Professional Documents
Culture Documents
Every
Lean Six Sigma Belt
Should Learn
In Their Training
ForrestW.BreyfogleIII
Forrest@SmarterSolutions.com
SmarterSolutions,Inc.
www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
Lean Six Sigma training can impact how one looks at data and how an organization makes
improvements. However, there are differences between what is taught in various Lean Six Sigma
trainingofferings.LeanSixSigmatraineesbenefitwhenthefollowingareaddressedintheirtraining:
WefindthatmostLeanSixSigmastudentshavebeentaughttobaselinetheirprocessesaspartofthe
define or measure phases in a Lean Six Sigma DefineMeasureAnalyzeImproveControl (DMAIC)
projectexecution roadmap. Practitioners tell us that they often skip this step because existing data
appeartobeoutofcontrolornotinaformthatiseasilycharted.
Typical Lean Six Sigma training involves Statistical Process Control (SPC) charting along with process
capabilitystatements.Withthisinstruction,thestudentistaughttotestforprocessstability,andifthe
processisstable,oneistoreporthowtheprocessisperforming.Forprocessesthathaveacontinuous
response,theprocesscapabilityisthentobereportedinprocesscapabilityindicessuchasCp,Cpk,Pp,
andPpk.
However,thisapproachcanhaveissuesthatleadmanystudentstoskippingthisstep.Foronething,
process capability indices can be difficult to understand and explain to others. In addition, process
capabilityindicescanbedependentuponhowoneselectstosamplefromtheprocess.Theotherthing
that often occurs with processmetric reporting is that an xbar and R chart, along with a pchart, c
chart, or uchart, indicates that the process is not stable, when stability actually exists. Unlike an
individualscontrolchart,commonSPCchartssuchasXbarandRcharts,pcharts,ccharts,anducharts
donotconsidervariabilitybetweensubgroupsasasourceofcommoncausevariability.
This charting difference is a big deal when establishing a process baseline, since betweensubgroup
variabilityistypicallyasourceofcommoncausevariability.Thereasonformakingthisstatementisthat
withbaseliningwearenotreallytryingtocontrolaprocessbetweensubgroupvariability,whichisa
primaryobjectiveforusingxbarandRcharts,pcharts,ccharts,anducharts.
SmarterSolutions,Inc. 2 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
Anotherimportantaspectof30,000footlevelreportingisratherthanareportusingprocesscapability
indices,whichcanbeafunctionoftheprocedureforsampling aprocess,theestimated percent non
conformanceisreported,whentheprocessisstable,inaformatthatcaneasilybeunderstood.
With30,000footlevelreporting,regionsofprocessstabilitycanbedetermined.Whenthereisarecent
region of stability, one can state that the process is predictable. With this form of reporting, if the
processhasaspecification,thenanestimatednonconformanceratecanbereported,asillustratedin
Figure 1 where there are specification limits of 80 100. If there is no specification (e.g., lead time,
inventory,cycletime,andprofitmargins),thenamedianresponsecanbereportedwith80%frequency
ofoccurrence,asillustratedinFigure2.
SmarterSolutions,Inc. 3 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
100 7.5
_
X=96.95 _
5.0 X=4.75
95
2.5
90 LCL=89.89
28 67
1 35 22 15 09 00
6 42 59 06
7
0.0 LCL=0.00
.8 0. .2 .5 .8 .6 6. .2 .8 8.
97 10 91 96 93 95 10 94 91 10 97.828 100.671 91.235 96.522 93.815 95.609 106.006 94.242 91.859 108.067
Subgroup Period Subgroup Period
Probability plot
Normal
Mean 96.95
99 StDev 5.016
N 150
90 AD 0.335
72.873 P-Value 0.506
Percent
50
10
100
0.036
80
0.01
80 90 100 110
Data values
Figure1:30,000footlevelperformancereportingillustrationwhenaspecificationexists.
SmarterSolutions,Inc. 4 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
100 7.5
_
X=96.95 _
5.0 X=4.75
95
2.5
90 LCL=89.89
28 67
1 35 22 15 09 00
6 42 59 06
7
0.0 LCL=0.00
.8 0. .2 .5 .8 .6 6. .2 .8 8.
97 10 91 96 93 95 10 94 91 10 97.828 100.671 91.235 96.522 93.815 95.609 106.006 94.242 91.859 108.067
Subgroup Period Subgroup Period
Probability plot
Normal
99.9
Mean 96.95
99 StDev 5.016
N 150
90 90 AD 0.335
P-Value 0.506
Percent
50 50
10 10
103.37
90.52
96.95
1
0.1
80 90 100 110
Data values
Figure2:30,000footlevelperformancereportingillustrationwhennospecificationexists.
Onlydatafromtherecentregionofprocessstabilityshouldbeusedtotestvarioushypothesesofwhat
might be done to improve a process; e.g., differences between departments, personnel, days of the
week,andtimeofday.
References:
XbarandRControlChart:IssuesandResolution
ProcessCapabilityCp,Cpk,Pp,PpkIssuesandResolution
SmarterSolutions,Inc. 5 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
30,000footlevel Performance Metric Reporting, Forrest W. Breyfogle III, Six Sigma Forum
Magazine,February2014,pages1832.
