Lean logistics

153

Lean logisticsDaniel T. Jones, Peter Hines and Nick Rich

Cardiff Business School, Cardiff, UK

Problem definitionThe last ten to 15 years has seen a series of “new” business solutions – each ofwhich contributed a new perspective – but each of which may have now run itscourse certainly as a stand-alone solution. These include:

• the Tom Peters approach which brought the customer back into primefocus and told us to “thrive on chaos” and break up traditionalmanagement structures[1,2];

• the TQM movement, often incorrectly labelled Japanese management,which advocated the power of variance analysis and control and thenecessary involvement of shopfloor teams in eliminating root causes ofvariance[3-5];

• the production control perspective which sought to solve the chaos andvariance problems in supply chains by better forecasting and materialsrequirements planning (MRP)-type planning and control systems[6];

• the purchasing emphasis, which involved a transformation topartnerships with suppliers rather than the previous adversarial stylerelationships[7,8];

• the business process re-engineering (BPR) viewpoint which stressed theimportance of the process and offered ways of automating theseprocesses to cut costs[9].

With hindsight, all of these tended to be partial solutions focused somewhatnarrowly on particular aspects of the complex process of running a business.Popular criticisms of such approaches are that:

• Total quality management (TQM) raised morale on the shopfloor butcould not cumulate the gains to real bottom-line savings while buyer-supplier partnerships again often failed to deliver[10-12];

• MRP and MRP II created monolithic systems that cannot respond torapidly changing demand or unpredictable interruptions and add costwhile failing to improve process utilization[13];

• BPR’s credibility was undermined because it was used as a crudeheadcount reduction device, particularly in the USA.

One likely reason these waves of change have had such a limited impact is thatcompanies are still in a world of batch and queue processing, with armies ofexpediters and progress chasers working to beat the system and cope with thechaos. The second reason is that these remedies do not dig deep enough really

International Journal of PhysicalDistribution & Logistics

Management, Vol. 27 No. 3/4, 1997,pp. 153-173. © MCB University

Press, 0960-0035

IJPDLM27,3/4

154

to transform the ways companies operate – they have too often been seen asbolt-on extras. The third reason is that consultants have only really been able tocarry out in-house change programmes or to sort out the logistics betweenpairs of companies, when in fact the changes need to be replicated by all firmsinvolved up and down the supply chain.

So it is perhaps not surprising that, despite a lot of talk, general businessperformance is seen by many as stagnating and not improving by leaps andbounds, leading to employee cynicism and uncertainty about employmentprospects.

Yet there are examples of organizations which stand out from the generalstatus quo and these offer promising avenues that may help the observer tomake sense of what went right and what went wrong with the solutionsmentioned above. To get beyond these partial solutions more fundamentalquestions about the organization of value creation have to be asked. In additionit is necessary to look behind the tools and techniques being used by theseexemplar firms to discover the underlying logic behind them.

A new logicA new logic was examined by Womack and Jones[14,15]. A natural startingpoint is with value creation – from the customer’s perspective the only reasonfor a firm to exist. If subjected to a careful review, many of the steps required inthe office to translate an order into a schedule and many of the steps required inthe factory physically to create the product, add little or no value for thecustomer.

Taiichi Ohno defined seven common forms of waste, activities that add costbut no value: production of goods not yet ordered; waiting; rectification ofmistakes; excess processing; excess movement; excess transport; and excessstock[16-18]. It is common to find that in a factory less that 5 per cent ofactivities actually add value, 35 per cent are necessary non-value-addingactivities and 60 per cent add no value at all[14,15]. It is easy to see the stepsthat add value, but it is much more difficult to see all the waste that surroundsthem. Our thesis is that beginning to eliminate this 60 per cent of activities andcosts offers the biggest opportunity for performance improvement today.

The provision of most goods and services stretches across severaldepartments and functions inside the firm and across many different firms.Making an aluminium drinks can, for instance, involves six manufacturingfirms from the mine to the supermarket. Each of these elements in the chain –the reduction mill, the smelter, the hot roller, the cold roller, the can maker andthe bottler – is busy trying to optimize its own performance, typically throughbigger machines with faster throughput times and bigger batches. This in turnleads to a series of intermediate stores getting bigger all the time. The net resultis that it takes 319 days to make the can, that itself takes only about three hoursof processing time!

Optimizing each piece of the supply chain in isolation does not lead to thelowest-cost solution. In fact it is necessary to look at the whole sequence of

Lean logistics

155

events, from the customer order right back to the order given to the rawmaterials producer, and forward through all successive firms making anddelivering the product to the customer. In trying to identify possibilities foreliminating waste this makes most sense if it is done for one particular productor product family – and for all the tributaries that flow into this stream of valuecreation.

