Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. POMS Applied Research Challenge, POMS Atlanta, May 9. 2014 How Company-Specific Production Systems Affect Plant Performance: The S-Curve Theory Torbjørn H.Netland Ph.D.,Postdoc NorwegianUniversityofScienceandTechnology(NTNU),Trondheim,Norway [email protected] KasraFerdows HeisleyFamilyChairofGlobalManufacturing McDonoughSchoolofBusiness,GeorgetownUniversity,WashingtonD.C. Ebly Sanchez VPSDirectorAmericas VolvoTrucksNorthAmerica,Greensboro,NC Company-specific Production Systems are essentially Corporate Lean Programs aiming to improve the operational performance of all plants in the company’s global network 1 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Does implementing an XPS improve plant performance? 1. Theempiricalliteraturesaysitdoes TQM(e.g.Sila,2007,JOM;Kaynak,2003,JOM;Black&Porter,1996,DS) Lean/JIT(e.g.Shah&Ward,2003,JOM;FullertonandMcwatters,2001,JOM) Sixsigma (e.g.Swink&Jacobs,2012,JOM;Shafer&Möller,2012,JOM) TPM(e.g.McKone,Schröder &Cua.,2001,JOM) 2. Practicesaysitdoes CompaniescontinuedevelopingXPSs Companypresentationsreportingmillionsofdollarssaved Popularliterature(TheEconomist;theLeanManagementJournal;etc.) 3. Ourownresearchsaysitdoes ResearchinVolvoABandJotunAS(Netland &Aspelund,2013,IJOPM; Netland&Sanchez,2013,TheTQMJournal) OursecondpaperatPOMS:"Incentivesforimplementingcorporate leanprograms" (Netland,Schloetzter &Ferdows,2014) Plant performance Plant performance f(x) Plant performance Exactly how does the implementation of an XPS affect plant performance? The relationship between XPS implementation and plant performance Rate of improveme Rate of improveme f '(x) Rate of improveme The pattern suggests what rate of improvement we should expect 2 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Why should we care about the pattern? Misplaced expectations of how quickly these programs can improve performance can make their implementation difficult and reduce their benefits. From: "How to Implement a Corporate lean Program" MIT Sloan Management Review, forthcoming Summer 2014 What do existing theories predict? Fourtheoriespredictthetotaleffectofdepth andspread 1. Thelearningcurve Effect of depth of XPS implementation 2. Thetheoryofperformancefrontier 3. Organizationalinertia Pattern of spread of XPS implementation in a plant 4. Epidemiologytheory We hypothesize that the combined effect of these theories is likely to result in an S-shaped performance curve 3 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Our Research Method In-depth case study research (Barratt, M., Choi, T.Y., Li, M., 2011, JOM; Yin, 1994; Eisenhardt, 1989, AMR) The Volvo Group StillSwedish…but global! HQGothenburg,Sweden Founded 1927 About 115.000employees Salesin180+markets Truck brands: Largest truck manufacturer in the world 4 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. VTC Umeå VCE Arvika Volvo's network VPT Köping VCE Eskilstuna VPT Reman Flen VCE Hallsberg VPT Skövde VAC Trollhättan VPC Vara VTC Tuve VPC Gothenburg VBC Borås VCE Braås VAC Kongsberg VCE Wroclav VBC Wroclav VCE Hameln VCE Konz VTC Gent RTC Blainville RTC Limoges RTC BeBresse VPT Madrid VBC Quebec VBC Montreal VBC Plattsburg VAC Newington VCE Shippensburg VPT Reman Middletown VTC Macunige VPT Hagerstown VTC New River Valley, VPT Reman Charlotte VPC Lexington VTC Kaluga VCE Changwon RTC Vénissieux VPT Vénissieux VCE Belley VPT Konouso VPT Hanyu VPC Ageo VCE Linyi VTC Ageo VPT Pithampur* VBC Mexico City VTC Bangkok (VCE Mexico City) VTC Banagalore VCE Banagalore VTC Las Tejerias VBC Xian VBC Banagalore VCE Pederneiras VTC Curitiba VPT Curitiba (VBC Curitiba) VCE Shanghai VPC Shanghai VBC SAIC Shanghai* VTC Brisbane VTC Durban VTC Umeå VCE Arvika Volvo's network VPT Köping VCE Eskilstuna VPT Reman Flen VCE Hallsberg VPT Skövde VAC Trollhättan VPC Vara VTC Tuve VPC Gothenburg VBC Borås VCE Braås VAC Kongsberg VCE Wroclav VBC Wroclav VCE Hameln VCE Konz VTC Gent RTC Blainville RTC Limoges RTC BeBresse VPT Madrid VBC Quebec VBC Montreal VBC Plattsburg VAC Newington VCE Shippensburg VPT Reman Middletown VTC Macunige VPT Hagerstown VTC New River Valley, VPT Reman Charlotte VPC Lexington VTC Kaluga VCE Changwon RTC Vénissieux VPT Vénissieux VCE Belley VPT Konouso VPT Hanyu VPC Ageo VCE Linyi VTC Ageo VPT Pithampur* VBC Mexico City VTC Bangkok (VCE Mexico City) VTC Banagalore VCE Banagalore VTC Las Tejerias VCE Pederneiras VTC Curitiba VPT Curitiba (VBC Curitiba) VCE Shanghai VPC Shanghai VBC SAIC Shanghai* VBC Xian VBC Banagalore VTC Brisbane VTC Durban Volvo Production System 5 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Our Data VCE Arvika VTC Umeå VPT Köping VCE Eskilstuna VPT Reman Flen VCE Hallsberg VPT Skövde VAC Trollhättan VPC Vara VTC Tuve VPC Gothenburg VBC Borås VCE Braås VAC Kongsberg VCE Wroclav VBC Wroclav VCE Hameln VCE Konz VTC Gent 40 factory visits 200 interviews Survey 312 responses VPS assessment database RTC Blainville