 0 Schedule Variance in time units 7

8 Integral Earned Schedule (IES) Camp Integration of Time & Cost The idea is to identify the point in time where the area under planned value curve (AUPV) equals the current area under earned value curve (AUEV). Integral Earned Schedule (IES) looks at when the current AUEV was to be accomplished. The area under PV and EV curves can be computed using integral approach (trapezoidal rule). 8

9 Integral Approach (Trapezoidal Rule) EV % or PV% EV(ATn-1)% EV(ATn) % ATn-1 ATn Time (Periods) ATn% -ATn-1% 9

10 IES Metrics and Indicators calculation Calculate cumulative values of AUPV% and AUEV% using Integral Approach (Trapezoidal Rule) Now other calculations is similar to ES calculation! In Lipke s ES Calculator: Substitute PVcum with AUPV%cum Substitute EVcum with AUEV%cum Substitute ES with IES Substitute SV(t) with SV(IES) Substitute SPI(t) with SPI(IES) 10

11 Integral Earned Schedule (IES) Formula IES cumulative = Whole time (months or periods) completed for ΣAUEV% Σ AUPV% + Fractional Time Completed (I) IEScum = X + [(ΣAUEV%t Σ AUPV%x) / (Σ AUPV%x+1 Σ AUPV%x)] (x = whole time (month, period,..) earned ; x+1 = month or period following; t = Time Now) Indicators: Schedule Variance (IES): SV(IES) = IES AT Schedule Performance Index (IES): SPI(IES) = IES / AT where: AT: Actual Time 11

12 Prediction Terminology Prediction Method Anbari Jacob Lipke Zarei Planned Value Rate Earned Duration Earned Schedule Integral Earned Schedule P.F. = 1 IEAC(t)ED1 = PD + AT (1 -SPI) IEAC(t)PV1 = PD - (SV/(BAC/PD)) After PCD: IEAC(t)ED1 = AT (2 - SPI) IEAC(t)ES1 = IEAC(t)IES1 = AT + (PD - ES) AT + (PD - IES) IEAC(t)ED2 = PD / SPI P.F. = SPI IEAC(t)PV2 = PD / SPI After PCD: IEAC(t)ED2 = AT / SPI IEAC(t)ES2 = PD / SPI(t) IEAC(t)IES2 = PD / SPI(IES) 12

13 Analogous ways to think of integration of time and money (IES) For labor-intensive projects IES (our integration of time and money) is analogous to estimating work in terms of staff-hours; Thus: the work duration is estimated by the number of staff available More people-less time, less people-more time; So the value of the work is the product. IES: the same approach in EVA analysis; The product of time and money (PV, staff) defines work scope; The product of EV and time is the work completed. 13

14 Real-life Project Data The Real-life Project: Reporting 100% complete as of Jul ,208 million pounds contract Planned Completion Date = Feb09 Actual Completion Date = Jul09 5 Periods Slip 14

15 Schedule Variance Comparison Periods SV(t)cum SV(EAS)cum SV(IES)cum Target SV SV( )cum Zarei Time (Periods) The SV(t) starts to show the real final delay from 97% completion point! Incapability in providing the real final delay in early / middle stage of project. At period 30, or the 50% completion point, the SV(IES) shows the more reliable and consistent SV trend, it correlates very well. SV(IES) predict also the final delay at this point reliably! SV(IES): always normal & consistent trend SV(t): upward and downward trends! inconsistent trend 15

16 Schedule Performance Index Comparison 1.20 SPI(t)cum SPI(EAS)cum SPI(IES)cum SPI( )cum Target SPI Zarei Index Value Time (Periods) SPI(IES): always overall normal & consistent trend SPI(t): always overall abnormal & inconsistent trend: Look at the upward and downward trends! Ooops! So confusing for PM! 16

17 Schedule Variance SV( ): SV(t): SV(IES): Expressed in pounds units; in the vertical direction Expressed in time units in the horizontal direction Expressed in time units Consider both vertical and horizontal direction (integration of time and cost together), then looks at horizontal direction 17