Focusing on the whole chain is the first step; focusing on the product is thesecond and focusing on the flow of value creation, and not on the moretraditional performance measurement of departments and firms, is the third.This we term a “value stream” – a new and more useful unit of analysis than thesupply chain or the individual firm. Focusing on the flow of value creationimmediately challenges the notion that batches are necessary and better. Asimplified version of this is illustrated in Figure 1. The point that this valuestream concept extends both upstream from the product assembler into the“supply chain” and downstream into the “distribution chain” is reinforced inFigure 2.

Taiichi Ohno demonstrated that if one looks from the perspective of thewhole value stream it is possible, and indeed vastly superior, to organizeactivities so that the work moves from step to step in an uninterrupted flow ata rate that exactly matches the pull of the customer. This was always thought

Figure 1.The auto value stream

3rd

2nd

1st

Assembly

Distribution

Dealer

Partsdistribution

Materials

Productdevelopment

CustomerKey

Product definitionfrom concept through detailed designand engineering to production launch

Information managementfrom order taking through detailedscheduling to delivery

Physical transformationfrom raw materials to a finishedproduct in the hands of the customer

IJPDLM27,3/4

156

of as a special case that only applied to very high volume production of astandardized product, such as a car. Ohno proved the general case anddeveloped the tools and techniques necessary to achieve it. The tool kit for themanufacturing context has come to be called the Toyota production system.Womack and Jones[19] describe its generic form as lean thinking and haveelsewhere documented the impressive gains achieved by those firms whichhave followed his example.

The key elements of Toyota’s tool box are:• Level the flow of orders and work by eliminating all causes of demand

distortion or amplification.• Organize the work so the product flows directly from operation to

operation with no interruptions – shortening set-up or response times tomake or deliver every product every day/week and ensuring that nobreakdowns occur though preventive maintenance.

• Only make or deliver what is pulled by the upstream step – no more andno less – sell one, order one.

• Work throughout the system to the same rhythm as customer demand.• Standardize the best work cycle for each task to ensure consistent

performance.• Standardize and minimize the necessary safety stock between

operations.• Make every operation detect and stop when an error occurs so it cannot

be passed on – making it possible for one employee to supervise severalmachines or making it possible to detect rogue orders against a historicordering profile.

• Manage progress and irregularities using simple visual control devices.• Log irregularities and prioritize in order to conduct root cause elimination

to prevent recurrences and to remove waste from the flow.

Figure 2.Total domain of thevalue stream

Assembler

3rd

2nd

1st

3rd

2nd

1st

Assembler

Supplytiers

Support

SupportAftermarket

Rawmaterial

Tiers

Distributiontiers

Lean logistics

157

The interesting thing that happens when the lean principles are applied, usingOhno’s tool kit, is that one begins to rethink not only the organization of thework but the appropriateness of the size of machines, warehouses and systemsto fit the flow. As people, machines, warehouses and systems are rethought andcombined in different ways, layers of previously hidden waste tend to beuncovered and removal – perfection – becomes the appropriate goal and notwhat your competitor is doing today. Perfection is defined as the completeremoval of waste until every action and every asset adds real value for theultimate customer. In theory, waste removal is a continuous process, operatingcyclically and without end.

A distribution example – Toyota parts supply systemToyota developed its Toyota production system (TPS) in the late 1940s in itsengine shop. It was fully developed and applied across Toyota’s manufacturingoperations by the late 1960s. Then, for the first time, it was written down and agroup was formed to teach it to Toyota’s first-tier suppliers, who were allmembers of the Toyota Group. By the mid 1970s the TPS perspective hadspread to its other first- and second-tier suppliers in Japan primarily throughtheir kyoryoku kai or supplier associations[20]. After the merger of Toyota’smanufacturing and sales companies in 1982 they began to apply the same logicto the aftermarket parts operation in Japan. This took from 1984 to 1990, afterwhich the logic began to be spread to overseas parts suppliers (of original andreplacement parts), and to the aftermarket distribution systems abroad.

Womack and Jones[19] studied the transformation of Toyota’s US partsdistribution system (illustrated in Figure 3), right back to the second-tier partsmanufacturer of a replacement bumper and compared it with its system inJapan. The picture is an interesting one (see Table I).

Figure 3.The downstream lean

value stream

Lean manufacturing Lean warehousing Lean retailing

Lean ordering

Lean logistics

USA 1994 USA 1996 Japan 1990

Service rate 98 per cent in 7 days 98 per cent in 1 day 98 per cent in 2 hoursSystem stock index 100 33 19Throughput time (weeks) 48 8 4

Source: [19, p. 86]

Table I.Toyota parts

distribution systemefficiency

IJPDLM27,3/4

158

The following summarizes the main actions taken to make the dramaticimprovements to the value streams described above.

