forecasting with collaboration? - lancaster university · forecasting with collaboration? ... best...

Forecasting with Collaboration? © 2013 Crone & Weller, Lancaster Centre for Forecasting

Forecasting with Collaboration? The Gap between Theory and Practice

23 April 2013, The Work Foundation, London, UK

Dr. Sven F. CronePhD candidate Matt Weller

Assistant Professor - Director

Note: a complete written report is available! Please contact Dr. Sven F. Crone [email protected] to receive it

23 April 2013, The Work Foundation, London, UK

Dr. Sven F. CronePhD candidate Matt Weller

Assistant Professor - Director

This material has been created and copyrighted © by Lancaster Centre for Forecasting, Lancaster University Management School, all rights reserved.

You may use this material for your private educational purposes, so long as they are clearly identified as being created and copyrighted © by Lancaster Centre for Forecasting, Lancaster University Management School. You are not permitted to alter, change, or enhance them.

You may not use any of the content, text and images, in full or in part, in a tutorial, training, education, written papers, videos or other recordings. You are not permitted to distribute or make available directly or indirectly, within or outside your company, nor exploit them without explicit prior written permission from Lancaster Centre for Forecasting, Lancaster University Management School, email: [email protected]

Forecasting with Collaboration? © 2013 Crone & Weller, Lancaster Centre for Forecasting Page 2

Forecasting with Collaboration?

ePhysical flow of goods

Flow of Information

$$


e

$$

Forecasting in Silos

e

$$

S&OP

Sales & Operations Planning


7‐Eleven, Inc.AAA HangersAbility Counts Incorporated CompanyAcme Distribution CentersADT SensormaticAIT Trimmings, Inc.Alfred Dunner, Inc.Alliance Hangers (Far East) LimitedArmy & Air Force Exchange ServiceBanner PharmacapsBest Buy Co., Inc.Black & Decker CorporationBon‐Ton Stores, Inc., TheBoscov's Inc.BraiformBrown Shoe Company, Inc.Byer CaliforniaCabela's, Inc.California Label ProductsCapco Wai Shing, LLCChevron Phillips Chemical CompanyColgate‐Palmolive CompanyCVS Caremark | CVS/pharmacyDefense Logistics AgencyDellDick's Sporting Goods, Inc.Dillard's, Inc.DIY HangersDonna Karan InternationalDynamic Worldwide Logistics, Inc.e‐Long Asia LimitedElizabeth Arden ‐Makeup, Cosmetics & SkincareERA GROUP INC.Ermenegildo Zegna CorporationEsquel Enterprises Ltd.Estee Lauder CompaniesEvergreen Plastic Bags Trading LimitedFedEx Freight

FineLine TechnologiesFishman & Tobin, Inc.Garmin InternationalGOH International, Ltd.Greatwide Logistics Services CompanyH.H. Brown Shoe Company, Inc.Hanesbrands, Inc.Hangerlogic, Inc. (formerly Batts Canada)Hangers PlusHewlett‐Packard CompanyHo Fat Hangers & Button Products Co., Ltd.Holt Renfrew & Company LimitedHome Depot, Inc., TheHudson's Bay CompanyIdeavillage Products Corp.JCPenneyJDM Sorting LLCJockey International, Inc.Johnson & Johnson Sales & Logistics Co.KANE IS ABLE INC.KDS AccessoriesKimberly‐Clark CorporationKohl's Department Stores Inc.Kraft Foods North AmericaLevi Strauss & Co.LF USA Inc.Limited Brands, Inc.Liz Claiborne Inc.Lowe's Companies, Inc.LVMH GroupMacy's, Inc.Mainetti GroupMark Bric, Inc.Meijer, Inc.Michael Kors LLCMotorola, Inc.National Retail Systems, Inc.NEO Zipper Co., Ltd.Nestle' PurinaNew Wide Group

