A Continual Improvement Approach: Closing the Loop in an EPCM environment

Introductionexpected to happen.
In his Quality Handbook Juran predicted thatLesson Learned
“The 20th century will be remembered as theIf the expectations have been met, the solution
Century of Productivity, whereas the 21stshall be standardized and Lesson Learned
century will come to be known as the Century ofDatabase shall be updated with the solution. The
Quality”(Juran, 1999). Maybe the world needsProject learns when someone, other than the
another decade for his prediction to come true.initial learner, adopts and adapts the new learning
Engineering companies in the Oil and Gas industryor prevention (Wieneke, 2008).
have an enormous challenge ahead, in their effortThen, the organization is doing Knowledge
to achieve the desired client success in a dynamicManagement and transforming its own history in a
and competitive environment.list of "Best Practices".
In the effort to become a fact-driven company,Example of a Performance Metric System
they implement Key Performance Indicators (KPI).(KPI)
However, one could finds more than 20 articles ofLet’s continue working with the project called
firms that are using Bar-Charts to monitor and‘SMART’ and the definition of one KPI:
display them in a bulletin board.7- Schedule Performance Index (SPI): Is a
Since the 1920s, Shewhart stressed themeasure of schedule efficiency on a project.
importance of controlling the process. Conversely,An SPI equal to or greater than one indicates a
they overlook the Quality Controls Charts thatfavourable condition: The Project is ahead of
shall be used to determine if a variable orschedule (SPI > 1)
attribute is predictable, stable over time.An SPI less than one indicates an unfavourable
Many professionals used to react to the naturalcondition: The Project is behind of schedule (SPI
variation as though it were a special cause. Even< 1)
an out of specification situation where the goal or(I) SPI = Actual Days / Baseline Days
target is not being met, is not special causeSPI can be calculated also using the formula:
(Breyfogle, 2008). Therefore, there is nothing toSPI=BCWP/BCWS
worry about unless the variable is out of control.Where:
Last but not least, these kinds of improvementBCWP- Budgeted cost of work performed
initiatives lack of a holistic view, the necessaryBCWS- Budgeted cost of work scheduledif
alignment, linkage and replication around theSPI<1 means project is behind schedule
business strategy (Juran, 1999). They only lead toI used the former one in this example because its
firefighting; a reactive culture instead of thesimplicity
prevention that organizations need to assureUsing a color-code in a dashboard leads to
accountability and value added creation acrossfirefighting, creating a finger pointing culture
engineering disciplines and project phases.among the project members (Breyfogle, 2008). It
The paper structure is as follows:is necessary to use a professional quality tool: the
General Objective:individual control chart.
- To develop a continual improvement frameworkIndividual Control Chart (XmR)
that close the loop, translating problems into anIt is important to emphasize that there are
action plan and preventing their recurrencecertain crucial assumptions, which allow the use of
Specific Objectives:this techniques:
1.1. The process is in statistical control.
To design a continual improvement framework2. The distribution of the process considered is
describing its implementation steps.Normal.
2.If these assumptions are not met, the resulting
To develop a Policy Deployment that alignsstatistics may be highly unreliable.
project objectives to client expectations.The results in Excel are similar to a Professional
3.Statistical Software, the variable SPI is in statistical
To design a dashboard performance metricControl.
system (KPI) that lead to root causeTo prove the variable ‘SPI’ follows a
analysis and improvement action plans.normal distribution, the hypothesis is:
Continual Improvement ApproachH0: The SPI sample follows a Normal distribution.
Almost all Engineering Firms in North America holdHa: The SPI sample does not follow a Normal
ISO 9001 Certification, mostly for marketingdistribution.
purposes. Consequently, only a few haveFigure V shows the results of the normality Test
succeeded implementing continual improvement‘Anderson-Darling’. As the computed
principles due to the complexity of projectp-value (0.317) is greater than the significance level
management activities. The following study,alpha=0.05, one should accept the null hypothesis
summarized in the figure I, provides anH0. The risk to reject the null hypothesis H0 while
improvement model for such a complexit is true is = 31.7%. The variable follows a normal
environment.distribution.
The metric is stable. Therefore, the process
Figure I describes a continual improvementcapability index (Cpk) can be determined
approach designed for EPC firms deliveringMinimum accepted capability:
engineering services in the Oil and Gas sector. ItProcess capability attempts to answer the
has nested PDSA cycles, during all the stages.question: can we consistently meet customer
The challenges that engineering firms are facingrequirements?
are immense due to each method of contractingIn a capable process Cpk is 1 or greater. Cpk will
affects in its own way the allocation ofbe higher only when the variable is meeting the
responsibility and the demands on the client sidetarget consistently, with minimum variation.
for coordination and integration of the projectHowever, considering the characteristics of the
flow of knowledge and information in threeEPC industry already mentioned, it is
dimensions: vertically, horizontally, and longitudinally.recommended to accept as a Minimum capability
EPC Continual Improvement Approach – ToCpk = 1. After the organization has applied this
PlanApproach in a whole project phase, it is ready to
The Project Scope of Work (SOW) and Contractreplicate the knowledge, reducing variation and
are among the inputs for starting the Continualincreasing the minimum capability target.