Manufacturing – stamping and chromingA six step approach to improvement was enacted on the factory shop floorinvolving:

(1) establishing target set-up times to make every product every day;

(2) reducing set-up times;

(3) changing the layout of machines to create single-piece flow of partsthrough the process;

(4) moving from monthly to weekly and then daily orders;

(5) stopping the use of MRP systems for production control;

(6) teaching the chroming subcontractor to process parts in a single-pieceflow.

The result of these actions was that throughput time was reduced from sixweeks to 48 hours, with zero defects and big cost reductions.

DeliveryToyota picks up both its OE and aftermarket parts from suppliers in the localarea on a “milk round” at regular and relatively short intervals. Figure 4compares this with a traditional auto assembly plant receiving full loads of alimited number of part numbers every three to four days directly from over 500suppliers. Having reduced the number of first-tier suppliers it deals with,through tiering, modularization and sourcing, many more part types from eachsupplier, Toyota can pick up small lots of each part number from fewer

Figure 4.Toyota JIT logistics

Direct full load

Milk roundexact quantities

170 suppliers45 parts/supplier

500 suppliers9 parts/supplier

every 4 hoursToyota

every 3/4 daysWestern assembler

Lean logistics

159

locations on a milk round every four hours. It then never has more than a half ashift’s worth of incoming parts at any one time.

The supplier has a relatively stable volume and range of parts, though theprecise mix of each part required can be determined a day or two ahead –overall volume and sequence are relatively fixed but mix and volumes per partare flexible. This distribution system also results in much higher transportutilization, as the capacity required can be fine tuned to relatively steady levels– and anticipated increases planned for – while at the same time realizing thebenefits of small lot JIT delivery. The extra cost of emergency shipments, whichare prevalent in a batch and queue system, is removed.

OrderingIn the ordering area it is necessary to move to a daily ordering system fromsuppliers. In the aftermarket parts area, for instance, this is done on a “sell oneorder one” basis rather than traditional standard reorder quantities with longlead times. These orders are then delivered to Toyota at predictable arrivaltimes in the warehouse, avoiding delays to lorries and warehouse staff alike.

At the same time a move from monthly to daily orders from the dealers isrequired so that a steady and relatively even flow of demand is inputted intoToyota’s system. The standard frequent deliveries of goods to dealers eliminatesthe need for emergency vehicle-off-road orders, order peaks and troughs as wellas the requirement to offload surplus stock at discount prices common inEurope and the USA.

Warehouse managementA similar type of logic to that applied in the factory is also used in thewarehouse involving:

• reduced bin sizes;

• storage by part type with frequently used parts near the front or aisleend;

• standard binning and picking routes for each part type;

• a division of the working day and tasks into standard work cycles;

• synchronized order-pick-pack-despatch and delivery steps for eachdelivery route (again a milk round) out to a group of local dealers;

• staggered outward delivery routes;

• controlled progress and irregularities through binning or picking ticketbundles for each cycle (preventing working ahead) and visual controlboards;

• the logging of irregularities and prioritization in order to conduct rootcause elimination of the most frequent problems to prevent recurrencesand hence improve the process.

IJPDLM27,3/4

160

The result of the warehouse management and ordering improvements is thatstock in Toyota’s regional distribution centre (RDC) is down from 24 to fourweeks, while the service rate and productivity are improved to three times asimilar, traditionally organized facility – with no automation! This gives a stockturn of 13 – equivalent to the best US supermarket chain!

DealersDaily delivery allows the dealer to reduce overall stock levels by well over halfwhile carrying a wider range of parts. It also improves service rates to waitingcustomers and eliminates delays in binning parts and cuts wasted walking andwaiting time for the mechanics picking up parts. Freed up space can be put torevenue-earning use. One of the key roles for the dealer service manager is tolevel the workflow and plan regular service work in the same manner as thewarehouse.

Network structureToyota uses a network of RDCs in Japan as well as the USA, handling 50-60,000part types. In Japan the process has been extended to the elimination of almostall parts stock from the dealer. Instead this stock is held at a local distributioncentre, which carries about 15,000 parts and can deliver parts to dealers fourtimes a day.

Dealers place timed orders on the system for regular servicing, which canflow at the right time from the RDC. They can get additional parts, requiredafter the car has been stripped down in time to complete the job within the day.This eliminates a further four weeks’ stock from the system. However, theability to do this depends on the density of the vehicles and dealers within ageographical region. The system works at its best within medium to high urbanpopulation densities.

The entire Toyota Parts Warehousing network in Japan (Figure 5) is part-owned by Toyota and part by its large dealer groups, each responsible for many

Figure 5.Toyota partsdistribution system inJapan

TMS slowmoving partswarehouse

33 KyohanPDCs

273 Kyohanbranches

4,776 Toyotadealers

Toyota andsuppliers

1-2 × day 4 × day

4 × day

4 × day

Lean logistics

161

dealer outlets within a Prefecture. Just as Toyota is dealing directly with alimited number of parts suppliers it is also dealing with a limited number ofdealer groups in Japan, and not with thousands of separately owned dealers asin the west. This makes shared ownership of the distribution chain and milkround delivery systems easier to organize.