Nike, Inc.Nintendo of America, IncNordstrom, Inc.Nygard International PartnershipOld Dominion Freight Line, Inc.Perry Ellis International Inc.Phillips‐Van Heusen Corp.Polo Ralph Lauren CorporationPoly‐Pak Industries, Inc.Procter & Gamble Company, TheQVC, Inc.r‐pac International Corp.Remac, Inc.RG Barry CorporationRyder Systems, Inc.Saks Inc.Samsung Electronics AmericaSauder Woodworking Co.Schering‐Plough Healthcare ProductsSchneider National, Inc.Sears Holdings CorporationSewing Collection, Inc.Sheen Master Industrial (Hong Kong) LimitedSimmons Bedding CompanySony Electronics, Inc.Sourcing Solutions InternationalSun Wing Hanger & Button Mfy. Ltd.SZKJ Plastic Corp.Target CorporationTerpac Plastics International Inc.The Accessory Corp / Hanger HeadquartersThe Clorox CompanyThe Jones GroupThe TJX Companies, Inc.TIC GROUP USA LLCUnileverUniplast Industries Inc.UPSUSA TACOMVF CorporationVisconti Garment Hangers, Inc.Wai Yip Plastic Mfy.Walmart Stores, Inc.Warnaco Group, Inc., TheWest Marine, Inc.West Point Home, Inc.William Carter CompanyWise Well Plastic Products Trading LimitedYRC Worldwide Corporation

Associate

AccentureAlixPartners, LLPAvery Dennison CorporationCheckpoint Systems, Inc.Data2 CorporationDemandTec, Inc.DICentral Corp.eBizprise Inc.GXSIBMJDA Software Group, Inc.Kurt SalmonManhattan Associates, Inc.Oliver Wight Americas, Inc.Oracle America, Inc.PTCRedPrairie CorporationSeeonic, Inc.SPS CommerceTrade & Standards

American Apparel & Footwear AssociatioChina National Textile and Apparel CounCouncil of Supply Chain Management Professionals (CSCMP)EPCglobal USGS1 CanadaGS1 USNational Retail FederationRetail Council of Canada (RCC)Retail Industry Leaders Association (RILARosettaNetOther

Edgewood Consulting GroupItem‐Level RFID Committee

VICS Members

S&OP is well established & Best Practices exists:

• Use statistics for baseline forecasting• Limit & constrain expert Judgment • Segment assortment ABC XYZ and focus on important items (in S&OP)

• Measure & track accuracy (FVA …)• …

S&OP focusses on exchangeof company internal information lack of external (downstream) information! How to integrate this?


e

$$

Sales & Operations Planning plus Contiinuous Replenishment

ePOS

Manufacturer receives ePOS data of consumer sales from retailer• Uses ePOS data to improve forecasting & reduce lead time

• Synchronises internal production scheduling & inventory activities with actual sales at retailers

• Retailer still prepares their individual


e

$$

Vendor managed Inventory(& Consignment Selling)

Manufacturer decides on appropriate inventory levels & policies of products • Manufacturer is responsible for creating & maintaining inventory plan

• Manufacturer creates orders (within previously agreed upon bounds)

• Ownership of inventory remains with downstream (Retail / 3PL) partner

Stocks

e

$$

Collaborative Planning, Forecastingand Replenishment


What info sharing & collaboration activitiesare you currently engaged in?

Disconnect between theory and practice requires attention (Fildes et al, 2008) Empirical evidence sought regarding forecasting practice under collaboration

Motivation and Research Questions

►So what about the forecasting part? Model selection is important in sharing forecast information (Zhao,2002) Lack of integration of downstream data into forecasting(Smaros 2007, Lapide 2005)

►Research questions addressed1. How are manufacturers collaborating with downstream partners?2. How are manufacturers forecasting and using downstream data?3. What factors are most important to accuracy?

Prior Research Focus Limitations

Modelling Studies:Aviv 2001, 2002Zhao, 2002, Disney 2003Chen, 2007, Waller 2011

Supply chain benefits• Inventory costs• Service levels• Bullwhip reduction

Often unrealistic assumptions & simplificationsBasic demand processes e.g. iid, AR(1)Unrepresentative e.g. seasonality, promos, level shiftsLimited forecasting methods

Empirical Studies:VICS, variousECR Europe 2001, 2002GMA 2003Smaros 2007

• Benefits• Barriers• Enabling factors• Collaborative process

Often idealised cases (practitioner literature)Single case studiesVery limited coverage of forecasting!