Chapters 12 and 13: Integrated Enterprise Excellence, Volume III: Improvement Project
Execution,ForrestW.BreyfogleIII,CitiusPublishing,2008.
But too often, organizations might execute a kaizen event or some other total quality management
(TQM) improvement effort without the result of these efforts being reported with a true statistical
analysis.Youmayjustseeareportwrittenwithtestimonialsoftheeffortsbenefit.Withallofthetools
we have been taught, why do many improvements finish with a qualitative assessment of the benefit
ratherthanaquantitativemeasure?
The best practices would be to monitor the process response over time to ensure that the improved
process performance remains at its improved level with a proper process control chart, such as the
30,000footlevelmethoddescribedabove.Whenstabilityisobserved,thenahypothesistestshouldbe
performedtoestimatethesignificanceofthechange.WhenIseeabusinessimprovementprojectavoid
these two steps, as a manager, I question if the improvement only feels good and does not actually
improveperformance.
Figure3sresponseafterachangecouldbecomparedtoFigure1sresponsetovisuallyrepresentthe
change along with recognition that the new process performance is consistent and predictable. This
predictability assessment would allow a new capability assessment to be estimated and a hypothesis
testtobecompletedtoshowthesignificanceofthechange.
SmarterSolutions,Inc. 6 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
95 5.0 _
_
X=92.79 X=3.83
90 2.5
0.0
85
LCL=83.67 LCL=-1.14
1 5 9 13 17 21 25 29 33 37 1 5 9 13 17 21 25 29 33 37
Subgroup Period Subgroup Period
Probability plot
Normal
99
Mean 92.79
95.488 StDev 4.258
90 N 50
AD 0.224
P-Value 0.816
Percent
50
10
100
80
0.1 0.134
80 85 90 95 100
Data values in current stage
Reference:
Chapter37:IntegratedEnterpriseExcellence,VolumeIII:ImprovementProjectExecution,ForrestW.
BreyfogleIII,CitiusPublishing,2008
SmarterSolutions,Inc. 7 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
I am sure you have seen many improvement projects about which no one seems to care if they are
completedsuccessfully.Someorganizationswillallownonimportantprojectstobeusedforatraining
project,butIconsideritirresponsibletoallowabusinesstoexecuteanimprovementeffortthatwillnot
provideameasurableimpactontheenterprise.Ifanorganizationisgoingtoassignresourcestosolvea
problem, there should be a management expectation that the cost of the effort is recovered at the
enterpriselevel.
Ourpreferredwaytoensurethateveryimprovementprojectislinkedtoanenterprisegainusesaform
ofafaulttreeoradecisiontreediagram.IntheIntegratedEnterpriseExcellence(IEE)LeanSixSigma
system,wecallthistheEnterpriseImprovementPlan(EIP),asillustratedinfigure4.ThisEIPdiagram
demonstrates a clear linkage between the existing organizational goals and strategies to all
improvement effort, even nonsixsigma efforts. Using this tool to derive the needed improvement
effortsfromtheacknowledgedbusinessgoalsandalignedstrategiesleadstoauniqueconditionwhere
thebusinessmeasurementimprovementneedspullsforimprovementprojectstomeetthesegoals,
ratherthanthecommonsituationwheretheimprovementteamshuntforprojectstowork.
SmarterSolutions,Inc. 8 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
Figure4:AnEnterpriseImprovementPlan(EIP)
References:
ProjectSelectionwithWholeenterpriseBenefit
Chapter12:IntegratedEnterpriseExcellence,VolumeII:BusinessDeployment,ForrestW.
BreyfogleIII,CitiusPublishing,2008.
SmarterSolutions,Inc. 9 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
Figure9.3:TheBusinessProcessManagementGuidebook:AnIntegratedEnterpriseExcellence
BPMSystem,ForrestW.BreyfogleIII,CitiusPublishing,2013.
There are DOE uses that benefit data collection efforts. One simple use is to create a data collection
plan for a project that includes the potential causes as twolevel factors. Each DOE observation
becomesa targeted datacollectionevent. Toaccomplish this, onecouldmodifyoperationstocreate
thesetofconditionsneededintheexperimentorevenadjusttheworksequenceinordertocreatethe
experimentallevels.WithafractionalfactorialDOE,onecouldgainvaluableinsightabouttheeffectsof
upto15twolevelfactorsinonly16trials.
AnothergreatuseiswhatwecallahistoricalDOE.Theuseofthistechniqueisvaluablewhenthereisa
verylargedatabasetoanalyzeandonewishestosimplifytheanalysis.OnewouldcreatethisDOEasin
the previous example and then utilize a database administrator to submit SQL queries that pull data
fromthelargedatabasethatmatchtheplannedDOEobservations.
