So, we have budgeted costs of a project and established methods to collect data. Now, it is time for finding ways of cost reporting.
Let's look at some known techniques on evaluating performance against plans in terms of costs. We might also attempt to predict the costs in the future.
Cost reporting and analysis is crucial in project management. Accurate and current data allows for corrective actions in a timely manner.
Milestones or no milestones?
One of the most common ways methods is to compare the planned costs and actual costs at the time of completed milestones. This is the easiest way used and take the lest effort by the management. But, what do we do if there is no milestones and more accurate reports are required?
Below figure presents an example of graph how the costs can be plotted in more detail.
All of the cost estimates are included in the curve. The project manager should consult the graph at anytime to monitor and analyse the costs and that it's all is going to plan. In the graph above the actual costs have been recorded in the line. One missing information in the graph, however, to successfully evaluate the project is the actual project progress. Low spending (below the estimate) might indicate poor project performance.
It would be rather wrong to assume that if the costs are going to plan, the project is up to schedule as well.
We should not only should know how much we spent to date but what we should spend to date and what have we achieved so far.
Therefore, milestones monitoring might help to answer those questions.
But, how do we identify milestones?
Milestone is an easily identifiable stage of a project towards completion. All good project management software would allow to identify milestones. Once all activities within milestones are finished, the milestone is achieved.
Milestones activities might be started with 0 or unit duration in within the network. This might involve creating dummies. Milestones also should be achieved by completing packages from the WBS.
For a milestone, we require two sets of data:
- the date on which the milestone is scheduled to be achieved,
- the estimated cost or budget for associated work package.
Once the milestone data is collected the graphs can be plotted. We than match the cumulative costs against the planned achievements dates for those work packages. We must ensure that no estimated costs are left out. Symbols might be added to the budget graph to represent each milestone. Each milestone has the time scheduled for its completion.
Now, we must determine the actual costs. Here we need:
- the date when milestone was actually achieved,
- the costs actually incurred (including committed costs and purchased items) at the end of each cost monitoring period.
The actual costs can be plotted as a graph on the same axes as the budgeted lines. Milestones might have allocated numbers.
Now, the graph has been enriched with the milestone information.
Time to analyse...
Let's say that 13 milestones has been identified for a project in construction industry. Each milestone has been added on the graph with a number.
The milestones and their budgeted-actual costs are listed below.
Actual cost data has also been gathered up to date. The milestones completed have turned into diamonds on the graph.
Now...
If all the planned costs and actual costs were exactly the same and the milestones completed on time - this means that all is going up to the plan.
However, when there is a difference (variation), investigation is needed.
In the graph we know that the milestone 8 was achieved few weeks late. (We can see that by the position of diamond and the circle.) Therefore we know that the project has started late and we need to catch up. The very low cost recorded in week 8 indicates that a very little activity or no activity was taking place.
Important thing to remember here is that without milestones being present analysis would of been rather difficult. This process, however has its limitations. We will now look at another technique, with more derailed evaluations called Earned Value Analysis.
Earned Value Analysis
In order for Earned Value Analysis to work we must have few things in place:
- a detailed WBS,
- correspondingly correct cost coding system,
- timely and accurate collection of cost data,
- method of monitoring work done, including work in progress.
It aims to compare the actual cost for accurately identified amount of work with the costs budgeted for the same amount of work.
The procedure uses Cost Performance Index (CPI). If all is going to plan the CPI os 1.0. Index less than 0.1 indicated that the value is less than expected.
The first most important factor in the process is accurately defined Work Breakdown Structure. The costs of each package in Work Breakdown Structure must provide the framework. It is important that is is carried out to the level of activities performed by each individual manager so each can be quantified by Earned Value Analysis.
As part of the procedures, there are some important definitions to explore.
ACWP- Actual Cost of Work Performed
BCWP - Budgeted Cost of Work Performed on the measured date including work in progress
BCWS - Budgeted Cost of Work Scheduled
CPI - Cost performance Index indicates the measure of success in achieving results against budget
SPI - Schedule Performance Index - measure the progress performance against plans
CPI = BCWP/ACWP
SPI = BCWP/BCWS
Predicting costs
In order to illustrate how this could work, we could imaging a project of building a wall.
We need to built 1000m of wall for the total project cost of £40,000 within 10 weeks in working days (50 days).