200 complete surveys from forecasters in Manufacturing 173 classified in comprehensive group of collaboration

• Questionaire Design– Pilot study in 2011 (to ensure validity)– Final version pre‐tested with 18 FMCG forecasters– Questionnaire implemented online– Conducted January 2012‐August 2012

• Survey Sample Design– Specified target group: demand planning &

forecasting professionals (in manufacturing)– LCF Mailing list, forecasting lists / blogs (ISF, SAS)– 100s of LinkedIn Groups – 2000+ personalised LinkedIn invites– Multiple reminders sent

• Response– 540 responses

• 260 incomplete (reminders send, to only speculative interest, unwilling to give email address (although not mandatory), Atrophy (number of repeated questions), unsuitable respondent (industry sector & position)

• 15 complete responses discarded (Consultants/academics, rushed surveys (10‐15 mins), highly inconsistent answers, middle‐clicking { same answer for every question in groups)

Study MethodologySurvey of Practitioners

mostly large manufacturers active in the FMCG / CPG industry

Study MethodologyDemographics of Respondents


OrganisationalView

Data View

FunctionalView

Control(Process) View

Survey ResultsOrganisational Setup

Companies are engaged in multiple organisational schemes

85% of companies are engaged in multiple schemes(S&OP plus >>Info >>CPFR >>VMI)

Still forecasting in Silos? You are the last one … ;‐)

Most companies are engaged in S&OP (unclear how formalised)

O

D FP

Companies are most strongly

engaged in S&OP(least in CPFR)


grouped responses : CPFR and S&OP dominate sample reponses

6259

2718

277

27 responses did not clearly fit a scheme & were

excluded

Survey ResultsOrganisational Setup

O

D FP

most companies are engaged in multiple collaboration schemes in parallel … should receive substantial amounts of internal + external data!

… run multiple planning processes in parallel. SUPRISING!

8%

6%

34%

52%

Number of IS&C Schemes engaged in

No external IS&C

S&OP + 1 externalIS&C scheme

S&OP + 2 externalIS&C schemes

S&OP + 3 externalIS&C schemes

Over half of companies

engage in S&OP, IS, VMI and CPFR simultaneously!

ResultsO

D FP


Planners run S&OP, CRP, VMI, CPFR … in parallel!

(different data exchanged, different client contacts, different planning frequencies …)

… requires goodMultitasking!

(It can be done, with standardised Data & supporting IT Systems)

O

D FP

On average, more internal data is available than external (despite IS&C) (as expected) IS&C schemes have more external data available

Survey ResultsData Exchange

Less for silos ….

Internal data widely available (across schemes, effect of S&OP)

**

Never for silos ….

O

D FP

External info shows different availability by CPFR >> VMI …




…


Company internal data soucres!

VMI has least info on DC orders & shipments (?!?) Ta

nsa

ctio

nal

Dat

aC

au

sa

lDa

ta

Silos have only little/least internal

info available

CPFR (Marketing focused) receives more Promo info


info available


info available

VMI has least info on DC orders & shipments (?!?)



O

D FP

…


Company external data soucres!

Surprisingly many external data

types available !

CPFR has more info available

> VMI > IS > S&OP

essential VMI info is not available

essential VMI info is not available

Limited ePOS data available across schemes

S&OP receives little external information Surprisingly many

external data types available !

Surprisingly many & heterogeneous external data

types available !

No external data is always available (Promos for CPFR. DC stock for VMI) !

O

D FP


Choice of 6 transactional data sources is always available Up to 18 (5 internal + 13 external) data sources sometimes available


Internal info more available than external info

Internal info more available than external info

Myriad of information is available!

O

D FP

Amount of information = Information overload!

How do I chase the missing data?

What else do I expect will likely happen?

What will rather not happen?

Why do I have to spend my time on this?

I should be out there making forecasts!

Maybe next month ….


What data do you share?

…


Info is shared only medium horizons of 5‐8 weeks out

Data is exchanged by unstructured means (eMail …)

Structured Data is mostly for CPFR & VMI (but still rare for a standard)

O

D FP


…


Data is not exchanged in a standardised or consistent form

O

D FP

Data arrives in heterogeneous formats …

… how to harmonise& integrate into systems?


Repetitive data collection & transformation

(productive) Value Added activity?