RTC Limoges RTC BeBresse VPT Madrid VBC Quebec VBC Montreal VBC Plattsburg VAC Newington VCE Shippensburg VPT Reman Middletown VTC Macunige VPT Hagerstown VTC New River Valley, VPT Reman Charlotte VPC Lexington RTC Vénissieux VPT Vénissieux VCE Belley VPT Konouso VPT Hanyu VPC Ageo VCE Linyi VTC Ageo VPT Pithampur* VBC Mexico City (All original data) (VCE Mexico City) VTC Bangkok VTC Bangalore VCE Bangalore VTC Las Tejerias VCE Shanghai VPC Shanghai VBC SAIC Shanghai* VBC Xian VBC Bangalore VCE Pederneiras VTC Curitiba VPT Curitiba (VBC Curitiba) VTC Kaluga VCE Changwon VTC Brisbane VTC Durban Explaining the analyses Performance Independentvariableismeasuredby the VPSassessmentscores High Low Limited Assessmentlasts4days,5‐6trainedassessors Standardizedauditofabout130items ImplementationlevelofVPSon0‐5LikertScale >100plantassessmentsfor49plants XPS impementation z-scores Extensive 6 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Explaining the analyses Safety|Quality|Delivery|Cost Performance Dependentvariableismeasured infourdifferentways High 1. Plantperformancelevel:fromassessment data (latestversionofassessment,N=25) 2. Rate ofimprovementinplantVPSscore: fromassessment data (plantswith2ormoreassessments,N=35) 3. Rate ofimprovementinplantperformance: fromsurvey (7items,N=32) 4. Rate ofimprovementinplantperformance: fromvisits,observations,andinterviews(N=40) Low XPS impementation Limited Extensive Statistical technique used for pattern recognition Locally weighted regression (LOESS) LOESSisatechniqueforfittingthebestcurvedepictingtheshapeofthe relationshipbetweentwovariables(ClevelandandDevlin,1988). Amajoradvantageisthatitdoesnotneedapriorispecificationofafit function:Itdiscoverstheformfromthedataitself. Using a kernel function as a smoothing algorithm, LOESS computes a center for each neighborhood of data points (decided by the smoothing parameter alpha) that minimizes the weighed distances between the center and the points in that neighborhood. It then draws a curve through these local neighborhood centers. Parametersinouranalyses Epanechnikovkernelfunctionhasrobustproperties (Gasseretal.,1985) Areasonablevalueforalphais0.40<α <0.80(Jacoby,2000) 7 Plant performance f(x) Plant performance Test 1 of 4 Plant performance Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Plant performance level (Score 1-5 from latest assessment) Safety|Quality|Delivery|Cost LOESS curve fitted to scatter plot ( α=0.40) Leanimplementation VPS implementation Rate of improveme f '(x) Rate of improveme Tests 2, 3 and 4 Rate of improveme Plants (N=25) (z-scores from Assessments) Rate of performance improvement All four tests independently suggest the S-curve (bell-curved rate of improvement) 8 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Conclusions XPS implementation affects performance non-linearly like an S-Curve Performance Rate of performance improvement Stage I Beginner Stage II In-transition Stage III Advanced Stage IV Cutting-edge Level of XPS implementation Performance improves slowly in initial stages of XPS implementation, then improves rapidly and eventually improves slowly again Managerial implications Plants in each stage should be managed differently Don’t apply the same action plan in all plants in the global network Plant performance • Continue to allocate budget for the program even though rate of improvement slows down • Allow and encourage these plants to establish more direct linkages outside the firm • Leverage plant’s distinct capabilities strategically • • • • • • • • • • • Increase allocated budget for continuous improvement projects Give local managers more autonomy in choice of projects Use these plants as benchmarks for other plants Set stretch targets but expect declining rate of improvement Set stretch targets and expect accelerated rate of improvement Publicize improvement successes Watch for creeping complacency Hold extensive training sessions in pilot areas of the plant Establish dedicated implementation teams to drive and coach the program Allocate budgets, but set small targets for improvement Follow progress closely (e.g., show up in the plant frequently), but be patient Stage 1 Beginner Stage 2 In‐transition Stage 3 Advanced Stage 4 Cutting‐edge 9 Netland, T. H., & Ferdows, K. (2014) How company-specific production systems affect plant performance: The S-curve theory. Working paper. Norwegian University of Science and Technology, Trondheim / Georgetown University, Washington, D.C. Be prepared for critical transitions! There are danger zones at each stage Plant performance Why spend more for little additional improvement? We are not getting the 4 needed resources We are doing enough 3 We are different 1 Stage 1 Beginner Stage 2 In‐transition Stage 3 Advanced Stage 4 Cutting‐edge Thank you! & Torbjørn H. Netland, NTNU Kasra Ferdows, Georgetown U. Ebly Sanchez, Volvo More about this research at: www.better-operations.com 10
© Copyright 2024 ExpyDoc