Targeting improvements within the lean enterpriseThe basis of adding value rather than just cost is the elimination of the sevenwastes as discussed above[16-18]. However, as illustrated in the above example,this waste elimination needs to be on a wider scale than simply removing wasteinside organizations as advocated by traditional texts on the subject[21]. It isnecessary to address both the intra- and inter-company value (or waste) addingprocesses.

At present there is an ill-defined and ill-categorized tool kit to understand thevalue stream, although several researchers have developed individual tools[22-25]. In general these authors have tended to propose their creations as theanswer, rather than a part of the jigsaw. In addition the existing tools derivefrom functional ghettos and so, on their own, do not fit well with the more cross-functional tool box required by today’s best companies.

In order to do this, a seven-element “toolbox” of methods derived from avariety of functional or academic disciplines has been assembled and has beenapplied under the term value stream mapping[26]. The selection of which of theseven tools to use is made on the basis of the seven wastes prevalent in the valuestream concerned and the usefulness of each tool in helping to understand suchwastes. (A full description of the method can be found in reference[26].)

At this point it will be useful briefly to review the tools before providing acase example of their application in the UK distribution industry.

Process activity mappingThe first of the tools is the traditional industrial engineering tool, processactivity mapping (Figure 6). There are five stages to this general approach:

(1) the study of the flow of processes;(2) the identification of waste;(3) a consideration of whether the process can be rearranged into a more

efficient sequence;(4) a consideration of a better flow pattern involving different flow layout or

transportation routeing; and(5) a consideration of whether everything that is being done is really

necessary.In order to do this, several simple stages must be followed. First, a preliminaryanalysis of the process is undertaken, followed by the detailed recording of allthe required items in each process. The result of this is a map of the processunder consideration (Figure 6). As can be seen from this internal process

IJPDLM27,3/4

162

industry example, each step of the flow has been categorized from 1 to 23within a variety of types of activity (operation, transport, inspection andstorage). The machine or area used for each of these activities is recorded,together with the distance moved, time taken and number of people involved. Asimple flow chart of the types of activity being undertaken at any one time canthen be drafted.

After this the total distance moved, time taken and people involved can becalculated. This calculation is then recorded. The final diagram (Figure 6) canbe used as the basis for further analysis and subsequent improvement. This isoften achieved through the use of techniques such as 5W1H (asking Why doesan activity occur, Who does it, on What machine, Where, When and How). Thebasis of this approach is therefore to try to eliminate activities that areunnecessary, simplify others, combine others and seek sequence changes thatwill reduce waste.

Supply-chain response matrixThe origin of the second tool lies in the time compression and logisticsmovement and goes under a variety of names. It was used by New[22] and Forza

Figure 6.Process activitymapping – a processindustry example

Lean logistics

163

et al.[23] within a textile supply chain setting. In a more wide-ranging work,Beesley applied what he termed time-based process mapping to a range ofindustrial sectors including automotive, aerospace and construction[24]. Asimilar approach was adopted by Jessop and Jones in the electronics, food,clothing and automotive industries[25].

The fundamentals of this mapping approach, as shown in Figure 7, are thatit seeks to portray in a simple diagram the critical lead time constraints for aparticular process. In this case it is the cumulative lead time within adistribution company, its suppliers and its downstream retailer. In this figurethe horizontal measurements shows the lead time for the product bothinternally and externally. The vertical plot shows the average amount ofstanding inventory (in days) at specific points in the supply chain.

In this case the horizontal axis shows the cumulative lead time is 42 workingdays. In addition the vertical axis shows that another 99 working days ofmaterial is held within the system. Thus a total response time of 141 workingdays can be seen to be typical in this system. Once this is understood, each ofthe individual lead times and inventory amounts can be targeted forimprovement activity as was shown with the process activity mappingapproach.

Production variety funnelThe production variety funnel is shown in Figure 8. This approach originates inthe operations management area and has been applied by New in the textileindustry[22]. A similar method is IVAT analysis, which views internaloperations in companies as consisting of activities that conform to an I, V, A or

Figure 7.Supply chain responsematrix – a distribution

example

Retailer

First tier supplier

Retail premisesDistributionPicking

ShipPrimaryoperationsReceive

Cumulative inventory 99 days

41.5

82

1.5

10

Main store

Supplier lead time

40Companyoperations

Cumulative leadtime 42 days

Total:141 days

2

IJPDLM27,3/4

164

T shape[27]. “I” plants consist of unidirectional, unvarying production ofmultiple identical items such as a chemical plant. “V” plants consist of a limitednumber of raw materials processed into a wide variety of finished products in agenerally diverging pattern. “V” plants are typical in textile and metalfabrication industries. “A” plants, in contrast, have many raw materials and alimited range of finished products with different streams of raw materials usingdifferent facilities. Such plants are typical in the aerospace or other majorassembly industries. Lastly, “T” plants have a wide combination of productsfrom a restricted number of components made into semi-processed parts heldready for a wide range of customer-demanded final versions. This type of site istypical in electronics and household appliance industries.