… rather a time killer!


…


Spreadsheets are most widely & regularly used

O

D FP

…


Few individuals are really happy

;‐)

O

D FP


What systems do you use?

25% of all judgment + 76% of stats (57%) 82 % of forecasts are made with simple Methods

25% pure expert

Judgment

43% use simplest methods only for level time series

(28% MA+15% Naive)

43% use simplest methods only for level time series

(28% MA+15% Naive)

Only 10% of methods are capable to integrate causal

drivers in forecasting (Promos, DC stock …)

Only 10% of methods are capable to integrate causal

drivers in forecasting (Promos, DC stock …)

Survey ResultsAlgorithm Usage

O

D FP


What algorithms do you use?

Use of judgment on myriad of data sources?

Survey ResultsData Usage

O

D FP

NONE of the data is not used purely

statistically!


…

Survey ResultsData Usage

O

D FP

Most data is considered

important, but used only

Judgmentally

New Tools?


Naive Method(s)

New Tools!

“It can drive a 6‐D nail thru a

2 X 4 at 200 yards”


Brown (1956)

Box & Jenkins (1970)

Rumelhart & McClelland (1986).

“the number of components in integrated circuits (IC) has doubled every year from its invention in 1958 until 1965” predicted that the trend would continue "for at least ten years"

Moore (1965)

Breiman(1996)

Breiman(2001)

Freund & Schapire(1997)

Boser, Guyon, and Vapnik (1992)

Pearl & Reed (1920)!

Does your vehicle fleet look like this?

BMW Isetta (1956-1962)

Austin Healey 3000 (1959)


Man and woman shown working with IBM type 704 electronic data processing machine used for making computations for aeronautical research.

Date 21 March 1957Working with IBM type 704 electronic data processing machine used for making computations for aeronautical research.21 March 1957, NASA

Does your hardware look like this?

Do your employees look like this?


Then why do youralgorithms look like this?

∙ ∙

∙ ∙

∙ ∙

∙ ∙

© Scorcher publications (1956) Pamphlet reproduction on the revolution in Hungary 1956, Series of Council Communist Pamphlets.


… talking about TECHNOLOGY!

Ask yourself …

Can you affordto ignore 60 years

of progress?


ARE you asking yourself?(… just checking!)

Review youralgorithm‘s

maturity!


Upgrade to 1970s technology … ?

Siemens System 4004

Intel 4004 (4-bit bus, 108-KHz CPU chip, 1971)

OCC1 Osborne 1 Personal Business Computer

busicom calculator 141P(Intel 4004 powered Datapoint 2200 (1972)

…ARIMAX

Multiple Regression

…

Survey ResultsAccuracy

O

D FP


…

Survey ResultsAccuracy

O

D FP

…

Survey ResultsHow to improve?

O

D FP


…

Survey ResultsHow to improve?

O

D FP

Take aways

• Companies are engaging in multiple Info Sharing & Collaboration schemes simultaneously does this Multitasking add value?

• Demand Planners are overloaded with Data which data sources add value & Information?

• Demand planners receive unstructured Data tedious data processing jobs kill valuable time

• Lack of systems support Excel is great for data gathering, not forecasting!

• Comanies ignore newer algorithsm that can help! algorithms exist to integrate data automatically

• Review your Collaborative practices try new approaches to forecasting & using info!

… making Collaboration work !


… Discussion?

Matt Weller

Dr. Sven F. CroneAssistant Professor, Director

Lancaster University Management SchoolResearch Centre for Forecasting

Lancaster, LA1 4YXTel. +44 1524 5‐92991

[email protected]

Questions?


Matt WellerDr. Sven F. Crone

Assistant Professor, DirectorLancaster University Management School

Research Centre for Forecasting

Lancaster, LA1 4YXTel. +44 1524 5‐92991

[email protected]

Copyright

These slides are from a workshop on collaborative forecasting.

You may use these slides for internal purpose, so long as they are clearly identified as being created and copyrighted © by Matt Weller and Dr. Sven F. Crone, Lancaster Centre for Forecasting.

You may not use the text and images in a paper, tutorial or external training without explicit prior permission from the centre.

forecasting with collaboration? - lancaster university · forecasting with collaboration? ... best...

Documents