Week 13_ SeeGod Meregini_ Determining the Control Limits (variance threshold) in order to timely implement corrective action.

Problem Recognition/Evaluation

The entire process of developing an approved baseline plan and then determining status to measure against that plan has a primary purpose which is to establish a baseline against which “significant” variances can be identified and allow for implementation of corrective action.

In a new project, I was tasked with recommending an appropriate method for determining the variance threshold or control limit which if exceeded required some type of corrective action and documentation.

The root cause analysis for this problem was:

·         Previous similar projects lacked good visibility into current status of work done.

·         Historical data showed a system break down due to inability to process new reporting period variances before the next reporting period appeared.

·         Previous similar projects had too many variances addressed at the same time resulting to ineffective resolution of the variances.

·         An attempt to investigate all variances at the same time for each given reporting period created much documentation with little time to follow up and a lot of these variances poorly addressed.

·        

Development of Feasible Alternatives/Solutions

In solving this problem, examination of six possible methods was used.

Alternative 1- Combination of Absolute Amount and Percentage of Budget/Earned Value

 This method combines the use of both a percentage and an absolute amount threshold. For example, a threshold could be set as $20,000 AND for 200hours assuming $200 per hour AND at least a 25% variance of a control account.

Alternative 2 – Percentage of Budget or Earned Value to Date

This approach lays emphasis on percentages as a variation threshold i.e. if 20% of a controls account is exceeded a variance analysis should be performed.

Alternative 3- Absolute Amount

This approach uses a certain agreed upon absolute value as the threshold.. This could be a $15,000 amount and if exceeded a variance analysis carried out.

Alternative 4- Top 5/10 Approach

This  method uses a philosophy stating that there is a limited number of variances which can be properly investigated, documented and addressed in any reporting period, and whatever that number is, (five, ten , or otherwise), should not be exceeded in that reporting period.

Simply put, if 5 variances are what can be properly addressed at any given reporting period, then the number of variance in any given period should not exceed 5.

Alternative 5-   EVM Approach

This method measures the value of work accomplished in a given period compares it with the planned value of work scheduled for that period and with the actual cost of work accomplished.

This indicates how past performance may affect future performance by objectively measuring project status using weighted milestone and discrete units based on specific criteria defined before work starts.

Alternative 6- Multiple Combinations

This method considers more factors in determining “significant “threshold. Factors considered include the following:

·         Cumulative (lower percentage tolerance i.e. 10% and “looser” absolute value) vs. Incremental (higher percentage tolerance i.e. 20% and “tighter” absolute value)

·         Favourable (“looser”  tolerance i.e. +25%) vs. Unfavourable (“tighter” tolerance i.e. -15%)

·         Milestone Crediting. This means if an agreed upon number of milestones are exceeded, then variance analysis must be performed.

·         Technical Criticality. This means that a tighter threshold is placed for an account with a particular high risk technical work which if delayed would cause ultimate failure or major delay of the overall project.

·         Using the level of a WBS or OBS being analysed. This means that a higher percentage or “looser” tolerance is used for the lowest account level i.e. 20 % for level 4 and a lower percentage or “tighter” tolerance used for higher summary levels i.e. 10% for level 3, 5% for level 2 and so on.

Probable Outcomes of Alternatives / Solutions.

Alternative 1

While this is more complicated than alternative 2 and 3, therby yielding better results, there is still a problem for large budget accounts due to the percentage threshold. I.e. for the example given in this alternative, we must have the 25% deviation regardless of how large the absolute dollar variance.

 Alternative 2

This is a fairly simple method but has an equivalent disadvantage to the Absolute Amount approach. For a very large account, such as $1,000,000, a variance of $190,000 would require no explanation with a 20% threshold as exemplified in this alternative above.

Alternative 3

This is the simplest approach, and variances exceeding a chosen value can quickly be identified regardless of the account. However, it is deficient in recognising the size of the account and the percentage complete progress of work done. This would result in needless variances being explained in some cases and others deserving attention overlooked.