This would indicate approx. £800 budget and 20m of wall per day, plus cost of 1 meter of wall of £40.
Lets say we are carrying out Earned Value analysis on a day 20.
We have finished 360m of wall, ACWP: £18,000.
The BCWS for day 20 is 20 X £800 per day = £16,000
The BCWP for day 20 is 360m with £40 per meter = £14,400
Therefore:
CPI = BCWP/ACWP = £14,400/£16,000 = 0.8
Now, we could divide the cost of the project by CPI (£40,000/0.8 = £50,000) which would indicate that the cost pf project increased to £50,000.
But we can also say that £18,000 has been spent to day so we can work out the remaining cost.
360m costs £18,000 = £50 per meter
We have left 640m for £50 per meter indicates another £32,000 cost is remaining.
Predicting schedule
Earned value data can also be used to predict completion date by calculating the SPI.
SPI = BCWP/BCWS = £14,400/£16,000 = 0.9
The original estimate of 50 days can be divided by SPI giving 56 days.
In most projects, however, there is many different activities in different stages.
These can be activities completed, not started and activities in progress.
The activities in progress can be measured by assessing actual quantities of work done or percentage of work done.
Engineering design and IT design projects
When measuring the progress of engineering design activities we can identify how many specifications and drawings are to be produced. Then divide this number by the number of drawings actually issued, multiply by 100 and obtain the % rate. This can be used in department when drawings are similar and there is a lot of them. However, the amount of work to produce one drawing cannot be compared at all times to amount of work needed for another drawing. In this case the best approach would be to ask the engineers for estimate of % of tasks achieved and how much longer the job will take.
In IT design projects it can be quite difficult to estimate tasks for projects required. The engineer might be guilty of optimism or poor estimation. But, better to have any answer than have no answer at all.
We know that work-in-progress will be a lot less accurate than completed activities. But they only account for a small percentage of the tasks, therefore these will diluted with larger and completed work.
Budgets, progress and actual costs can be compared on a sheet such as figure below in predicting the project expenditure.
In the example the total hours budgeted is 2,975 and the earned value BCWP is 1,389. However, against the cost the ACWP is1,142 man-hours. Although 46.7% of work has been done only 38.4% of budget has been spent. This generally means that the work costed less than budget. the CPI is 1.22.
Manufacturing projects
There is no reason for not using the same methods for manufacturing projects to cover production activities.
Production activities, however, often employ different labour grades drawn from several departments. It might be unreasonable to add each possible cost into the analysis. This might involve too much effort and little returns. In production, it might be wiser to analyse the performance of departments rather than individual labour grades. In this way each manager will be presented with their performance feedback against budgets.
The collection of production cost data is important here and need to be aligned with the Work Breakdown Structure. This can be achieved by using structured cost coding system and paying attention to allocation of job numbers. The time should be recorded against each job number which incorporates identifier code enabling the work and its costs to be related to correct project task.
The basis for allocating work values for each production activity derives from the project Work Breakdown together with associated estimates. For the Earned Value Analysis the project manager does not need to consider production task breakdowns in more details than in the Work Breakdown Structure and overall project schedule.
Subcontractors
Earned value analysis might also consider subcontractors activities. However, if the subcontractor work is planned in the fixed-price basis the subcontractor carries the financial risk of earned value measurement for internal control. The project manager needs to pay close attention to the measurement techniques used by subcontractors whenever they submit claim for payment for the progress. The progress payments should be made to measured achievable events or supported by certified measurement.
Earned Value Analysis Reports
Early predictions of final costs always tend to be unreliable. There reasons for these are:
- Estimates of progress are only judgements,
- During the first months of large projects the sample of work analysed is too small to produce valid calculations,
- There is no guarantee that the performance levels at the beginning of a project will remain the same as the remaining tasks,
- Sometime "false claims" are made on tasks presumably 100% completed as they often might need corrections or modifications later on.
Bad predictions
If the Actual Values significantly exceed the Earned Value so that the final result will be in excess of the budget, we should be grateful for these early warning. Stricter control of modifications should avoid unnecessary expenditure. Here the change control and register is imperative. Of course, changes requested by customers will be paid for, but internal changes should be scrutinised through authorisation. Only essential unfunded changes should be allowed.
The demands made to individuals will need to be more strident. This will require all participants to know about the situation. It is important to gain the full co-operation.