Such a delineation using the production variety funnel (Figure 8) allows themapper to understand how the firm or the value stream operates and theaccompanying complexity that has to be managed. In addition, such a mappingprocess helps potential research clients to understand the similarities anddifferences of their industry compared to one that may have been more widelyresearched. The approach can be very useful in helping to decide where totarget inventory reduction and making changes to the processing of products. Itis also useful in gaining an overview of the company or supply chain beingstudied.

Quality filter mappingThe fourth value stream mapping tool is the quality filter mapping approach.This is a new tool that is designed to identify where quality problems exist inthe value stream. The resulting map itself shows where three different types ofquality defects occur in the value stream (Figure 9). The first of these is productdefects. Product defects are defined here as defects in goods produced that are

Figure 8.Production varietyfunnel – a brewingindustry case

Pack size

Can height and size

Actual gravity brew

Core high gravity brews

Variantand leadtimeanalysis

Materialsmixing Fermentation Conditioning Filling and packaging

Process path

Lean logistics

165

not caught by in-line or end of line inspection and are therefore passed on tocustomers.

The second type of quality defect is what may be termed service defects.Service defects are problems given to a customer that are not directly related tothe goods themselves, but due to the accompanying level of service. The mostimportant of these service defects are inappropriate delivery (late or early)together with incorrect paperwork or documentation. In other words suchdefects include any problems that customers are caused that are not concernedwith faulty physical goods.

The third type of defect is internal scrap. Internal scrap refers to defectsmade within a company that have been caught by in-line or end of lineinspection. The in-line inspection methods will vary and can consist oftraditional product inspection, statistical process control or through poke yokeor foolproofing devices. Such methods may help, used respectively, to inspect inquality, control quality within given tolerances or prevent defects from beingmade in the first place.

Each of these three types of defect is then mapped along the value stream. Inthe automotive example given (Figure 9), this value stream is seen to consist ofdistributor, assembler, first-tier supplier, second-tier supplier, third-tier supplierand raw material source. This approach has clear advantages in identifying

Figure 9.The quality filter

mapping approach: anautomotive example

1,000,000

100,000

10,000

1,000

100

10

1

Distrib

utor

Assem

bler

First t

ier

Secon

d tie

r

Third

tier

Raw m

ater

ials

Defect rate (PPM)

KeyProduct defectsService defects (paperwork, delivery, etc.)Scrap defects

IJPDLM27,3/4

166

where defects are occurring and hence in identifying problems, inefficienciesand wasted effort. This information can then be used for subsequentimprovement activity.

Demand amplification mappingDemand amplification mapping has its roots in the systems dynamics work ofForrester and Burbidge[28,29]. What has now become known as the Forrestereffect was first described in a Harvard Business Review article in 1958 by JayForrester[28]. This effect is primarily linked to delays and poor decision makingconcerning information and material flow. The Burbidge effect is linked to theLaw of Industrial Dynamics, which states “if demand is transmitted along aseries of inventories using stock control ordering, then the amplification ofdemand variation will increase with each transfer”[29]. Thus, in unmodifiedsupply chains, excess inventory, production, labour and capacity are generallyfound as a result. It is therefore quite likely that on many day-to-day occasionsmanufacturers will be unable to satisfy retail demand even though they are onaverage able to produce more goods than those being sold. Within a supply-chain setting, manufacturers have, as a result, sought to hold, in some casessizeable, stocks to avoid such problems. Forrester likens this to driving anautomobile blindfolded with instructions being given by a passenger[28].

The basis of the mapping tool in the value stream setting is given in Figure10. In this instance an FMCG (fast moving consumer goods) food product is

Figure 10.Demand amplificationmapping: an FMCGfood product example

180

160

140

120

100

80

60

40

20

0

28.8

.93

25.9

.93

23.1

0.93

20.11

.93

18.1

2.93

15.1

.94

12.2

.94

9.4.

94

7.5.

94

4.6.

94

2.7.

94

30.7

.94

27.8

.94

24.9

.94

22.1

0.94

19.11

.94

19.11

.94

17.1

2.94

Retail sales

Weeks

Pallets

KeyRetail salesOrders to suppliers

Source: [30]

Lean logistics

167

being mapped through its distribution through a leading UK supermarketretailer. In this simple example two curves are plotted. The first, the solid line,represents the actual consumer sales as recorded by EPOS (electronic point ofsale) data. The second, the dotted line, represents the orders placed to thesupplier to fulfil this demand. As can be seen, the variability of consumer salesis far lower than for supplier orders. It is also possible subsequently to map thisproduct further upstream. An example may be the manufacturing plant of thecleaning products company or even the demand they place on their rawmaterial suppliers.

