Long term capacity management by manit1111

VIEWS: 258 PAGES: 11

More Info
									                                          Int. J. Production Economics 69 (2001) 215}225

Long-term capacity management: Linking the perspectives
from manufacturing strategy and sales and operations planning
                              Jan Olhager*, Martin Rudberg, Joakim Wikner
                                                             (                                           (
               Department of Production Economics, IMIE, Linkoping Institute of Technology, SE-58183 Linkoping, Sweden
                                              Received 2 April 1998; accepted 2 February 1999


   E$cient long-term capacity management is vital to any manufacturing "rm. It has implications on competitive
performance in terms of cost, delivery speed, dependability and #exibility. In a manufacturing strategy, capacity is
a structural decision category, dealing with dynamic capacity expansion and reduction relative to the long-term changes
in demand levels. Sales and operations planning (S&OP) is the long-term planning of production levels relative to sales
within the framework of a manufacturing planning and control system. Within the S&OP, resource planning is used for
determining the appropriate capacity levels in order to support the production plan. Manufacturing strategy and sales
and operations planning provide two perspectives on long-term capacity management, raising and treating di!erent
issues. In this paper, we compare and link them in a framework for long-term capacity management.          2001 Elsevier
Science B.V. All rights reserved.

Keywords: Capacity management; Manufacturing strategy; Sales and operations planning; Production planning; Resource planning

1. Introduction                                                               city is most often treated at an aggregate level,
                                                                              dealing with key work centres rather than all indi-
   The management of capacity in a manufacturing                              vidual resources and based on forecasts of product
"rm is often divided into three or four stages,                               families rather than of individual products. More
ranging from long-term capacity planning to                                   speci"cally, long-term capacity management is
short-term capacity control and execution. Inter-                             most interested with the capacities that take a long
mediate capacity management is related to rough-                              time to change, either to acquire new capacity or to
cut capacity planning, linked to the master sched-                            reduce capacity levels. Typically, the planning hor-
ule, and capacity requirements planning, linked to                            izon is 1}5 years and the planning period is
the material requirements plan. In this paper, we                             a month, at least for the "rst year, and then possibly
deal with the longest-term perspective. The relevant                          quarters or even longer periods.
issues treated at the long-term capacity manage-                                 The input to long-term capacity management is
ment level are related to determining when and by                             a sales plan, based on a demand forecast. Such
how much the capacity levels should change. Capa-                             a sales plan must at least cover the time perspective
                                                                              for acquiring new capacity or reducing the relevant
  * Corresponding author.                                                     capacity. The sales plan can be translated into
  E-mail address: jan.olhager@ipe.lin.se (J. Olhager).                        a corresponding capacity plan. However, decisions

0925-5273/01/$ - see front matter             2001 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 5 - 5 2 7 3 ( 9 9 ) 0 0 0 9 8 - 5
216                         J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225

regarding the production plan, in terms of e.g. pro-            category from a manufacturing strategy perspect-
duction smoothing, means that production will not               ive. Nowadays, many MPC systems support the
be identical to the sales plan. Furthermore, capa-              MRPII structure, i.e. manufacturing resource
city may be acquired or reduced at times or quant-              planning, from where the term sales and operations
ities other than those required by the sales plan.              planning originated. In early MRPII terminology,
Such issues are treated from two separate perspect-             the term production planning was used instead of
ives, on the one hand the manufacturing strategy                S&OP, but today the latter is being used to a larger
perspective and on the other the sales and opera-               extent. At the S&OP level, a production plan is
tions planning perspective, representing the highest            developed based on a sales plan. Here the issue is
planning level in a manufacturing planning and                  related to the production level relative to the de-
control system.                                                 mand level in various periods. The principal op-
   In a manufacturing strategy, capacity is con-                tions available are level, chase and mix (or
sidered as one of approximately seven decision cat-             combination). These are sometimes called planning
egories, for which the manufacturing "rm must                   strategies. Level means that a production rate is
have some policies. The other decision categories               established over the planning horizon, whereas
are facilities, production process, vertical integra-           chase implies that production matches demand in
tion, quality, organisation and personnel, and                  a manner such that all demand in a period is
"nally information or planning and control systems              produced in the same period (typically month). In
(see e.g. [1,2]). Capacity, facilities, production              the mix option a production rate is used for a few
process and vertical integration are considered                 periods and then changed. Any di!erence between
structural decision categories, dealing with the                the sales plan and the production plan will result in
long-term manufacturing operations structure. The               a corresponding inventory or backlog plan * an
latter three decision categories deal with the manu-            inventory plan in a make-to-stock situation and
facturing infrastructure. The capacity issues are re-           a backlog plan for make-to-order or engineer-to-
lated to the strategic relationship between capacity            order environments. The production plan is trans-
and demand levels, speci"cally translated into ca-              lated into a capacity requirements plan in terms of
pacity expansion or reduction strategies. The basic             aggregate resources in the so-called resource plann-
foundation for such a strategy is that capacity                 ing. Then, under- and over-capacities are identi"ed
comes in large, discrete steps rather than in small             in the resulting capacity plan. Thus, the focus of
increments. Therefore, it is of strategic importance            capacity management from an S&OP perspective is
to decide whether capacity should come "rst, i.e.               on the rate of production relative sales.
prior to expected changes in demand, or if capacity                Both perspectives deal with the long-term man-
should be acquired "rst when the corresponding                  agement of capacity, i.e. how to best utilise &slow-
level of demand has been acknowledged. There are                moving' resources for manufacturing operations.
three di!erent strategies in principle: lead, lag or            Still, the issues, objectives and options di!er. Ac-
track. Lead means that capacity is added in antici-             cording to the decision categories identi"ed by
pation of increasing demand, whereas lag means                  Hayes and Wheelwright [1] we have structural
the opposite. Track is a switching strategy, where              decisions related to capacity (expansion or reduc-
the di!erences between capacity and demand levels               tion strategies) and infrastructural decisions related
are kept to a minimum. There is a corresponding                 to MPC systems (e.g. sales and operations plann-
discussion for capacity reduction. Thus, the main               ing). This paper links these two decision categories
focus on capacity management from a manufactur-                 traditionally treated as belonging to di!erent
ing strategy perspective is the timing of capacity              groups of categories. This raises interesting ques-
changes.                                                        tions regarding the integration of capacity issues
   Sales and operations planning (S&OP) is the                  and whether they should be treated in sequence or
longest-term planning level in a manufacturing                  not. The purpose of this paper is to discuss the
planning and control (MPC) system. Thus, S&OP                   relationships between the two perspectives, in order
belongs to a manufacturing infrastructure decision              to integrate strategic and planning/control issues
                            J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225                    217

into a framework for long-term capacity manage-                 amount of capacity changes as given (which it typi-
ment. To start with, we treat each respective per-              cally is), the timing of capacity changes remains to
spective, "rst the manufacturing strategy view and              be decided upon.
then the S&OP view. Thereafter, we study the rela-
tionship and the interaction between the two per-               2.2. The timing of capacity changes
spectives, and "nally we propose a framework for
how to link them.                                                  The timing variable in a capacity strategy is
                                                                concerned with the balance between the (forecas-
                                                                ted) demand for capacity and the supply of capa-
                                                                city. If there is a capacity demand surplus the
2. Manufacturing strategy perspective
                                                                utilisation will be high, thus enabling a low cost
                                                                pro"le, but there is also a risk of loosing customers
  Hayes and Wheelwright [1] use three variables
                                                                due to e.g. long delivery lead times. A capacity
to describe a capacity strategy: the type of capacity
                                                                supply surplus on the other hand creates a higher
needed, the amount of capacity that should be ad-
                                                                cost pro"le but due to the surplus capacity it is
ded (or reduced), and the timing of capacity
                                                                easier to maintain high delivery reliability and #ex-
changes. Since the type of capacity strongly in#uen-
                                                                ibility. The capacity strategy can thus be ex-
ces the amount that is to be added or reduced, the
                                                                pressed as a trade-o! between high utilisation
"rst two are normally discussed together in the
                                                                (low cost pro"le) and maintaining a capacity
so-called sizing problem.
                                                                cushion (#exibility). Based on this, two &pure'
                                                                types of capacity strategies can be identi"ed,
2.1. The sizing of capacity changes                             usually referred to as leading demand (capacity
                                                                supply surplus) and lagging demand (capacity
   A central part of the sizing problem is scale                demand surplus) in line with the framework in [1].
considerations, where economies as well as dis-                 In their framework capacity strategy is, however,
economies of scale are weighted against each other,             only de"ned for increasing demand. Our frame-
which leads to concepts such as optimal step                    work contains both increasing and decreasing de-
change, optimal plant size, etc. Due to the inherent            mand patterns. In many cases it is not possible (nor
properties of most resources, capacity can normally             desirable) to maintain a pure strategy and a middle
only be changed in discrete steps with a consider-              way is chosen which contains aspects of both lead-
able lead-time. Adding a new machine or facility                ing demand and lagging demand. This approach
would probably mean a signi"cant capacity expan-                aims at "nding an e$cient trade-o! between the
sion making the capacity changing step-wise as                  pure strategies and is here referred to as tracking
illustrated in Figs. 1}3. Assuming the type and                 demand.

                               Fig. 1. Capacity leading demand (capacity supply surplus).
218                         J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225

                              Fig. 2. Capacity lagging demand (capacity demand surplus).

   The objective of the lead strategy approach is to            taken when the utilisation is still high. The guiding
maintain a cushion of capacity that e.g. can be used            rule in this case is that capacity should never exceed
to support volume #exibility and reliable lead                  demand, as illustrated in Fig. 2. Again, the reason is
times. If there is a positive trend in demand, capa-            to support the same order winning characteristics
city should be added in anticipation of demand, as              independently of the demand pattern. As discussed
shown in Fig. 1. A corresponding negative trend in              above in the case of the lead strategy, the lag label is
demand would mean that capacity should be de-                   also somewhat misleading. In a situation where
creased to the demand level when eliminating                    demand is declining, the decision to decrease capa-
a &step' of capacity. The guiding rule in this case is          city must be taken when the utilisation is still high
that the capacity always should be larger than or               and thus capacity is leading demand.
equal to demand, thereby supporting the same or-                   In some cases one of the two extreme strategies
der winning characteristics both in positive and                described above may be desirable and achievable,
negative trends of demand. Traditionally, this ap-              but in most cases there must be a trade-o! due to
proach has been referred to as a lead strategy since            e.g. di$culties in forecasting the future demand.
capacity leads the increase in demand to maintain               This combined approach is usually labelled track-
a capacity cushion. But, as can be seen in Fig. 1,              ing strategy. The objective is to track the demand as
when there is a decreasing demand pattern the                   close as possible, hence putting more emphasis on
objective of keeping a capacity cushion by necessity            the sizing problem. Reducing the size of the step
creates a lagging behaviour, not reducing capacity              changes facilitates a track strategy, which conse-
until a reduction can be conducted without endan-               quently minimises the deviations between demand
gering the maintenance of a capacity cushion.                   and capacity. An example of the tracking strategy is
Therefore, the label lead strategy may be mislead-              shown in Fig. 3 indicating the presence of both
ing, but it is used below anyhow since it is well               under- and over-capacity under this strategy.
established in the literature.                                     The three capacity strategies discussed above can
   The lag strategy is based on the objective of                be described as di!erent positions on a spectrum.
maintaining a high utilisation of resources. This is            The end-points of the spectrum would at one end
of particular interest in environments where price is           correspond to a situation where low cost, through
an order winner and thus the focus on a low cost                high utilisation, is the norm, and at the other end to
per unit is central. The basic principle is to produce          a situation where #exibility, based on over-capa-
as much as possible and still maintain full capacity            city, prevails. In between the end-points combina-
utilisation. Capacity should thus be added in reac-             tions, based on trade-o!s between the end-points,
tion to increasing demand. The lag strategy is,                 are found. Below, we discuss a corresponding spec-
however, challenging when demand is declining                   trum from a manufacturing planning and control
since the decision to reduce capacity should be                 perspective.
                           J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225                            219

                                          Fig. 3. Capacity tracking demand.

3. Sales and operations planning perspective                   long-term capacity planning via feedback from the
                                                               execution of the sales plan and production plan.
   In recent years the term sales and operations               Within the S&OP process the resource planning
planning (S&OP) has been more frequently used in               translates the production plan into a resource
articles and textbooks, although mostly related to             requirements plan, thereby reviewing the required
authors discussing MRPII or similar systems.                   capacity strategy for conformance. As tactical and
Some authors use S&OP interchangeably with the                 operational decisions are made each month, the
terms aggregate planning or production planning.               capacity plan gets executed or the need to review
However, we want to distinguish S&OP from the                  the capacity strategy becomes apparent.
others, where we see aggregate production plann-                  The issue in the S&OP process is to create a bal-
ing (APP) as a part of the S&OP process, which is              ance between the sales plan and the production
the long-term planning of production and sales                 plan. In doing this, planners must make decisions
relative the forecasted demand and the supply of               on marketing activities, output rates, employment
capacity.                                                      levels, inventory levels, backlogs, subcontracting,
                                                               etc. The decision options can roughly be divided
3.1. The S&OP process                                          into two types; those trying to modify demand to
                                                               match the production constraints, and those
   S&OP is often referred to as a fundamental that             modifying supply to match the sales plan (some-
maintains the balance between aggregate supply                 times referred to as aggressive and reactive alterna-
and aggregate demand, through monthly updates                  tives, see e.g. [5]). The supply option is normally
of the annual business plan, see e.g. [3]. Roughly,            further derived into the well-known methods where
S&OP can be divided into a sales plan (based on                production is trying to either level production or
forecasted demand) and a production plan, which                chase sales (i.e. forecasted demand), or some kind of
of course both a!ects inventory and/or order back-             combination or mix of the two (see Fig. 4). We label
log, and capacity requirements. The S&OP process               these methods planning strategies.
is the forum where di!erent functional strategies
meet for establishing a production plan that eco-
nomically serves the needs of the market, while
supporting both the strategic and "nancial plans of
the "rm.
   One of the most interesting features of S&OP is
its part strategic and part tactical nature [4].
S&OP is on the one hand constrained by the capa-
city strategy, but on the other also in#uences the                     Fig. 4. Tactical decisions in the S&OP process.
220                         J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225

3.2. Modify demand

   Modifying demand will not be discussed in any
detail in this paper. Nevertheless, using a set of
marketing tools to in#uence the sales pattern can
be seen as a way to achieve a more levelled produc-
tion. Such tools can for instance be campaigns, new
product introduction, complementary products,
                                                                Fig. 5. The production plan is levelling the sales plan (in the
pricing, etc.                                                   MTO case, an initial backlog is assumed).

3.3. Modify supply
                                                                this sense is the inverse of the inventory curve.
   The supply option can be divided into three                  However, #uctuations in sales can also be managed
categories } level, chase and combination/mix                   by in- and outsourcing, subcontracting, etc [8],
planning strategies. The two extremes (level and                without changing the internal production rates.
chase) are seldom used in isolation. The process                   When the sales plan is chased, output rates are
industry, however, is normally using a pure level               adjusted to match sales each period during the
strategy, and a pure chase strategy can for instance            planning horizon. Buxey [6] de"nes chase as entail-
be found in engineer-to-order environments. When                ing exact synchronisation between (forecasted) de-
discussing planning strategies, it is of great import-          mand and supply, while Pan and Kleiner [9] de"ne
ance to distinguish between di!erent market envi-               it as matching forecasted demand period by period.
ronments: engineer-to-order (ETO), make-to-order                The aim is to minimise investment in inventories
(MTO), assemble-to-order (ATO) (sometimes refer-                and/or the order backlog by producing to the ac-
red to as "nish-to-order, FTO) and make-to-stock                tual sales plan or customer orders. In the ideal case,
(MTS) all call for di!erent approaches when the                 with very short lead-times, the MTS and the MTO
supply is modi"ed. In the following, we let the                 environments collapse into a &make-to-require-
MTO case to also cover for the ETO case.                        ments' (MTR) environment without any inventory
   When production is levelled, constant output                 or order backlog. In the normal case inventory
rates are maintained during the planning horizon.               and/or backlog do exist, but are kept on a stable
Buxey [6] de"nes level as e!ectively decoupling                 level. Flexibility and adaptability are achieved at
supply from demand. A similar de"nition is also                 the expense of low utilisation of production re-
found in Johansen and Riis [7], where they state                sources and at higher costs associated with changes
that production is decoupled from forecasts and                 in output rates. The literature suggests numerous
actual market demand for a period of several                    ways of adopting the chase option; changing em-
months. The aim is to achieve a uniform and high                ployment levels (hire/"re), using overtime or idle-
utilisation of production resources, including                  time (excess capacity), #exible working hours, etc.
a minimisation of costs related to changes in pro-                 As mentioned above a combination of the two
duction rates. In an MTS environment, the tactics               extremes will commonly be observed in practice. In
can roughly be described as maintaining a steady                this combination/mix strategy, output rates are
production (output) rate throughout the planning                changed during the planning horizon, but not as
horizon, absorbing #uctuations in the sales plan by             frequent as when the sales plan is chased. Produc-
changes in inventory level. In MTO environments                 tion can for instance be levelled over a couple of
#uctuations are managed by changes in the order                 periods, followed by a chase strategy, then altered
backlog, i.e. increasing or decreasing the backlog,             back to a level strategy, and so forth (see Fig. 6,
which of course a!ects the delivery lead times and              showing an example of the combination/mix strat-
reliability (see Fig. 5 for an example of a level               egy). The combination/mix option requires some
strategy). Note that the backlog can be seen as                 anticipation inventory build-up during slack peri-
&negative inventory', and that the backlog curve in             ods (in an MTS environment), or increases in the
                              J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225                   221

                                                                  right option, but also suggests a couple of support-
                                                                  ive methods that can be used in connection with
                                                                  these. A more hands-on tool is the S&OP reports
                                                                  (see e.g. [3]), where past and current performance
                                                                  as well as future plans are easily visualised. How-
                                                                  ever, the cost trade-o! mentioned above has been
                                                                  an object of research for several decades, wherefore
                                                                  the largest portions of articles in the S&OP/APP
Fig. 6. The combination/mix planning strategy (in the MTO
case, an initial backlog is assumed).                             area focus on mathematical models trying to opti-
                                                                  mise the choice of planning strategy. The basic
                                                                  mathematical models can be described as follows:
order backlog during peak periods (in MTO envi-                   given a set of forecasts (F ), determine production,
ronments), but only a few changes of output rates                 inventory/backlog, and workforce levels (P , I , and
                                                                                                                R R
are necessary. This planning strategy thereby en-                 = , respectively), t"1, 2, 2, N, which minimises
ables a better correspondence between sales and                   cost subject to appropriate constraints. N is the
production output than does a level strategy, more-               time horizon, normally reaching over 6}18 months
over it creates a more stable production environ-                 [9]. In choosing P , I , and = relative to F , the
                                                                                       R R          R              R
ment than does a chase strategy. Decisions                        decision-makers choose the desired trade-o! be-
regarding how often production rates shall be alter-              tween the important cost parameters. Nam and
ed during the planning horizon, is based on the                   Logendran [10] made a comprehensive survey of
trade-o! between costs associated with maintain-                  APP models containing 140 journal articles (from
ing a steady output rate and costs associated with                17 journals) and 14 textbooks, ranging from the
changes in output rates.                                          early 1950s to 1990. However, all mathematical
   As in the case of capacity strategies, the planning            models for aggregate planning depend on the estima-
strategies can be described as di!erent positions on              tion of a number of relatively uncertain cost para-
a spectrum, with level and chase as the endpoints.                meters, and despite the numerous approaches and
Level would correspond to a situation where uni-                  techniques covered in the literature, only a few have
form and high utilisation is the norm, while chase                been implemented in an industrial situation [10].
would correspond to a situation where (volume) #ex-
ibility is the norm. Along the spectrum di!erent types
of combination/mix planning strategies are to be                  4. Interaction of the two perspectives
found, where the choice of strategy is based on the
trade-o! between the endpoints. After a short descrip-               As mentioned in the introduction the capacity
tion of models and methodologies used in S&OP, we                 strategy mainly focuses on the timing of capacity
then turn to discuss how the two perspectives (capa-              changes, whereas the S&OP's primary focus is on
city strategy and S&OP) interact with each other.                 the rate of production. Both perspectives do deal
                                                                  with long-term capacity management, but since
3.4. Models and methodologies                                     they belong to di!erent decision categories they are
                                                                  traditionally treated in sequence, where the S&OP
   Johansen and Riis [7] provide a framework for                  has to work within the frames set by the capacity
choosing appropriate demand management de-                        strategy. Yet, there is a strong interrelationship
pending on the strategy focus. In their framework                 between them, which is discussed below.
a number of company speci"c factors (like forecast
accuracy, intensity of season, process complexity,                4.1. Manufacturing strategy's impact on sales and
etc.) and political and social factors (like interest                  operations planning
rates, rate of unemployment, etc.) are used to guide
managers in the S&OP process. The framework                         The decision on when and how to acquire new
aims at helping managers not only to choose the                   capacity strongly in#uences the action space for
222                            J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225

             Fig. 7. Typical capacity strategies and planning strategies for di!erent product/process environments.

S&OP. This strategic manufacturing decision de-                    overtime at some instances and have over-capacity
termines whether capacity is adjusted before                       at others, typically for MTO situations, or build
a change in demand, or only once a demand change                   inventory in MTS environments.
has been experienced. Consequently, the planning
strategy of S&OP will be more or less forced to                    4.2. Sales and operations planning's impact on
work within the capacity levels set by the capacity                     manufacturing strategy
expansion or reduction strategy. Still, should a pro-
duction plan be chosen that will need extra capa-                     The decision on how to produce also in#uences
city, then that decision will implicitly imply that                the need for new capacity. Production smoothing
this extra capacity is of temporary nature, and not                will keep the maximum capacity requirements to
a new capacity acquisition.                                        a minimum, whereas more capacity is needed in
   If a lead strategy is followed, then S&OP is left               one form or another if production volumes vary
with much freedom. However, using a lag strategy                   a lot between periods. Also, in capacity reduction
will force the S&OP to act within much tighter                     situations a level production rate will facilitate the
capacity levels, restricting the use of a chase plann-             decision on capacity levels.
ing strategy. If a chase strategy is desirable, then                  If products can be produced in a long-term stable
there is a need for extra capacity in terms of sub-                production rate and there is an upward demand
contracting, short-term overtime, or capacity #ex-                 trend, then new capacity can be postponed. The
ibility. A track strategy usually means that capacity              option to use a level production rate is usually quite
can be acquired in smaller steps, allowing for                     good in an MTS environment. On the other hand,
a closer "t between capacity and demand. Such                      if products are made or engineered to order and
a capacity expansion strategy would still imply that               a chase planning strategy is pursued, then new
the planning strategy will use sub-contracting or                  capacity will be needed earlier as demand increases
                            J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225                    223

relative to an MTS situation. A mix or combination              work, it is necessary to treat the lead strategy as
strategy would lead to a situation between these                corresponding to a capacity supply surplus situ-
two other (extreme) planning strategies.                        ation, and the lag strategy must correspond to
   Even in MTO situations, new capacity invest-                 capacity demand surplus. Hereby, the capacity
ments can be delayed if production can be levelled              strategies support the same order winning charac-
rather than be chasing demand. The reason is that               teristics both for expanding and declining demand
the maximum capacity requirements are reduced,                  patterns.
thereby reducing the need for extra capacity until the             If capacity resources are acquired in large dis-
average capacity requirements indicate such a need.             crete steps, then it is important to distinguish be-
In one case, where one of the authors was involved,             tween capacity acquisition (or reduction) and
a company would have been able to delay a new                   capacity control. The former is dealt with from
investment by approximately one year, should they               a manufacturing strategy perspective, whereas the
have adopted a production smoothing option.                     latter is treated in the sales and operations plan.
                                                                However, these decisions need to be taken in an
4.3. A framework for combining the two perspectives             integrated fashion, and not in sequence. The reason
                                                                is that there is interdependency between these deci-
   The product/process matrix [1] can be used to                sions, such that both in#uence the decision space of
illustrate the correspondence between capacity                  the other.
strategies and planning strategies (see Fig. 7). Typi-             The main results from linking the perspectives of
cally, manufacturing "rms are found along the di-               capacity expansion strategies and planning strat-
agonal from the upper left-hand corner to the lower             egies are summarised in Fig. 8.
right-hand corner that is roughly represented by                   The upper left-hand corner and the lower right-
positions 1}4. Low-volume manufacturing of non-                 hand corner both indicate situations where the two
standard, one-of-a-kind products is typically per-              perspectives are mutually supportive. Lead and
formed in a job shop. In such an ETO environment                chase focus on resource availability and #exibility
(see e.g. [11]), a typical order winner is #exibility           to provide order winners such as #exibility, design
(see e.g. [12]). Then, a lead strategy (according to            and quality. Lag and level combine their respective
our de"nition of capacity supply surplus) is prefer-            focus on resource utilisation, in order to support
able, to allow for volume #exibility. Also, due to the          competition on price. The lower left-hand corner
inherent nature of ETO not to manufacture to                    illustrates a situation that is neither supportive nor
a forecast, a chase strategy is preferable. At the              con#icting. Using a combination of lead and level
other end, high-volume, standard commodity items                means that production rate changes can be accom-
are typically manufactured in continuous pro-                   modated if needed, i.e. a possibility to move to-
cesses. Such items are typically made-to-stock (see             wards a mix or chase planning strategy. Also, using
e.g. [11]) and the predominant order winner is price            a level strategy means that the maximum capacity
(see e.g. [12]). Due to the focus on price, resource            requirements are kept to a minimum and thereby
utilisation is most important leading to a lag strat-           new capacity acquisitions can be postponed. The
egy (capacity demand surplus), in order to avoid                real problem is the upper right-hand corner where
over-capacity. Also, the MTS situation and the                  the strategies are con#icting. Using capacity de-
importance of cost-e$ciency lead to a level produc-             mand surplus means that there are very limited
tion rate. The main reason is again to avoid unnec-             opportunities to execute a chase strategy. Instead,
essary temporary over-capacity. In this paper, we               the company must rely on other, external sources
add capacity strategy and planning strategy to the              for additional capacity when needed to support
list of characteristics related to the integrated               a chase plan. Still, the company is likely to face
choice of product/process environment.                          frequent overloads and subsequently major deliv-
   As can be seen in Fig. 7, there is a correspond-             ery problems. The positions for track and mix/com-
ence between, on the one hand, lead and chase, and              bination can be viewed as intermediate relative to
on the other, between lag and level. In this frame-             the other extreme compositions.
224                          J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225

                      Fig. 8. Illustration of the e!ects of combining capacity and planning strategies.

5. Summary and extensions                                         strategy is synonymous to capacity supply surplus,
                                                                  whereas a lag strategy assumes capacity demand
   In this paper we have linked a structural and an               surplus.
infrastructural decision category from the manufac-                  An extension to this research is to relate the
turing strategy framework. The structural perspect-               long-term capacity management issues to the prod-
ive deals with capacity in terms of capacity levels               uct life cycle, to investigate the sensitivity relative to
and expansion/reduction strategies. The infrastruc-               the various life cycle stages, e.g. taking market
tural perspective deals with S&OP in terms of                     growth rate into account. Furthermore, the design
planning strategies for production relative to sales,             of decision support systems for long-term capacity
inventory and/or backlogs. Together they provide                  management, i.e. how to combine and integrate
two perspectives on long-term capacity man-                       these issues, needs further attention.
agement. The decisions related to capacity ac-
quisition/reduction and planning/control are inter-
related in such a way that they must be integrated                Acknowledgements
and cannot be dealt with in sequence. We have
plotted these two characteristics relative to the                   The research is supported by grants from the
product/process matrix [1] and found a strong                     Swedish Foundation for Strategic Research, Eric-
relationship. Typically, lead and chase are related               sson Radio Systems AB and Volvo Research Foun-
because they both put a strong focus on resource                  dation, Volvo Educational Foundation and Dr
availability and #exibility. Moreover, lag and level              Pehr G Gyllenhammar Research Foundation.
both strongly focus on resource utilisation. When
discussing capacity reduction strategies relative
to the lead, track and lag strategies suggested by
Hayes and Wheelwright [1] for capacity expansion
strategies, we suggest that these three strategies,                [1] R.H. