1. Demand Planning– 8.2.18 notes
Laura Buck, Hyster-Yale
RobertByrne,e2open
ChrisReed,Ferguson
AndrewSmith,PepsiCo
Andrew Sokell,Hyster-Yale
VikramSriinivasan,Llamasoft(subforRyanChan)
ChipWhite,Ph.D.,GA Tech
AmyThorn, DBMA
Amy – I’ll bringyouupto date on what’s beengoingonsince the SCLA conference. Eachpeergroup hashad a follow-up
call to brainstormwhatthe format wouldlooklike forapresentation for2019 lab.Thisisgoingto be a keygroup. I look
forwardto yourengagement,involvement,oversightandleadershipaswe developthe topicsfor2019. The task forthis
groupis to continue the conversation,take the pressingissue youcame upwithanddetermine how toaddressthatin
an educational formatin2019. There’sbeenalotof interestinthe followingareas:visibility,AI,machine learning,big
data, andpredictive analytics.
The art and science of demandforecasting,whichcapturesmuchof the othertopicsthat were covered(e.g.,
correlationforimprovedforecasts)
Demandforecastingversusadaptability. Thisgoesbackto productdesign. If leadtimesare shortand the
productis designedtobe differentiatedlate inthe supplychain(BOToralmostBOT; pull supplychains),then
accurate forecastsaren'tas needed. Perhapsasteptoconsideristo do whatcan be done inproductdesignso
accurate demandforecastingisn'tcritical tosuccess.
Andrew – Howto applyAI to demand planningalgorithmsistopof mind.
Amy – We’re lookingforalignmentonthe potential topicsanda deeperdive.Whatare youdoingwithinyourcompany?
Chip– Howcan you more accuratelypredictdemand,forecasting? Lookingforcorrelationswhichwouldhelpdemand
forecastsbe more accurate – weatherconditions,etc. If youcan redesignyourproductinsupplychainsoandshorten
leadtimesdemandforecastingbecomeslessimportant. Asaresultyouhave more visibilityonthe demandside. That
seemedtorise above the general topic. Howdoyoumonetize the datayoucollect andextravalue?Insome casesmore
accurate forecastsdon’thelpmuch.Whatwouldhelpisif youcoulddifferentiateyourproductlate inthe processbefore
youcomplete the product. Are there toolstobetterhelpunderstandthe value of demandforecastsgiventhe same lead
timesandno designmodificationvs.redesigningthe productandshortenedleadtimes. Itmightbe more expensivebut
wouldhave animpact inthe future. Also,there isinterestinmobilityinsupplychains –the abilitytoquicklymove
aroundproductionor stage capacityto bettermatch demandtrends.
Amy – Do youfeel thatyouhave a bestpractice inthisarea?
Laura – Rightnowmy projectiscentralizingsystems.Everythingasupplierwouldneed –it’sa one stop shop – POmgmt.
system,etc.I’mworkingtoestablishthatforall of oursuppliers.We’llstarttotie that intoour systemtoconnectwith
Ryderand Cevato increase visibilityandsee whereourproductsare at. US isRyderand CevainEurope.
Amy – Ryderisan SCLA member.
Chris– I’mnot workingwithanyof those companies.
Chip– IsXPO involvedwithSCLA? Theyhave lotsof assetsinEurope andthe US withrecentacquisitionsinthe lastfew
years.
Amy – No.
Laura – I was surprisedattheirfootprintinEurope andwhattheyare able tohandle forus. Theirfootprinthere inthe
US is not thatlarge. In termsof cost, Ryderhas the largerfootprint forwhat we needtoreduce ourcost.
Chip– Do youexport/importfromsouthernFL?
2. Laura – Some – several forkliftsforourdealersinthe islandsandso.America.
Chris– We talkedaboutnewproductintroductionsandthe rightmanaging,shaping,demandsensingof thatinour
forums. I didn’thearthatmake the list.
Chip– That’sa topic that came up fairlyregularly. Youmayhave a differentsupplychainfornew productintroduction
that mightchange basedon endof life cycle of the product.
Chris– Ferguson(wholesaler) isnotactivelyinvolvedindesigningthe productbutunderstandinghow the marketwill
react – deployingtothe rightplace, availability thatcomesfromthe demand,etc.Andrew sharedanumberof potential
demandcurves.We have to identifythe bestfitcurve ondayone thenhow to sense andadapt torealityas itoccurs
weeklyandmonthlytoupdate the curve.
Andrew - There wasdiscussionof pluggingintoproductdesign(sales,marketing,etc.) sothatyouunderstandsales
targets,etc.That goesinto whatsupplychainoutlooklookslikeandthe range of things.
Chip– the art and science of demandforecastingandhow tomanage yoursalesand marketingpeople,incentivize them
so everyone ismovinginthe same direction –salesandintroductions.
Chris- You wantthe rightdynamictensiontogetto the rightoutcome.Be biasedwithsuccessfirstthenacceptreality if
youhave to but do whatyou can to be successful first.
Chip– Fail fast. Case studiesaroundnewproductintroductionseemedtohita harmonicwithinthe group.
Pickingupon the case studypiece (focusonnew product),how youmap processand mapa diagnosticof howdemand
sensingtravel intoandthroughyourorg thenback out to yoursupplybase.Doyou have a processof inquiringwhatbest
practice lookslike?
Chip– Howdo youavoidthe bullwhipaffectbysendingthe rightsignalstoyoursupplybase?
Amy – Have you done anyobjective analysisof the demandplanningtoolsoutthere?Woulditbe valuabletopresent
differenttools?
Vikram– We have toolsonthe designside of thingstoprovide modelingonhow youcan run scenarios. You’re notonly
lookingatdemandvariabilitybutthe external factors.We alsoprovide life cycle modelingcapabilities.
Chip– I thinkthatwouldbe veryinteresting. We cancontribute fromthe researchside. Ispoke toone of mystudents
followingthe SCLA event.
Amy – Conceptually,come upwithhowyouwouldaddresssome of the challengesandsolutionswe talkedabout –a
paragraph or two.That’ll be a startingpointfor future conversation. Chip,iswhatyou’re doingsharable?It’dbe nice to
see that laidoutas well.Forthe othersinthe group,is there a willingnessforyoutobe one of the case studiesand
share what hasworked,hasn’tworked,etc. I’dlike tosee new productlaunchasone of the case examples. Let’ssetour
nextcall for September.
Nextconference call:Thursday, September27th at 11:00 AM Eastern