55r 09 Analysing s Curves

Copyright 2010 AACE International, Inc. AACE International Recommended Practices

AACE International Recommended Practice No. 55R-09

ANALYZING S-CURVES TCM Framework: 10.1 – Project Performance Assessment

Acknowledgments: Jeff Goodman, PSP (Author) Timothy T. Calvey, PE PSP Christopher W. Carson, PSP Andrew Dick Anthony L. Jervis Dennis R. Hanks, PE CCE Donald F. McDonald, Jr. PE CCE PSP Vicente A. Ruiz

Donald E. Parker, PE CCE Hannah E. Schumacher, PSP William James Simons, PSP Jerry L. Vogt, PSP Ronald M. Winter PSP David C. Wolfson James G. Zack, Jr., CFCC


AACE International Recommended Practice No. 55R-09 ANALYZING S-CURVES TCM Framework: 10.1 – Project Performance Assessment

November 10, 2010

PURPOSE This recommended practice (RP) for analyzing S-curves is intended to serve as a guideline, not to establish a standard. As a recommended practice of AACE International, analyzing S-curves provides guidelines for stakeholders of a project to evaluate the current status and trends of a project in a simple graphical format. S-curves are usually developed by a project scheduler or cost engineer and can be applied on a variety of project types. The product is generally used as a project management and/or total cost management (TCM) tool for graphic representation of project performance. The RP provides descriptions of S-curves with the intent to improve understanding and communication among project participants and stakeholders when preparing and analyzing graphics based upon project schedule information. The RP describes different types of S-curves that may be generated from a schedule provided the proper information is loaded into the schedule and the status of the information is maintained throughout the duration of the project. OVERVIEW An S-curve is a graphic display of cumulative costs, labor hours, progress, or other quantities plotted against time. The term derives from S-like shape of the curve, flatter at the beginning and end and steeper in the middle, which is typical of most complex projects. Most projects start slowly, accelerate throughout the majority of the work and then slow down again near the end as productivity declines and work runs out. The term S-curve can also be used to indicate an S shaped chart resulting from a cumulative likelihood distribution. In this function, an S-curve is a tool of quantitative risk analysis which project management would use to determine the possible dangers of any given course of action. S-curves are also called, “cumulative distribution charts,” “velocity diagrams,” and “SPLOTS” (S-plots.) RECOMMENDED PRACTICE AACE recommends that schedulers and other project team members develop and use S-curves to plan, monitor, analyze, forecast and control project progress. Project managers should request that project controls personnel produce and use this graphical technique as a tool for briefing stakeholders on project status and trends in a quick and intuitive manner. DEVELOPING S-CURVES Prior to developing an S-curve, a project baseline schedule needs to be developed. The baseline schedule should employ best scheduling practices (i.e. documentation of scheduling basis, start, finish, no open ends, minimal constraints, a defined critical path, etc.). The baseline schedule should also contain cost and/or quantity data information if that type of S-curve is desired. The S-curve produced from the baseline early dates is often referred to as the “target S-curve” which reflects projected or planned progress on the project if all tasks are completed on their original early finish dates. This curve represents the “best” progress or productivity that can be expected. It is important that the cost and/or quantity information loaded into the activities represents the appropriate time-scaled values for those activities.


Analyzing S-Curves

November 10, 2010

2 of 14

For example, if electrical service equipment will be delivered in one large shipment and then installed over a three month time period, the large cost of the equipment will be earned when delivered, so that cost should be loaded into the delivery activity, and the balance of the costs can be loaded into the installation activity. When loading costs into activities, the activities should reflect a straight-line consumption of those costs (a ten day activity with a $10,000 cost should reflect installation of $1,000 per day). This methodology of loading costs into the schedule will result in reasonable and appropriate S-curves. S-curves were generated traditionally within the C/SCSC (Cost/Schedule Controls Systems Criteria) process that evolved into the present Earned Value Management System. 1. Common S-Curves: A variety of S-curves exist, the most common being man hours versus time, and costs versus time. While the S-curves generated using any quantities versus time that are useful for comparison, such as volume of concrete or linear footage of pipe, the S-curves generated using cost versus time may be helpful in developing the projects overall cash flow. The time unit used is typically monthly to coincide with normal monthly project status updates. Weekly and even daily time units are also used. The smaller the time unit between readings, the “smoother” the S-curve will be and easier to use for forecasting trends. The shape of the S-curve may yield significant information about the nature of the project. Curves that start up steeply and then flatten out may indicate front-loading, which could be projects that require little or no planning time, repair of disaster damage, or they could indicate a schedule that has been accelerated from the beginning. Curves that start out with a lower slope and then run steeply to completion could indicate a large planning or design time in the beginning and a reduced construction time. If the project is primarily labor, generally those curves are typically back-loaded, initially flat and increasingly steep towards the end of the project. Projects where the costs are mostly installed materials and labor tend to show a fairly linear distribution of costs over time. Mobilization costs and deposits will cause more front-loading of the curve. 1.1. Progress S-Curves: After creating a baseline schedule, a baseline S-curve should be generated. Baseline S-curves provide the basis on which to compare a project's actual status to its planned progress. There are two types of comparisons that can be developed (for simplicity’s sake Figure 1 and 2 below reference hours but, just as easily could be replaced with costs):