This simple analysis tool can then be used to show how demand changesalong the value stream within varying time buckets. This information can thenbe used as the basis for decision making and further analysis to try to redesignthe value stream configuration, manage the fluctuations, reduce the fluctuationor to set up dual mode solutions where regular demand can be managed in oneway and exceptional or promotions demand can be managed in a separate way.

Decision point analysisDecision point analysis is of particular use for “T” plants or value streams thatexhibit similar features, although it may be used in other types of industries.The decision point is the point in the value stream where actual demand pullgives way to a forecast-driven push. In other words, it is the point at whichproducts stop being made due to actual demand and start being made againstforecasts alone[31]. Thus with reference to Figure 11, an example from theFMCG industry, the decision point can be at any point from regionaldistribution centres to national distribution centres through to any point insidethe manufacturer or indeed at any tier within the supply chain.

Figure 11.Decision point analysis:

an FMCG example

Assemble to order

Make to order

Purchase and make to order

The pullMake and ship to local stock

Make and ship to national stock

Purchasedgoods

Work inprogress

Finishedgoods

Nationaldistribution

centre

Regionaldistribution

centre

FactoryStores

The push

KeySupply stream decision pointForecast drivenCustomer driven

Suppliers

Source: [32]

IJPDLM27,3/4

168

Gaining a basic understanding of where this point lies is useful for two reasons.First, at the immediate level, it becomes possible to assess the processes thatoperate both downstream and upstream from this point. The purpose of this isto ensure that they are aligned with the relevant pull or push philosophy.Second, at a more fundamental and longer-term level, it is possible to designvarious “what if” scenarios to view the operation of the value stream if thedecision point is moved. This may allow for a better design of the value stream.

Physical structurePhysical structure mapping is a new tool that has been found to be useful inunderstanding what a particular value stream looks like at an overview orindustry level. This knowledge is helpful in understanding what the industrylooks like, how it operates and in particular in directing attention to areas thatmay not be receiving sufficient developmental attention.

The tool is illustrated in Figure 12 and can be seen to be split into two parts,namely: volume structure and cost structure. The first diagram (Figure 12a)

Figure 12.Physical structuremapping: an automotiveindustry example

Assembler

3rd

2nd

1st

3rd

2nd

1st

Assembler

Supplytiers

Tiers

Aftermarket

Rawmaterials

Support

Support

3rd

2nd

1st

Assembler

1st

2nd

3rd

Rawmaterials

SupportAftermarket

Support

(a) By number of firms involved

(b) By cost adding

Distributiontiers

Lean logistics

169

shows the structure of the industry according to the various tiers that exist inboth the supplier area and the distribution area, with the assembler located inthe middle point. In this simple example there are three supplier tiers as wellthree distribution tiers. In addition, the supplier area is seen to include rawmaterial sources and other support suppliers (such as tooling, capitalequipment and office supplies firms). These two sets of firms are not given a tierlevel as they can be seen to interact with the assembler as well as the othersupplier tiers.

The distribution area of the figure includes three tiers as well as a sectionrepresenting the after-market (in this case for spare parts) as well as variousother support organizations providing consumables and service items. Thiscomplete industry map therefore captures all the firms involved with the area ofeach part of the diagram proportional to the number of firms in each set.

The second diagram (Figure 12b) maps the industry in a similar way withthe same sets of organizations. However, instead of linking the area of the figureof the diagram to the number of firms involved it is directly linked to the value-adding process (or more strictly speaking to the cost-adding process). As can benoted in this automotive case, the major cost adding occurs within the rawmaterial firms, the first-tier suppliers and the assembler themselves. In this casedistribution costs are not significant.

However, the basis of use of this second figure is that it is then possible toanalyse the value adding that is required in the final product as it is sold to theconsumer. Thus value analysis tools employed by industrial engineers can befocused at the complete industry or supply-chain structure[21]. Such anapproach may result in a redesign of how the industry itself functions either ifthe sector is dominated by one firm such as Benetton or if all the firms in theindustry can be brought together to achieve such an aim. Thus, similarly to theprocess activity mapping tool discussed above, attempts can be made to try toeliminate activities that are unnecessary, to simplify others, to combine others,and to seek sequence changes that will reduce waste.

The use of this seven-element tool kit is not confined to any particulartheoretical approach to ultimate implementation. Thus the options can be leftopen at this stage whether to adopt a kaizen, kaikaku (major step changeimprovement) or business process re-engineering approach once the tools havebeen used[19,33,34].

A distribution caseIn order for the reader to gain a better understanding of the approach, a briefindustry case will be reviewed. The company involved is a highly profitableleading industrial distributor with over 60,000 products and an enviable recordfor customer service.

After undertaking preliminary discussions it was decided to focus on theupstream value stream to the point where goods were available for distributionby the firm. Nine products were chosen based on a Pareto analysis from oneparticular value stream, namely the lighting product range. Preliminary

IJPDLM27,3/4

170

interviews with key cross-functional staff showed that unnecessary inventory,defects, inappropriate processing and transportation were the most seriouswastes in the system. Based on the knowledge of the supporting research teamit was decided to adopt five of the tools: process activity mapping, supply-chainresponse matrix, quality filter mapping, demand amplification mapping anddecision point analysis.

The choice of these particular tools was based on two factors. First, of theavailable tools, it was felt that the five chosen would be most appropriate inidentifying and understanding the particular types of waste that had beensuggested by the firm as being present. Second, the use of further tools wouldprovide only small additional benefit and so they were not used in this case.

The on-site mapping work was carried out over a three-day period andproved that each of the tools were of value in analysing the selected valuestreams. An example of how the interplay of the tools was effective was that thesupply-chain response matrix suggested that the key priority for the firm wassupplier lead time reduction. However, when the data from the quality filtermapping were added it was found that the real issue was on-time deliveryrather than lead time reduction. Thus, if the supply-chain response matrix hadbeen used on its own it might have resulted in shorter lead times butexacerbated the true problem of on-time delivery.

The work assisted the firm to conclude that although there was no internalcrisis warranting radical change, there was considerable room forimprovement, particular regarding the relatively unresponsive suppliers. As aresult, attention has been paid to the setting up of a cross-functionally drivensupplier association[20] with six key suppliers in one product group area for thepurpose of supplier co-ordination and development. Within this supplierassociation is an awareness raising programme of why change is required,involving ongoing benchmarking. In addition, education and implementationare being carried out using devices such as vendor managed inventory, due dateperformance, milk rounds, self-certification, stabilized schedules and EDI.

The company found the ongoing mapping work very useful and one seniorexecutive noted that “the combination of mapping tools has provided aneffective means of mapping the[company’s] supply chain concentratingdiscussion/action on key issues”. Another described the work as “not rocketscience but down to earth common sense which has resulted in us setting up afollow up project which will be the most important thing we do between nowand the end of the century”. A conservative estimate of the savings that couldbe reaped as a result of the follow-on work was in excess of £10 million per year,equivalent to a 20 per cent improvement in profitability. Such work will helpposition the company within a truly lean logistics value stream.

Where do we go from here?This paper has sought to provide a framework for a new way of thinking aboutthe supply chain termed lean logistics. Lean logistics takes its fundamentalphilosophy from the Toyota production system and is based around extended

Lean logistics

171

TPS right along supply chains from customers right back to raw materialextraction. Such an approach has been developed in order to overcome some ofthe fragmentation problems of traditional functional and business thinking.Within lean logistics the key concepts of value, value streams, flow, pull andperfection have been discussed. Each of these key concepts was illustratedwithin a case study of Toyota’s parts supply system.

The starting point for Toyota in developing such a system was its work inunderstanding waste and inefficiency in existing value streams. In order for thereader to see how this type of work may be done, a new framework waspresented here called value stream mapping. This approach can been used todiagnose waste and help organizations and value streams to make subsequentradical or incremental improvements. A distribution case has been provided toillustrate the method as a basis for an extended improvement programme thatthe company is now undertaking.

As this paper has been designed to be largely conceptual in nature it may beuseful to suggest a number of missing pieces or unresolved issues in ourknowledge and application of lean logistics. This offers a basis for future work.The questions and issues are:

• How can we define the value and logistical implications of the newproduct and service bundles required by consumers?

• How do we define the value stream when in reality value streamsoverlap, split and coalesce?

• How do we create an effective “wake up call” for value streams to makethe necessary improvements?

• What are the right warehouse types, locations and operating systems forlean logistics?

• How do we design value streams that dampen chaos rather than amplifyit?

• How do we understand the change management steps required to movefrom batch and queue to flow and pull logic?

While this list is not exhaustive it is designed to capture some of the importantissues as supply chain management moves towards value stream management,industry moves from a mass production paradigm to a lean productionapproach and the unit of competitive advantage moves from the company to thecomplete lean enterprise. The solutions to these questions are not readilyapparent but are the subject of a current research initiative at the LeanEnterprise Research Centre. This is within an industry-focused programme ofresearch based around a small group of firms dedicated to developingthemselves and their respective value streams into a world class status.

References1. Peters, T. and Waterman, R., In Search of Excellence, Harper Collins, New York, NY, 1982.2. Peters, T., Liberation Management, Alfred Knopf, New York, NY, 1992.

IJPDLM27,3/4

172

3. Deming, E., Out of Crisis, Massachusetts Institute of Technology Centre for AdvancedEngineering Study, Cambridge, MA, 1982.

4. Juran, J., Juran on Leadership for Quality, Free Press, New York, NY, 1989.5. Crosby, P., Quality is Free, McGraw Hill, New York, NY, 1979.6. Hill, T., Manufacturing Strategy: Text and Cases, Macmillan, Basingstoke, 1995.7. Ellram, L., “A managerial guideline for the development and implementation of purchasing

partnerships”, International Journal of Purchasing and Materials Management, Vol. 27 No.2, Summer 1991, pp. 2-8.

8. Partnership Sourcing Ltd, Making Partnership Sourcing Happen, Partnership SourcingLtd, London, 1992.

9. Hammer, M. and Champy, C., Reengineering the Corporation, Harper Business, New York,NY, 1993.

10. De Meyer, A. and Wittenberg-Cox, A., Creating Product Value: Putting Manufacturing onthe Strategic Agenda, Pitman Publishing, London, 1992.

11. Southey, P., “Partnership sourcing failures – some lessons for the 90s”, Proceedings of theInternational Symposium on Logistics, University of Nottingham, 6-7 July 1993, pp. 341-7.

12. Sako, M., Lamming, R. and Helper, S., “Supplier relations in the UK car industry: goodnews-bad news”, European Journal of Purchasing and Supply Management, Vol. 1 No. 4,1994, pp. 237-48.

13. Schonberger, R., Building A Chain of Customers Linking Business Functions to Create theWorld Class Company, The Free Press, New York, NY, 1990.

14. Womack, J. and Jones, D., “Beyond Toyota: how to root out waste and pursue perfection”,Harvard Business Review, September-October 1996.

15. Womack, J. and Jones, D., “From lean production to the lean enterprise”, Harvard BusinessReview, February-March, 1994.

16. Japan Management Association, Kanban: Just-in-Time at Toyota, Productivity Press,Cambridge, MA, 1985.

17. Shingo, S., A Study of the Toyota Production System from an Industrial EngineeringViewpoint, Productivity Press, Cambridge, MA, 1989.

18. Monden, Y., Toyota Production System: An Integrated Approach to Just-In-Time, 2ndEdition, Industrial Engineering and Management Press, Norcross, GA, 1993.

19. Womack, J. and Jones, D., Lean Thinking: Banish Waste and Create Wealth in YourCorporation, Simon & Schuster, New York, NY, 1996.

20. Hines, P., Creating World Class Suppliers: Unlocking Mutual Competitive Advantage,Pitman Publishing, London, 1994.

21. Miles, L., Techniques of Value Analysis and Engineering, McGraw-Hill, New York, NY,1961.

22. New, C., “The use of throughput efficiency as a key performance measure for the newmanufacturing era”, The International Journal of Logistics Management, Vol. 4 No. 2, 1993,pp. 95-104.

23. Forza, C., Vinelli, A. and Filippini R., “Telecommunication services for quick response inthe textile-apparel industry”, Proceedings of the 1st International Symposium on Logistics,The University of Nottingham, 1993, pp. 119-26.

24. Beesley, A., “A need for time based process mapping and its application in procurement”,Proceedings of the 3rd Annual IPSERA Conference, University of Glamorgan, 1994, pp. 41-56.

25. Jessop, D. and Jones, O., “Value stream process modelling: a methodology for creatingcompetitive advantage”, Proceedings of the 4th Annual IPSERA Conference, University ofBirmingham, 1995.

Lean logistics

173

26. Hines, P. and Rich, N., “The seven value stream mapping tools”, International Journal ofProduction and Operations Management, Vol. 17 No. 1, 1997, pp. 44-62.

27. Macbeth, D. and Ferguson, N., Partnership Sourcing: An Integrated Supply ChainApproach, Pitman Publishing, London, 1994.

28. Forrester, J., “Industrial dynamics: a major breakthrough for decision makers”, HarvardBusiness Review, July-August 1958, pp. 37-66.

29. Burbidge, J., “Automated production control with a simulation capability”, Proceedings ofthe IFIP Conference, Copenhagen, 1984.

30. Jones, O. (1995), “Forrester mapping project”, Supply Chain Development Programme I,Workshop 6, Hertford, 25-26 January, Cardiff Business School.

31. Hoekstra, S. and Romme, S., Towards Integrated Logistics Structure – DevelopingCustomer-oriented Goods Flows, McGraw-Hill, New York, NY, 1992.

32. Rich, N. (1995), “Supply team responsiveness project”, Supply Chain DevelopmentProgramme I, Workshop 6, Hertford, 25-26 January, Cardiff Business School.

33. Imai, M., Kaizen: The Key to Japan’s Competitive Success, McGraw-Hill, New York, NY,1986.

34. Hammer, M., “Re-engineering work: don’t automate, obliterate”, Harvard Business Review,July-August, 1990, pp. 104-12.