INT. J. PROD. RES., 2000, VOL. 38, NO.

17, 4061 ± 4070

Lean, agile or leagile? Matching your supply chain to the marketplace

RACHEL MASON-JONESy, BEN NAYLORz and

DENIS R. TOWILLz*

Many enterprises have pursued the lean thinking paradigm to improve the e -
ciency of their business processes. More recently, the agile manufacturing para-
digm has been highlighted as an alternative to, and possibly an improvement on,
leanness. In pursuing such arguments in isolation, the power of each paradigm
may be lost, which is basically that agile manufacturing is adopted where demand
is volatile, and lean manufacturing adopted where there is a stable demand.
However, in some situations it is advisable to utilize a diŒerent paradigm on
either side of the material ¯ ow de-coupling point to enable a total supply chain
strategy. This approach we have termed the Leagile Paradigm. This paper there-
fore considers the eŒect of the marketplace environment on strategy selection to
ensure optimal supply chain performance. Real-world case studies in the mechan-
ical precision products, carpet making, and electronic products market sectors
demonstrate the new approach to matching supply chain design to the actual
needs of the marketplace.

1. Introduction
The success and failure of supply chains are ultimately determined in the market-
place by the end consumer. Getting the right product, at the right price, at the right
time to the consumer is not only the lynchpin to competitive success but also the key
to survival. Hence, customer satisfaction and marketplace understanding are crucial
elements for consideration when attempting to establish a new supply chain strategy.
Only when the constraints of the marketplace are understood can an enterprise
attempt to develop a strategy that will meet the needs of both the supply chain
and the end consumer.
Supply chain performance improvement initiatives strive to match supply and
demand, thereby driving down costs simultaneously with improving customer satis-
faction. This invariably requires uncertainty within the supply chain to be reduced as
much as practicable so as to facilitate a more predictable upstream demand.
Sometimes, however, uncertainty is impossible to remove from the supply chain
due to the type of product involved. If a product is highly fashionable then, by its
very nature, its demand will be unpredictable. Hence, speci® c supply chains are faced
with the situation where they have to accept uncertainty but need to develop a
strategy that enables them still to match supply and demand.
Those companies who design business strategies that acknowledge the presence
of uncertainty and provide mechanisms for pro-actively tackling it are rewarded by
an opportunity to enable best practice ahead of competitors whose responses are

{ Logistics Systems Dynamics Group, CardiŒUniversity, Wales, UK.

{ Logistics Systems Dynamics Group, CardiŒUniversity PO Box 907, CardiŒ, CF1 3YP,
Wales UK.
* To whom correspondence should be addressed.

International Journal of Production Research ISSN 0020± 7543 print/ISSN 1366± 588X online # 2000 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
4062 R. Mason-Jones et al.

purely reactive (Mason-Jones and Towill 1999). The message here is that supply
chains need to adopt a strategy that suits both their particular product and market-
place. This paper analyses the lean, agile and `leagile’ paradigms and their roles in
tackling diŒering marketplace uncertainty scenarios. It includes case studies illustrat-
ing the approach and representing real-world supply chains in three diŒerent market
sectors.

2. Supply chain uncertainty

Much uncertainty evident in supply chains is system induced and magni® ed by
the `Bullwhip eŒect’ , as opposed to being present in the marketplace. The supply
chain dynamics wave propagation observed by Forrester (1961) used to be called
`demand ampli® cation’ . Nowadays it is more commonly termed the `Bullwhip eŒect’
(Lee et al. 1997) due to the characteristic of increasingly magni® ed and hence wor-
sening behaviour observed upstream from the source of the disturbance. System
induced uncertainty is inherent within many supply chains due to the strategies
and relationships involved and is therefore within the direct control of the companies
involved. Hence, it is our experience that many of the detrimental eŒects of un-
certainty can be alleviated by working hard to reduce the system-induced eŒects
(Towill and McCullen 1999). Done properly, this leaves the supply chain to develop
a strategy that needs only to deal with marketplace uncertainty.
The combined eŒect of system-induced and marketplace uncertainty typically
leads to the type of Bullwhip eŒect supply chain dynamics shown in ® gure 1
(Fisher, 1997). This shows the resultant behaviour due to a system-imposed uncer-
tainty resulting from a supplier discount scheme operating in a retail supply chain.
As can be seen, the enticement of a discount oŒered to the retailer caused an unpre-

8
FA LSE DEMA ND
PEA K INDUCED
B Y SUPPLIER SHIPMENTS
DISCOUNT!
CASES OF CHICKEN NOODLE

6
SOUP (X105 )

4
A CTUA L
CONSUMPTION

0
0 10 20 30 40 50
July 1st June 30th
WEEKS

Figure 1. An example of a system-induced bullwhip eŒect in a commodities supply chain.