Figure 1 – Target Plan Man Hours vs. Actual Man Hours Expenditure


Analyzing S-Curves

November 10, 2010

3 of 14

Figure 2 – Target Plan Man Hours vs. Earned Man Hours (Based on Plan) The two comparisons can be combined with a target or earned vs. actual comparison which can provide insight as to manpower and financial resources required to complete the project. Schedules with status applied to them are called updated schedules. One can produce the same type of S-curves as produced with the baseline schedule to track actual progress and to forecast upcoming progress. S-curves produced from update schedules are typically similar but different than those produced from the baseline schedule in that they also display curves derived from earned and actual data. The baseline schedule provides target progress, typically in costs or man hours. The update schedule may provide three basic metrics; the value of the work that was planned to be achieved at the time of the update, the actual value of the work achieved at the time of the update and the earned progress typically shows earned costs or earned man hours at the time of the update. Payrolls or other man hour reports provide actual man hours expenditures. Updated project schedules and payrolls/timecards provide the actual data that is then compared to the baseline S-curves. This allows the progress of a project to be monitored and quickly reveals any divergence from the baseline schedule. S-curves may also be used to depict project growth, slippage, and progress. The two values that are generated in updated schedules are different in that the earned value is derived from the accumulation of the percent complete of the individual activities times their planned quantities or costs, and the actual value is derived from either manual entry of actual job costs or the calculations of the completed activities’ values along with the calculation based on the estimates of remaining durations of those activities. Obviously, if the project management team can provide actual job values or quantities to be entered into the system, the actual curves will be accurate and valuable. Without actual real-time job data, the software used generally defaults to its own algorithms for the calculations of completion in the schedule. If the remaining duration and the percent complete components of the schedule are linked such that one calculates from the other, then the two available curves will be identical and offer no separate analysis ability. For this reason, it is important that the schedule software is set to calculate percent complete and remaining duration separately, allowing time and values to be represented in the curves. When discussing cost loaded schedules, the actual costs curves are generated from percent complete of activities (hopefully based on quantities or costs) and the earned value curves are generated from remaining duration calculations based on time.


Analyzing S-Curves

November 10, 2010

4 of 14

1.2. S-Curves with Early and Late Dates: Most scheduling software can provide the information that calculates the cost, hours, and/or quantity data in the schedule over the both the early dates (forward pass of the schedule) and the late dates (backward pass of the schedule). These are sometimes referred to as “banana curves”. The two curves will typically only overlap at the beginning and end of the project. The envelope produced by these two curves represents the range of possibilities that the project can expect if it is to be delivered on time. It is important to note that the lower range of the banana curve, the late date curve, is the prediction of progress where every single activity is performed on the last available day and all work is on the critical path. This means that if any single activity on the late date curve slips, the project slips. Stakeholders should be concerned about performance risk when a project planned or actual curve is tracking close to the late date curve. This information can then easily be put in a graphical format as shown in figure 3 below:

Figure 3 – Typical S-Curve with Early and Late Dates 1.3. Quantity S-Curve: Another common useful type of S-curve found in the manufacturing and construction industries is the quantity S-curve. It uses the production schedule and graphs the quantity versus schedule time. Here the quantity information loaded into the baseline schedule which is tracked along with actual quantities. This allows the progress of a project to be monitored, and quickly reveal any divergence from the baseline schedule. These production S-curves may also be used to depict project growth, slippage, and progress to date. For comparative purposes one could plot planned versus actual quantities.