The performance of individuals can always be achieved by setting small steps targeted objectives. These should be quantifiable and allow them to monitor their own performance. These objectives should be in correlation to WBS and the whole project. We must not set objectives which cannot be met.
If after all these efforts the project budget is still suffering we might need to re-open the fixed price negotiation with the customer when an opportunity arises.
The excuse might come from customer asking for substantial change or result of economic situation that is beyond the contractor's hand. Through this, smaller additions can be planned in generously to compensate the areas of loss.
We must however remember that the changes cannot be made too late. The project manager should control the spending at all times.
Cost performance for purchased items and materials
Measuring the performance of purchasing materials would be different than labour costs. Purchased orders are generally originated well in advance. The costs of purchase orders are allocated before they are being used.
The purchase order cost control starts when order is placed. If order issued is priced higher than the budget it is too late to avoid overspending.
A curve can be plotted to show the cumulative value of purchase orders as they are placed, as per figure below.
The budget comparison line should be first plotted on the graph. It will be calculated by adding the materials cost estimates for each task and timing them to when the purchase orders to be issued.
We must not forget about any common items which will not be purchased but taken from common company stock. The quantities of these items must be estimated and cost added to the cumulative totals so that all materials used in the project are added.
We can also do it more accurately, by tabulating actual costs against estimates whenever new purchase order is issued. This will allow to see a risk of overspending clearly and allow to review future purchases.
We need data to use this method. It includes:
- the total value of all orders already placed,
- the total estimated value of the purchase orders yet to be placed,
- cost of any materials already issued from general stock,
- estimated cost of any materials still to be used from general stock.
We also need to record any surplus into the total. They might of result from over-ordering or other reasons. Especially, we need to keep record of those materials which can be returned to supplier for full credit.
Project changes and Earned Value Analysis
When change happen, we must think on the effects on cost. At times, customer might be held responsible for additional costs. Changes can be defined as "funded" or "unfunded". Im most cases, changes increase the workload. We need to make corrections by adding modification to the task list with the additional cost needed. The costs are generally rather tricky to estimate. If the changes are not too large or cancel work already marked as achieved in most cases, these can be ignored.
If the changes, however, affected tasks marked as completed, evaluation should be done for every department on the effect on costs. This might cut projects to less man-hours being completed or raising additional purchase order.
Funded changes can be considered as new tasks to be added to the task lists and budget. Customer should pay for any changes and the waste.
Predicting total project costs
Cost predictions can be plotted on graphs for direct comparison with the budget. Combined both man-hours costs and purchased materials. The man-hours should be converted into costs from labour grade.
We know that for some projects the profits become the end objective. The final graph should present the comparison of customs and budget to final selling price.
The graph below shows plotting of cost predictions for whole project.
The costs of materials have been included at the time of commitment. We can see that the second and third point in the prediction line already shows variation. Above week 24 we can see clear overspending. Corrections had been made and the overspent goes down, but than at the end of the project overspent occurs. This might be because additional resources were employed to finish the project on time.
The actual cost line is cumulative adding the costs. We can see how much more information we can gather from the Predicted cost curve. The actual cumulative cost would not show the overspend.
We often present these in spreadsheet forms. One example is presented below. These are often used to produce regular monthly cost reports.
Column A list the main work packages from Work Breakdown Structure. It must include any cost items including miscellaneous items.
Column B should give cost code for each item.
Column C shows the original budget for each item and adds up at the end of the column.
Column D lists budget increments due to customer changes that customer will pay.
Column E adds up column D and C.
We might also add another column to add advanced notes and information of costs "in pipeline" which are waiting for authorisation.
Column F shows the actual costs on the report date including all labour hours converted to standard used rate into project control currency, payments for direct insurance premiums, licences, legal fees, consultants fees, payments made to subcontractors, the cost of already committed materials, the value of other materials for which the purchases have been made including freight, packaging, insurance, duties, and any other costs incurred.
Column G shows the assessed Earned Value for all work performed on the date of evaluation.
Column H calculates CPI for each item from column F and G.
Column J forecast the work remaining by factoring estimates for work remaining by CPI.
Column K presents the best prediction possible of the total projects costs at the completion. As the time pass the predictions become more accurate.
At the end of a project when total costs become known we can compare the actual costs with estimated costs. This might help us avoid mistakes in the future.
Bibliography:
Lock D, 2013, "Project Management", Gower, 10th edition, ch.25
Comments