Improvement planning phase. The first step is theThe Figure VII has all the results in only one
development of a Project Policy Deployment (agraph.
Strategic Planning methodology, developed by Dr.The value of Cpk (0.08), in both figure VI and VII
Yoji Akao (1988).means that there is a lot of variation and it is
Engineering companies will find that Policyprobable that the organization could not
Deployment’s greatest strength is its abilityconsistently meet customer requirements in the
to translate qualitative, executive level Projectfuture. Being ‘0.96’ the Lower
goals into quantitative, achievable actions (SeeSpecification Limit for the metric “SPI”, it
Table I, below).is obvious that the performance have been behind
schedule in 5 of the last 8 months (SPI > 1).
Effective Policy Deployment starts with the clientHowever, this is just part of the natural variation.
requirements of the specific Oil & GasTherefore, it means the current process/system
Project that is being designed. It shall start withis incapable of being within specification limits; it
the Mission and Vision statements, and develophas to be redesigned in order to reduce variation.
Quality Policy, Critical Success Factors, objectives
and metrics. A future paper will explain the stepsAnalysis of ‘SPI’ Variation
in the definition of Key Performance Metrics forThe Project Management Institute has found that
an EPC project during its various stages andmature companies have a schedule performance
methods of contracting.index (SPI) variation of 0.08 and a cost
Performance feedback: Individual Control Chartsperformance index (CPI) variation of 0.11. Less
(XmR)mature companies have corresponding values of
The above step creates the foundation for0.16 for both indices.
managing, measuring and monitoring performance,The ‘SPI’ Standard Deviation in the
while the next step (“Feedback” block) is“SMART Project” is 0.0875325 (see figure
the improvement driver where all theV). The ‘SPI’ is close to the industry
project’s proactive initiatives will initiate.average, and the variation is due to common
The objective of this step is to identify when acauses.
metric is out – of – Control, and conduct aConclusions
detail analysis of the special causes of variation to1. The proposed framework closed the
find the Root Cause.improvement loop, integrating lessons learned and
Most of the metrics collected in an EPCreplicating it throughout the whole organization,
engineering firm are 'variables' instead ofwhile achieving the desired alignment with the
'attributes'. The Individual Control Chart – XmRproject strategic deployment.
is the one suitable for this kind of observations2. If applied properly it will solve critical problems,
(conforms to ANSI/ASQC B1-B3 1996).increasing communication between the project
The term special or assignable causes as opposedteam members, and increasing the
to chance or common causes was used bylinkage-integration between disciplines and project
Shewhart (1939) to distinguish between a processphases.
that is in control, with variation due to randomReferences
(chance) causes only, from a process that is outANSI/ASQC B1-B3-1996: Quality Control Chart
of control, with variation that is due to someMethodologies
non-chance or special (assignable) factors (cf.Akao, Yoji, edition (Jap: 1988, Eng: 1991) (in English
Montgomery, 1996, p. 102).(tr. from Japanese)). Hoshin Kanri, policy
The XmR utilizes three-sigma control limits anddeployment for successful TQM. New York:
indicates an out-of-control signal if a single pointProductivity Press (Originally Japanese Standards
falls beyond the control limits. The development ofAssociation). pp. xiii. ISBN 1-56327
the equations for computing the control chart andBreyfogle, F. W. (2008), Integrated Enterprise
its limits can be found out in Juran (1999)Excellence, Vol. III Improvement Project
Problem Definition and Description:Execution: A Management and Black Belt Guide
The best problem statements make nofor Going Beyond Lean Six Sigma and the
assumptions; they simply document the currentBalanced Scorecard, bRIDGEWAY bOOKS,
state.aUSTIN, tx.
Craig Cochran (2006) stated that crafting aConstruction Sector Council. 2006-2010 Alberta
problem statement is one of the most importantConstruction Workforce Supply/Demand Forecast.
steps in problem-solving.May 11, 2006.
EPC Continual Improvement Approach – ToCochran, Craig; (2006).Becoming a Customer
DoFocused Organization, Paton Press
The root cause analysis by means of an IshikawaCrosby, P.B; (1979) Quality is Free, New York:
Diagram will generate several possible solutions.McGraw-Hill.
This phase ends with the Improvement ActionDeming, W.E; (1986) Out of the Crisis, Cambridge,
Plan that will be followed until its final resolution.MA: Massachusetts Institute of Technology,
EPC Continual Improvement Approach – ToCentre for Advanced Engineering Studies.
Study & To ActISO 9001: 2008 - Quality management systems --
Study implies understanding the sources ofRequirements
variation in the process (Common vs. specialJuran, J.M; (1999) and F.M. Gryna. Juran’s
causes). Therefore, it requires executing theQuality Handbook, 5th Edition, New York:
nested PDSA.McGraw-Hill.
During this stage it is necessary to evaluate theMaynard's Industrial Engineering Handbook; (2001)
benefits of the solution (The “Benefits5th Ed, McGraw-Hill, New York, pp. 4.12-4.113
expected” shall be defined in theMontgomery, D. C; (1996). Statistical quality
Improvement Action Plan previously prepared).control (3rd. Edition). New York: Wiley.
Compare what has been found to what is being