Analyzing S-Curves

November 10, 2010

5 of 14

Figure 4 – Quantity Output S-Curve 1.4. Cash Flow S-Curves: Cash flow is the movement and timing of cash with respect to the events in a project. Depending on the ultimate purpose, constructing the cash flow curve has various requirements. The stakeholders of the project have various interests in the curve. One primary use of developing the cash flow curve is to see need for cash, and the timing of the payment obligations. It is significant to clearly understand that without good financial health there would not be a project. On a fixed price contract, stakeholders are concerned with the most complex timing considerations; the work actually performed or earned as well as related expenditures and revenue payments. Each of these may have multiple timing considerations. Other contract types have some, but not all of these considerations. Thus, in order to produce a realistic useful cash flow curve must consider the timing of each cost element. If the estimate classified cost by cost type (labor, material, equipment, subcontracts, etc.] and the cost is loaded into schedule as resources in the same categories, then the early/late curves can be graphed by cost type. Obviously, the S-curve on which all these calculations are based should be the optimal resource-based curve, on the median of the early and late curves. 1.5. Other Specific S-Curves: Depending on the data and information loaded into schedule activities there are numerous types of S-curves the can be developed. The following is a brief list of common types of S-curves: • Resources • Manpower • Installed quantities • WBS


Analyzing S-Curves

November 10, 2010

6 of 14

PROGRESSING S-CURVES Once the baseline S-curve is graphed and actual progress status information is obtained for the project, the actual progress can be graphed along with the baseline or target S-curve. The progress S-curve with the baseline/target S-curve provides stakeholders with information on variances and trends. Earned value and performance information may also be generated and displayed on the common S-curve graph. The variances in cost/quantity and schedule can readily be seen and by analyzing the results relative to the baseline/target S-curve, forecasts can be made of anticipated variations at completion. S- CURVE ANALYSIS S-curves based upon actual data can be used to monitor the project as it progresses and comparison of actual progress to the projected S-curve helps to determine whether the project will be completed within the time and budget limitations. There are several methods of analyzing the S-curve. The methods will be discussed in this RP are applying actual data, applying earned value, applying earned schedule and applying common shapes of early and late dates to the S-curve. The analysis method used may depend on the complexity of the project and contractual requirements. 1. S-Curve Analysis – Applying Actual data: The easiest S-curve analysis method to use is applying actual data to the S-curve. This provides an easily understood view of project status. Below are two examples of actual data applied to S-curves along with a brief analysis for the curve.

Figure 5 – Quantity Output S-Curve


Analyzing S-Curves

November 10, 2010

7 of 14

The above S-Curve suggests that since the actual data is below the planned value the project is behind schedule. Any given actual quantity figure is only reached later than it was planned to be. Conversely, if the actual data line was above the planned line, the graphic would indicate the project is currently tracking ahead of schedule.

Figure 6 – Early/Late Date S-Curve with Actual Progress In the above example, the actual data is plotted along with the early and late date information from the schedule. The analysis of this graph indicates that as long as the actual data falls between the early and late dates and the forecasted completion is on or before the end date, the project is in good standing. This plot does not guarantee project success; it only indicates that the actual work appears to be within planned boundaries. Again, note that the closer that actual data conforms to the late dates, the higher the risk of project delay, especially with resource limitations compared to the resource logic used in the schedule. A project that is tracking along the late dates in the first 85% of the project has a high likelihood of slipping past the predicted completion and should be analyzed carefully for the resource requirements to complete the project. 2. S-Curve Analysis – Earned Value: Another common method of analysis of the S-curve for a project is with earned value. Earned value management is a method of integrating scope, schedule, and resources into a discrete set of numbers and then objectively measuring the project performance and progress using these values. Below are two examples of applying actual and earned data to S-curves along with a brief analysis of the curve.


Analyzing S-Curves

November 10, 2010

8 of 14

Figure 7 – Earned Value S-Curve In figure 7 above, the “planned” line is the performance baseline from the schedule and is considered the benchmark S-curve against which actual performance is analyzed. The figure above shows the project currently anticipates an early finish but with additional costs or quantities from the performance baseline. This information should be evaluated with the project stakeholders to determine if this is the best course for the project. Figure 8 depicts a hypothetical situation where a known problem and its projected impact are being considered.


Analyzing S-Curves

November 10, 2010

9 of 14

Figure 8 – Earned Value S-Curve In the above graph, the “planned” line is the performance baseline from the schedule and is considered the primary S-curve being analyzed. The graph above shows the project currently anticipates a cost or quantity overrun, together with possible corrective action(s) as well as finishing beyond the original planned finish date. The reason for the delay and forecast overrun should be brought to the project stakeholders to determine the best course of action for the project. In order for the project to regain lost time in order to complete on time, it will be necessary for the performance to be strong enough to allow the S-curve to change slope enough to meet the completion milestone. This can be plotted and used to determine progress necessary to complete on time, or the S-curve for recovery schedules that are produced can be reviewed to see that the curves are sufficient. It is also possible to separate out individual trades to graph earned value curves and that can be very useful in analyzing the performance of those trades as well as predicting the impact of the trade’s performance on the project.


Analyzing S-Curves

November 10, 2010

10 of 14

Figure 9 – Earned Value S-Curve basic analysis Figure 9 identifies from a graphical view some of the types of analysis that can be performed using earned value. For the purpose of this RP, earned value terms will be introduced and a high level of variance analysis will be identified. The axis are defined as X=Time and Y=Cost/Quantity

Budgeted Cost of Work Scheduled (BCWS)/Planned Value (PV) – The sum of all budgets for work scheduled to be accomplished within a given time period. Budget at Completion (BAC) – The total authorised budget for accomplishing the project scope. Actual Cost of Work Performed (ACWP)/Actual Cost (AC) – The costs actually incurred in accomplishing the work performed. Budgeted Cost of Work Performed (BCWP)/Earned Value (EV) – The value of the completed work expressed in terms of the budget assigned to that work. Estimate to Complete (ETC) – The expected remaining cost needed to complete an activity, group of activities or the project. Estimate at Completion (EAC) – The expected total cost of an activity, group of activities or the project. Generally AC+ETC Schedule Variance (SV) – The time difference between the planned and earned schedule - expressed as a time variance (days, weeks, months, years). SV indicates how much ahead or behind the project schedule is compared to the planned (Baseline). Generally a positive value indicates a favourable position and a negative value indicates an unfavourable position. Cost Variance (CV) – The difference between the EV and the PV expressed in dollars, generally a positive value indicates a favourable position and a negative value indicates an unfavourable position


Analyzing S-Curves

November 10, 2010

11 of 14

Variance at Completion (VAC) – The difference between the BAC and the EAC. A positive value indicates a favourable position and a negative value indicates an unfavourable position.

The BCWS (PV), BCWP (EV), and ACWP (AC) are graphically presented on an S-curve to identify variances and present the status of a project. 3. S-Curve Analysis – Earned Schedule: The concept of “earned schedule” is one where earned value is converted to progress in time along the projected baseline progress curve. This computation indicates what the date should be if earned progress were proceeding exactly according to plan. The comparison of the earned progress date and the status date will reveal how many days the project is behind or ahead.

Figure 10 – Earned Schedule S-Curve The “planned” line is the performance baseline from the schedule and is considered the benchmark S-curve against which actual performance is analyzed. The “time variance” between the “planned” and “earned” determines the forecasted completion date for the project. 6. S-Curve Analysis – Early and Late Dates: There are many shapes that “banana” curves may take depending on the project. Three examples are provided below. While any of these shapes may be acceptable depending on the project, the use of historical S-curves from similar projects will help validate the S-curve shape.


Analyzing S-Curves

November 10, 2010

12 of 14

Figure 11 – Example A

The banana S-curve shown above in Example A may indicate that there are excessive float values on the activities at the end of the project schedule. This could result from the logic ties particularly at the end project and these activities may require some analysis. The early date target line reaches completion well before the late finish date target line. This banana curve may indicate that the project scope is not required to finish anywhere near this early. So one may look at the underlying logic and ask the question, “Can this work be allowed to occur this soon as the result of the early completion of this scope?” This banana S-curve may be fine for the particular project or it may indicate something about the underlying schedule logic of the project.