18 Final Duration Predicting Comparison Duration (Periods) P.F. = IEAC(t)PV1 Anbari IEAC(t)EAS1 IES1 Zarei IEAC(t)ED1 Jacob Planned Duration IEAC(t)ES1 Lipke Final Duration Look at the upward and downward trends! So confusing for PM! 48 0% 20% 40% 60% 80% 100% Physical Percent Complete From 30% Completion point, IEAC(IES1): higher accuracy than others, consistent Early Warning Signal behavior / trend, correct final duration Other methods: abnormal, aberrant and inconsistent behavior / trend 18

19 Durat tion (Periods) Final Duration Predicting Comparison P.F. = SPI, SPI(t), SPI(IES) IEAC(t)PV2 Anbari IEAC(t)EAS2 IES2 Zarei IEAC(t)ED2 Jacob Planned Duration Look at the upward and downward trends! No consistency! IEAC(t)ES2 Lipke Final Duration Seeing is believing! 0% 20% 40% 60% 80% 100% Physical Percent Complete From 13% Completion point (early stage), IEAC(IES2): begins to show accurate and reliable prediction; consistent behavior and trend, the most accurate result, correct final duration. This is consistent and reliable EARLY WARNING SIGNAL! Other methods: abnormal, aberrant and inconsistent behavior / trend specially in early stage. 19

20 IES Summary Time & Cost Integration for assessing and measuring schedule performance and predicting final duration; IES: Integration of both time and cost together; Unique characteristics of the area Under Planned Value curve (AUPV) and area under EV curve (AUEV) in Real Projects (not applicable for artificial and unrealistic projects); SV(IES): normal & consistent trend from early stage of project; capability to predict real final delay of project (Early Warning Signal); CONSISTENCY AND ACCURACY IES method: consistent, stabilized and correct predictive behaviour from the early stage of project (Early Warning Signal). 20

21 Project Managers do not need the predictive results at the late stage of project At the late stage, PM cannot take corrective action; everybody is aware of delay IES method claims to predict correct final duration from early stage. Useful info for PMs! That is called: Early Warning Signal! IES method: IES Summary provide consistent results (real final delay, final duration) from early stage; CONSISTENCY AND ACCURACY advance programme control; give opportune time to project planners and PMs to take corrective action. The other methods do not have this capability! 21

22 Thank you to... Mr. Ray Stratton PMP, EVP for his book that motivated me to walk in EVM world; also for his technical advice during my research on EVM/IES; Prof. Mario Vanhoucke for his constant encouragement during my IES research; Mr. Walt Lipke for his technical advice during my work on ES implementation in Thameslink Programme (London) and for use of ES Calculator. 22

23 Thank you to... EVA Europe Geneva 2009, CERN L to R: Kym Henderson, Walt Lipke, Mojtaba Zarei-Kesheh, Stephan Vandevoorde, Prof. Mario Vanhoucke, Prof. Pierre Bonnal 23

24 References Anbari F. (2003) Earned value method and extension. Project Management Journal Fleming, Q. et al. (2005) Earned value project management, 3 rd ed, PMI Jacob, D. (2003) Forecasting project schedule completion with earned value metrics. The Measurable News (March) Jacob, D. et al. (2004) Forecasting project schedule completion with earned value metrics revisited. The Measurable News (Summer) Lipke, W. (2003) Schedule is different. The Measurable News (March) Lipke, W. (2009) Project duration forecasting a comparison of earned value management methods to earned schedule. The Measurable News (Issue 2) PMI-EVM Practice Standard, (2005), PMI Vanhoucke, M. et al. (2006) A simulation and evaluation of earned value metrics to forecast project duration. J Oper Res Soc Vanhoucke, M. et al. (2006) comparison of different project duration forecasting methods using earned value metrics. International Journal of Project Management 24 24

25 25