Alternative 4

This method relies on the underlying concept that it is better to address the most significant concerns well than to try to concentrate on a lot of variances poorly. This is can be done using Pareto Chart in order to help the team focus on those concerns that will have the greatest impact if addressed.

Alternative 5

This method imposes the discipline of planning all work in sufficient details so that the cost, technical effort, and schedule dependencies are known at the outset. These detail units are spread across the programme schedule to form the performance measurement baseline, which is used to detect any deviation from plan and immediately give insight into problems and potential impacts from the onset.

Alternative 6

Using the Cumulative and Incremental approach, wide fluctuations in a reporting period could make the data so unstable resulting in more headaches than it solves.

Using the milestone crediting as the earned value technique often leads to meaningless variances if there are a lot of milestones scheduled for a reporting period. Meaningless slippages of a single day will push all of the earned value credit into the next reporting period

Other parameters would lead to more complications in generating more variances leading to unnecessary paper work.

Selection Criteria in determining the solution

1.    Easy computation of realistic thresholds..

2.    Shows good visibility into current status at any given reporting period.

3.    Enables effective follow-up and resolution of the most significant variances.

Analysis and Comparison of the alternatives

 Alternatives 1, 2, and 3 have considerable deficient ways of indicating variances and are poor threshold methods. These methods also lack the ability to make the team correctly focus on the big contributors to the problems in order to resolve the variances.

Alternative 6 adds more work to the team due to its complex process of computation and creates more documentation and unnecessary paper work due to the number of variances that would be generated. It also would result in needless variances being explained in some cases while the more significant ones overlooked.

Alternative 4 is aimed at reducing the quantity of variances being analysed and increase the quality of that analysis including more real time correcting action planning and implementation. The Pareto Chart could be used as a tool to correctly identify the worst five or ten variances having the most impact on the goals of the business and customers. This method is also highly useful for external reporting of variances.

Alternative 5 can be seen to:

Ø  Prevent by:

·         Mitigating problems early because it isolates and detects any variance (cost, schedule to mention a few) by the WBS element from the baseline plan.

·         Creating upfront planning required in producing the system, this often addresses and prevents problems from surfacing later.

·         Reducing the time to deliver the end product because of the opportunity of circumventing costly and untimely rework due to early planning

Ø  Promote by:

·         Utilising Management by exception by directing attention to only the most critical problems thereby reducing information overload.

·          Allowing quick communication of cost and schedule variances relative to the baseline plan, hence, focusing management on most pressing problems first.

·         Allowing early detection and resolution of problems by anticipating what could go wrong, based on past trends.                    

Ø  Manage by:

o    Permitting management to defend and justify decisions based on the objective data obtained from an EVM system and hence determines the best course of action when problems arise.

Selection of Preferred Alternative

Based on the above analysis and criteria, it is recommended that ALTERNATIVE 5 be selected as the option to correctly determine the variance threshold or control limit in order to properly resolve variances.

Performance Monitoring/Post Evaluation

This would be monitored by constantly evaluating

·         The system to see how it effectively resolves variances completely in a given reporting period before the next reporting period.

·         The time it takes to properly investigate, document and address the number of agreed threshold variances in any reporting period.

·         The overall indicators of project success in terms of cost and schedule performance using earned value method at any given period.

References

Sullivan, W. G., Wicks, E.M., & Koelling, C.P. (2009). Engineering economy and design process. In M.J. Horton (Ed.), Engineering economy (14^th ed.) (chapter 1.3) (pp. 27). New Jersey, NJ: Pearson Education, Inc.

Brassard M, & Ritter D. (2010) Pareto Chart. The Memory Jogger 2 (2^nd ed) (pp.122). Salem,NH: GOAL/QPC

Humphreys, G.C (2002) Variance Analysis and Corrective Action. Project management using earned value (chapter 35). (pp.703-710). California, CA: Humphreys & Associates, Inc.

GAO, (2009). Managing Program Costs: Planning. Cost Estimating and Assessment Guide (chapter 18). (pp.202-204).Washington, D.C: GAO