Figure 12 – Example B


Analyzing S-Curves

November 10, 2010

13 of 14

The banana S-curve shown above in Example B is another indicator of potential problems, usually caused by excessive float due too relationship ties of logically staggered activates are linked into a common successor. The effect is a nice looking early target curve but a bloated looking late target curve. An obvious legitimate reason for this is when you have a detailed engineering schedule but the construction schedule is not yet fully developed. Also, there are two key points of analysis from the banana S-curve in Example B that may need further examination. First, there is potentially some excessive float in the schedule that goes against best practices which may contribute to the shape of the S-curve, but more importantly, this indicates that the late activities may not be resource leveled. This results in schedule activities sliding to the far right as possible (as shown in the late curve above). The project will then have too much work stacked up at one time at the end of the schedule and may become impossible from a resource perspective to complete. These types of banana curves at the project summary level may indicate that the schedule is not yet logically complete.

Figure 13 – Example C The banana S-curve in Example C tells us that the schedule is very tight (little activity float). This may not be reasonable unless it is being caused by excessive use of contingency, or a limited number of parallel activities. There are two potential methods to further analyze S-curves. First, is to consider internal rules of thumb for what percentage of activities in a schedule should have varying degrees of criticality (this will vary from industry to industry and project to project). The second is to review historical curves to see how close similar projects’ S-curves compare. 7. S-Curve Analysis – Cautions: It is possible for S-curve information to be misleading instead of instructive. For example, if the project is tracking costs and has pre-ordered several high-ticket items, then the cost S-curve may show excellent progress regardless of the current status of the critical work. In another example, poor productivity will also show a high cost S-curve while schedule progress will be lower than expected. Progress S-curves indicate cumulative progress of all work, not just critical work. The project may be


Analyzing S-Curves

November 10, 2010

14 of 14

behind due to over-emphasis on non-critical work at the expense of the critical work and still show excellent S-curve progress. The S-curve by itself and without any other project information (project status narrative, detail schedule analysis, etc.) may provide a false view of the project. Used in conjunction with other sound project controls practices, S-curves can provide project stakeholders quick and meaningful indications of the overall progress of the project. SUMMARY There are several different methods to generate and analyze an S-curve for a project. The complexity of the project, the project contract structure and value of the project are just a few factors in determining the best type of S-curve that can be utilized on a particular project. The S-curve together with a CPM schedule is a tool that may be used by the stakeholders to status, evaluate and forecast the project. REFERENCES 1. Hollmann, John K., PE CCE, Editor, Total Cost Management Framework: An Integrated Approach to

Portfolio, Program and Project Management, AACE International, Morgantown, WV, 2006. 2. AACE International Recommended Practice 10S-90, Cost Engineering Terminology, AACE

International, Morgantown, WV, (latest revision). 3. Project Management Institute, A Guide to the Project Management Body of Knowledge (PMBOK®

Guide), 4th Edition, Project Management Institute, Newtown Square, PA, 2008 4. Kerzner, Harold, Project Management: A Systems Approach to Planning, Scheduling, and

Controlling, 10th Edition, Wiley, Hoboken, NJ, 2009 5. O’Brien, James J., Plotnick, Fredric L., CPM in Construction Management, 6th Edition, McGraw-Hill,

New York, NY, 2006 6. Hutchings, Jonathan F., CPM Construction Scheduler’s Manual, McGraw-Hill, New York, NY, 1996 7. Lipke, Walter H. and Kym Henderson, Earned Schedule: An Emerging Enhancement to Earned Value

Management, CrossTalk - The Journal of Defense Software Engineering, Hill AFB, UT, November 2006

8. Midori Media, The Mysterious S Curve, PROJECTmagazine, 2003 9. AACE International Recommended Practice 38R-06, Documenting the Schedule Basis, AACE

International, Morgantown, WV, (latest revision). CONTRIBUTORS Jeff Goodman, PSP (Author) Timothy T. Calvey, PE PSP Christopher W. Carson, PSP Andrew Dick Anthony L. Jervis Dennis R. Hanks, PE CCE Donald F. McDonald, Jr. PE CCE PSP Vicente A. Ruiz Donald E. Parker, PE CCE Hannah E. Schumacher, PSP William James Simons, PSP Jerry L. Vogt, PSP Ronald M. Winter PSP David C. Wolfson James G. Zack, Jr., CFCC

55r 09 Analysing s Curves

Documents

Transcript of 55r 09 Analysing s Curves