The most common missing DOE concept in Lean Six Sigma training is that a DOE can be set up for
assessing an attribute or even a reliability response. Traditional DOE requires a continuous response,
butwithagoodunderstandingofDOE,thepractitionerwillfindthathe/shecanusetheDOEtodefine
the data collection and then analyze the data with any tool, such as logistic regression (for attribute
responses) and life regression (for reliabilitytypecensored data). He/she can even analyze the DOE
datawithregressionorANOVAifdesired.
References:
DataCollectionTools:Surveys,InfrequentSampling,DesignofExperiments(DOE)
MinimumSampleSize,isitrelevant?
Chapters2934:IntegratedEnterpriseExcellence,VolumeIII:ImprovementProjectExecution,
ForrestW.BreyfogleIII,CitiusPublishing,2008.
SmarterSolutions,Inc. 10 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
Many Lean Six Sigma programs will rush their students through the improve phase talking primarily
about testing and design of experiments with an assumption that the organization will accept the
change.Thisisnotbeneficialtothestudents.IntheearlydaysofSixSigma,beforetheGErollout,Six
SigmaBlackBeltsweretopickuptheiranalysisandprojectmanagementskillsontheirown,butthey
wereallrequiredtotakeaoneweekcourseinchangemanagementbecausethatwasrecognizedasthe
weakestpartoftheSixSigmamodel.ItistoobadthatmanyLeanSixSigmaprogramshaveforgotten
thislesson.
SmarterSolutionsrecognizesthatwithoutfocusingthestudentsonchangeachievement,manystudents
will become frustrated and not achieve their goals. We believe that change needs to have a
demonstratedbusinesssignificancethatalsohasanemotionalimpact.WebelievethateveryLeanSix
Sigmastudentshouldlearnatraditionalchangemanagementmodel.Kotterprovidesoneofthebest
modelsfortheresistancetochange.Kotterliststhefollowingeightstepsforchange:
1. Establishingasenseofurgency
2. Formingapowerfulguidingcoalition
3. Creatingavision
4. Communicatingthevision
5. Empoweringotherstoactonthevision
6. Planningforandcreatingshorttermwins
7. Consolidatingimprovementsandproducingstillmorechange
8. Institutionalizingnewapproaches
Byfollowingtheseeightsteps,nearlyeveryLeanSixSigmastudentcansuccessfullyimplementchanges.
In our Master Black Belt courses, we go even further to address the adoption of business changes
through the introduction of advocacy selling techniques, as described in John Dalys book, Advocacy.
SmarterSolutions,Inc. 11 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
Wealsofindthatthereisnotenoughemphasisgivenintrainingontheuseofpilottestingtoassessthe
impactofproposedchanges.
References:
LeadingChange,JohnP.Kotter,HarvardBusinessReviewPress,2012.
Advocacy:ChampioningIdeasandInfluencingOthers,JohnA.Daly,YaleUniversityPress,2011.
Summary
TakingLeanSixSigmatrainingorobtainingcertificationasaLeanSixSigmabeltisnotsufficienttobea
successfulpractitioner.BeingabletoeffectivelyapplytheabovefivetopicswillalloweverytrainedLean
SixSigmapractitionertobesuccessfulnotonlyintheirimprovementprojectsbutalsoinotherareasof
theirjob:
1. Baseliningskillscanbeusedatanytimeyouwanttounderstandhowaprocessisperforming.
2. Understanding how to demonstrate a change is a key tool to show the benefit from an
improvement.Youmayalsofindthisskillvaluabletoevaluatechangesthatwerenotsuccessful
inallowingtheimprovementefforttobeimplementedimmediately,limitingthegainsfromthe
processenhancementefforts.
3. UsingtheEIPtolinkanimprovementprojecttothebusinessgoalsandstrategiesmayallowthe
practitioner to gain more support to complete his/her projects, but it also may help him/her
understandwhatthebusinessgoalsandstrategiesareintendedtoaccomplish,whichmaylead
toabetterselectionofprojects.
4. Using Design of Experiments concepts for data collection and to limit the size of existing data
setsmayprovetobemorevaluablethantraditionalDOEuses.
5. AdoptionoftheKotter8stepstomanagechangeforLeanSixSigmaprojectscanexpeditethe
effectiveness of the improve phase. In addition, these steps can be useful in every type of
changeenvironment,whichcouldleadtothepractitionersbecomingaheroinhis/herbusiness.
Becoming proficient in the five areas that have been described will make the student not only a
betterbeltbutalsoabetteremployee.
SmarterSolutions,Inc. 12 www.SmarterSolutions.com
Five Things Every Lean Six Sigma Belt Should Learn In Their Training
References:
LeanSixSigmaBooks:ProjectExecutionGuide,ForrestW.BreyfogleIII,CitiusPublishing,2010.
TheIntegratedEnterpriseExcellenceSystem:AnEnhanced,UnifiedApproachtoBalanced
Scorecards,StrategicPlanning,andBusinessImprovement,ForrestW.BreyfogleIII,Citius
Publishing,2008.
ConnectwithForrestonLinkedIn,TwitterandGoogle+
FormorefreeresourcesbyForrestBreyfogle,jointheOnlineResourceLibrarytoday.
SmarterSolutions,Inc. 13 www.SmarterSolutions.com