[Source: Fisher 1997.]
 Lean, agile or leagile? 4063

dictable change in behaviour that was magni® ed throughout the chain by the
Bullwhip eŒect. This produced a typical dynamic pro® le with demand being ampli-
® ed as it is passed upstream. The resultant on-costs are considerable for all `players’
in the chain, including overtime, shift premiums, quality variances, and additional
distribution, handling and storage charges. However, the situation is made even
more inexplicable by the realization that true marketplace demand could have
been satis® ed with a very simple order placement pro® le consisting of just a few
ramps and plateaux. In other words, by responding to true marketplace demand, a
series of level schedules placed on suppliers would have su ced, a situation ideally
suited to lean supply (Suzaki 1987). Instead, the supply chain operated in the un-
necessarily costly agile mode.

3. Coping with diŒering product types

In his seminal paper, Fisher (1997) stated that when developing a supply chain
strategy that will facilitate matching supply and demand, the relationship between
product type, supply chain and sales predictability is pivotal to ensuring that the
optimal approach is adopted. He classi® ed products into two generic types, fashion
and commodities.

. Fashion products have a short life cycle and high demand uncertainty, there-
fore exposing the supply chain to the risks of both stock out and obsolescence.
A good example of a fashion product is trendy clothing. The challenge faced
by a supply chain delivering fashion products is to develop a strategy that will
improve the match between supply and demand and enable the companies to
respond faster to the marketplace.
. Commodities that are basic products, such as tinned soups, have relatively
long life cycles and have low demand uncertainty due to the fact they tend
to be well-established products with a known consumption pattern. The driv-
ing force for basic product supply chains is therefore cost reduction.
Note that in terms of Hill’s (1993) manufacturing strategy metrics, there is a con-
siderable diŒerence between the two groups of products. For example, in the classi-
® cation matrix of ® gure 2, the market winner for fashion products is availability,
whereas the market winner for commodities is price. Quality and lead-time are
market quali® ers in both cases: price and availability are market quali® ers for fash-
ion products and commodities respectively. What ® gure 2 emphasizes is that the
supply chain must excel at the market winner metrics and be highly competitive at
the market quali® er metrics, i.e. the minimum standard for entry into the market-
place.
These two product types respond to distinctly diŒerent marketplace pressures
and hence require a diŒerent supply chain approach to address their speci® c
characteristics. Only through understanding the particular characteristics of the
product type, marketplace requirements and management challenges can the correct
supply chain strategy be designed to ensure optimal performance and to establish
competitive advantage. This can be achieved via developing strategies that will
reduce the eŒect of the system-induced uncertainty, thereby reducing the Bullwhip
eŒect and, at the same time, actively coping with the particular marketplace un-
certainty pressures.
4064 R. Mason-Jones et al.

· Quality
Fashion Goods

· Price · Service Level

· Lead Time

· Quality
Commodities

· Lead Time · Price

· Service Level

Market Market
Qualifers Winners

Figure 2. An example of using the classi® cation matrix based on market winners and market
quali® ers. [Source: the present authors.]

4. Matching supply chain strategies with product type

Many organizations have adopted the lean thinking paradigm (Womack and
Jones 1994) in their drive to optimize performance and improve competitive posi-
tion. Recently, the agile manufacturing paradigm has been highlighted as an alter-
native to leanness (Richards 1996). It has also been suggested in some quarters that
agility is the next step after leanness. This could mean that, once leanness has been
achieved, an enterprise should strive for agility or even that agility should be the goal
of an enterprise and leanness as a primary objective should be forgotten. These
discussions oversimplify the situation as they fail to take into consideration the
generic product type and hence the business environment and response requirements
needed to match adequately supply chain design to the required structure.
The following de® nitions relate the agile and lean paradigms to supply chain
strategies and were developed to emphasize the distinguishing features of each
(Naylor et al. 1999).
. Agility means using market knowledge and a virtual corporation to exploit
pro® table opportunities in a volatile marketplace
. Leanness means developing a value stream to eliminate all waste, including
time, and to ensure a level schedule.
It should be noted that what may be regarded as `waste’ in lean production may
conversely be essential in agile production. As McHugh et al. (1995) have empha-
sized, one example is the question of capacity requirements. In lean production, the
customer buys speci® c products, whereas in agile production the customer reserves
capacity that may additionally need to be made available at very short notice.
As can be seen from the above de® nitions, the commodities are very well suited to
the lean environment as demand is relatively predictable and therefore facilitates the
level schedule requirements necessary for a lean supply chain (Suzaki 1987).
 Lean, agile or leagile? 4065

Conversely the characteristics of fashion products are more suited to the agile envir-
onment where the unpredictability of the demand is accepted as a business risk and
the strategy is developed to optimize performance in such an arena. A blanket
approach across the whole supply chain may, however, not be appropriate.
Leanness and agility can sometimes be combined with the strategic use of a de-
coupling point, thereby capitalizing on the bene® ts of both paradigms as shown in
® gure 3. Thus, there are some instances where there is an economic justi® cation for
engineering a `Leagile’ supply chain, thereby getting the best of both worlds.
This combined approach is known as `Leagility’ and, as a consequence, the
supply chain can thereby adopt a lean manufacturing approach upstream, enabling
a level schedule and opening up an opportunity to drive down costs upstream while
simultaneously still ensuring that downstream of the de-coupling point there is an
agile response capable of delivering to an unpredictable marketplace. The formal
de® nitions required are as follows.

`Leagile is the combination of the lean and agile paradigms within a total supply
chain strategy by positioning the decoupling point so as to best suit the need for
responding to a volatile demand downstream yet providing level scheduling
upstream from the marketplace.’ (Naylor et al. 1997)
`The decoupling point is the point in the material ¯ ow streams to which the
customer’ s order penetrates. It is here where order-driven and the forecast-
driven activities meet. As a rule, the decoupling point coincides with an import-
ant stock point ± in control terms a main stock point ± from which the customer
has to be supplied.’ (Hoekstra and Romme 1992)
customer
Satisfied
Material
supply

Lean P rocesses

(A) Lean Supply

customer
Material

Satisfied
supply

Agile Processes

(B) Agile Supply

Deco upling
poin t
customer
Material

Satisfied
supply

Lean A g il e
P ro ce ss e s P r o ce ss e s

(C) Leagile Supply

Figure 3. Block diagrams representing lean, agile, and leagile supply. [Source: the present
authors.]
4066 R. Mason-Jones et al.

Why call it Leagile? First, because in a Leagile supply chain, Lean material ¯ ow is
upstream of Agile material ¯ ow. Secondly, because to succeed as an Agile process it
must be fully documented, understood and engineered. This is readily enabled by
initially engineering a Lean process and then adapting it by removing speci® c con-
straints and capacity limitations, thus enabling Agility. Hence, Lean precedes Agile
on two counts; geographically and temporally.

5. Minimize uncertainty to maximize competitive advantage

Despite the diŒerences in the types of market uncertainty present in the lean,
agile and leagile paradigms, the Bullwhip mechanism and resultant detrimental
system induced uncertainties are the same for all the approaches. Therefore, it
does not matter which paradigm is adopted, the system-induced uncertainty eŒects
can still seriously hamper the eŒectiveness of the strategy. Therefore, it is crucial that
system-induced uncertainty is reduced to ensure that the performance opportunities
available via implementing a particular strategy are fully realized.
Fortunately, the removal of system-induced uncertainty is greatly aided by engin-
eering streamlined material ¯ ow, and thus designing in good practise. A suitable set
of tried and tested rules for simplifying and streamlining material ¯ ow is available as
shown in table 1 (Towill 1999). These rules have been derived from a wide-ranging
set of simulation studies and from industrial experiments and observations, and have
clearly been adopted in whole or in part in the Case Studies that appear later in this
paper. It is also possible to consolidate the rules into the following powerful set of
Material Flow Control Principles, which have already found widespread application
in supply chain BPR Programmes of proven eŒectiveness.
(1) Selection of good Decision Support Systems: if process lead times are reliable
and operations information of high quality, then good, robust control
systems can also be very simple.
(2) Slashing of material ¯ ow and information ¯ ow lead times: reduction of these is
within the technological and organizational remit of individual echelons,
thus requiring mainly internal action.

Rule 1 Only make products which you can quickly despatch and invoice to customers.
Rule 2 Only make in one period those components you need for assembly in the next
period.
Rule 3 Minimize the material throughput time, i.e. compress all lead times.
Rule 4 Use the shortest planning period, i.e. the smallest run quantity that can be managed
e ciently.
Rule 5 Only take deliveries from suppliers in small batches as and when needed for
processing or assembly.
Rule 6 Synchronize `Time Buckets’ throughout the Chain.
Rule 7 Form natural clusters of products and design processes appropriate to each value
stream.
Rule 8 Eliminate all process uncertainties.
Rule 9 Understand document, simplify and only then optimize (UDSO) the supply chain.
Rule 10 Streamline and make highly visible all information ¯ ows.
Rule 11 Use only proven simple but robust Decision Support Systems.
Rule 12 The Business process target is the seamless supply chain, i.e. all players `Think and
act as one.’

Table 1. Twelve rules for simplifying material ¯ ow necessary in lean, agile and leagile supply
chains. [Source: Towill (1999).]
 Lean, agile or leagile? 4067

(3) The widespread provision and integrity of operations information: the quality
and quantity of data available throughout the extended enterprise are, how-
ever, a political issue to be addressed and overcome by the supply chain
product champion.
(4) Elimination of redundant echelons: this removes a source of distortion and
delay but can give rise to ownership/political problems, which again need
solution by the supply chain product champion.
Note that the problem of `wide ownership’ (especially of sales data) requires an
attitudinal change on the part of all supply chain `players’ . Otherwise, the rapid
feedback of vital information from further along the chain will be inhibited and
dynamic performance considerably worsened. For example, actual sales data at
one echelon need to be identi® ed and transmitted alongside ® rm orders when
placed higher up the chain so that false demand signals can be discounted
(Wikner et al. 1991). We now proceed to demonstrate the `Lean’ , `Agile’ , and
`Leagile’ concepts applied to real-world supply chains.

6. Case Study A: the lean global supply chain: precision mechanical products
(source: Towill and McCullen 1999)
The OEM has three UK factories that manufacture complex mechanical systems
and export mainly to the USA, Japan, Korea and Europe. There are approximately
300 major product families, 1500 minor products and around 2000 regularly stocked
spares. The di culty of the materials management task is compounded by the local
mix of systems and components businesses. Due to delays in the supply chain, poor
coordination, and bad decision-making, overseas stocks intended to provide oŒ-the-
shelf availability were signi® cantly higher than planned. At the same time, customer
service fell far short of market requirements due to the associated build-up of stocks
being out-of-phase with market demand. In response to this situation, the company
having recognized that improving forecast accuracy alone would not in itself solve
these problems, instigated a long-term global supply chain BPR Programme. Hence,
the execution of a major re-engineering programme requiring sustained eŒort over a
period of years.
The BPR Programme has directly reduced manufacturing lead-times via kanban
line side replenishment and changing over to unit ¯ ow assembly. Furthermore, the
UK factories are now linked directly to international customer demand via an inte-
grated Distribution Requirements Planning (DRP) system. The DRP generates an
integrated production demand to cover customer requirements, replenishments, and
forecast sales. Time compression throughout the chain enables the DRP to run on
weekly (not monthly) time buckets. Consequently, present day planning is speeded
up and is much more closely related to true demand. The cumulative information
processing and manufacturing delays have been reduced from 23 weeks to 2 weeks
(over 90% time compression). This, in turn, has enabled a substantial revision of the
decision rules used to manage system inventory. These rules are now simpler and
more robust. Consequently, physical distribution in the global supply chain is now
much more streamlined and synchronized with actual requirements.
The BPR Programme visibly embraces the Material Flow Control Principles of
time compression; echelon elimination; transparent information ¯ ow and selection
of appropriate system controls. Observed results as a consequence of implementing
change include 50% reduction in both demand ampli® cation and stock ¯ uctuations
4068 R. Mason-Jones et al.

across a sampled range of products. There has also already been a doubling in stock
turns of global inventory. Moreover, these performance metrics are still improving
after two years of operation of the new system, with every indication that yet further
signi® cant gains may be expected. Customer service performance is greatly improved
as a result of the BPR Programme. Finally, the average number of days late against
customer due date has also been reduced by 90% and delivery variation similarly
reduced by 74% .

7. Case Study B: parallel, lean and agile supply chains: the USA carpetmaker
(source: Johansson et al. 1993)
There is a limit to the advances in performance enabled by a business becoming
`lean’ through eliminating obvious muda. A good example, where substantial inno-
vative process re-engineering is additionally needed, is the carpet industry. Here, the
traditional supply chain requires some 16 weeks from raw material supply to the
carpet being ® tted in the end-customers residence. This total cycle time covers six
major processes that add value. Elimination of muda reduces the cycle time to about
four weeks. Impressive though this 75% reduction may be, it hardly helps carpet
manufacturers respond in an agile mode to satisfy the current customer-imposed
time windowÐ this is one week from ordering a speci® c size, colour, pattern, and
qualify of carpet, to ® tting in the house. The one week cycle time means that the
production process must be totally re-engineered so that manufacture is completed
within three days, which is a far cry from the best `lean’ performance of four weeks.
The outcome of this marketplace pressure was a technology breakthrough with
the invention of a revolutionary new process, named Solution Dyed Nylon. This
enabled the ® bre to be dyed prior to being woven into the carpet. Together with
other essential re-engineering to reduce changeover times by 95% and downtime by
50% , plus integration of support processes, and streamlined material ¯ ow, the target
cycle time window was achieved, coupled with the additional bene® t of a slightly
reduced cost to the end customer. However, a Pareto Curve based product analysis
showed that the Agile Response Mode would only be justi® able for the top 10% of
the product range, which contributed 52% of the volume leaving the mill. The
remaining products have now been rationalized and are provided via an alternative
Lean Supply Chain that does not oŒer the same one-week guaranteed delivery. This
illustrates the importance of aligning the Agile Supply Chain with the business
priorities driving the change.

8. Case Study C: the leagile global supply chain: electronics products (source:
Naylor et al. 1997)
The Case Study company is a global player in a complex supply chain with at
least ® ve major echelons ranging from the hundreds of component suppliers through
® ve sub-assembly businesses to four ® nal assemblers. Depending on the particular
product sold, goods may be delivered directly to customers, or via their administra-
tion centres, or via independent distribution centres and thousands of authorized
dealers. In the mid-1980s, the OEM took the initiative in setting up a series of
improvement programmes, covering a period of a decade, leading to the establish-
ment of a leagile supply chain circa 1993. Initially, the supply chain consisted of a
large number of interacting but unintegrated `players’ . The ® rst re-engineering stage
consisted of eliminating OEM muda and reducing plant lead-times, followed by
vendors participating inconsequential `best practice’ initiatives including JIT.
 Lean, agile or leagile? 4069

The OEM then developed a simpli® ed but integrated approach to materials

planning at the supply chain level. This was supported by the simultaneous EDI
transmission of requirements to all plants. The result was that more of the total
chain was able to operate in the pull mode. Finally, this stage of the BPR
Programme was completed by EDI integration of the whole vendor base into the
supply chain, and hence greatly reduced lead times in the planning loop. Thus, the
OEM has visibly exploited all four Material Flow Control Principles within the BPR
Programme. To achieve leagility the de-coupling point was located at the ® nal
assembler, an action that usually requires associated product rationalization (Lee
and Sasser 1995). Speci® c products are now pulled by current sales demand whilst
upstream of the de-coupling point suppliers now work to level schedules, i.e. short
term variations of roughly §10% about the long term trend. The lean part of the
supply chain is demonstrably much faster (by a factor of about 10 to 1 with all the
consequential bottom-line bene® ts, Towill 1996). This is very important, since new
component technology could otherwise result in rapid and costly obsolescence even
upstream of the decouplng point.

9. Summary
We have seen that classifying supply chain design and operations according to
the Lean, Agile and Leagile Paradigms enables us to match the supply chain type
according to marketplace need. This results in three fundamental designs illustrated
in the real world by mechanical precision products (lean); carpet manufacture (agile);
and electronics products (leagile). Such a classi® cation proves clear rules for supply
chain engineering for each market segment. This enables us to apply lean principles,
agile principles and leagile principles according to the real needs of the speci® c
supply chain. However, for all three supply chain types it is essential to remove
system-induced uncertainty, as typi® ed by the `Bullwhip’ eŒect. This elimination is
greatly assisted via the proven Material Flow Control Principles consolidated from a
basic simpli® cation checklist, and which are visibly applied in all three Case Studies.
Leagile supply chains already exist in the real world. What is important is to recog-
nize when the new paradigm is the best way forward for a particular supply chain so
that it may be appropriately engineered from the outset.

Acknowledgements
Two of the authors (R. Mason-Jones and J. B. Naylor) were funded by EPSRC
Studentships during the execution of this research. This support is gratefully
acknowledged.

References
F ISHER, M . L ., 1997, The right supply chain for your product. Harvard Business Review,
March± April, pp. 105± 116.
F ORRESTER , J . W ., 1961, Industrial Dynamics (Cambridge, MA: MIT Press).
H ILL, T ., 1993, Manufacturing Strategy: Text and Cases, 2nd edn (London: Macmillan).
H OEKSTRA, S., and R OMME, J ., 1992, Integral Logistics Strictures: Developing Customer-orien-
tated Goods Flow (New York: Simon and Schuster).
J OHANSSON, H . J ., M CH UGH, P ., P ENDLEBURY, A . J ., and W HEELER, W . A ., 1993, Business
Process Re-engineering: Breakpoint Strategies for Market Dominance (Chichester, UK:
Wiley).
L EE, H . L ., and SASSER, M . M ., 1995, Product universatility and design for supply chain
management. Production Planning and Control, 6(3), 270± 277.
4070 Lean, agile or leagile?

L EE, H . L ., P ADMANABHAN, V., and W HANG, S., 1997, Information distortion in a supply
chain: the bullwhip eŒect. Management Science, 43(4), 546± 558.
M ASON-J ONES, R ., and T OWILL, D . R . 1999, Total cycle time compression and the agile supply
chain. Engineering Costs and Production Economics, 62, 61± 73.
M CH UGH, P ., M ERLI, G ., and W HEELER III , W . A ., 1995, Beyond Business Process Re-
Engineering ± Towards the Holonic Enterprise (New York: Wiley).
N AYLOR, J . B., N AIM, M . M ., and BERRY, D ., 1999, Leagility: integrating the lean and agile
manufacturing paradigm in the total supply chain, Engineering Costs and Production
Economics, 62, 107± 118.
R ICHARDS, C . W ., 1996, Agile manufacturing: beyond lean? Production and Inventory
Management Journal, 2nd Quarter, 60± 64.
SUZAKI, K ., 1987, The New Manufacturing Challenge (New York: The Free Press).
T OWILL, D . R ., 1996, Time compression and supply chain management ± a guided tour,
International Journal of Supply Chain Management, 1(1), 15± 27; 1999, Simplicity
wins ± twelve rules for designing eŒective supply chains. IOM Control Journal, 25(2),
9± 13.
T OWILL, D . R ., and M CC ULLEN , P ., 1999, The impact of an agile manufacturing programme
on supply chain dynamics. International Journal of Logistics in Manufacturing, 10(1),
83± 96.
W IKNER, J ., T OWILL, D . R ., and NAIM , M . M ., 1991, Smoothing supply chain dynamics.
International Journal of Production Economics, 22, 231± 248.
W OMACK, J . P ., and J ONES, D . T ., 1994, From lean production to the lean enterprise. Harvard
Business Review, March± April, 93± 103.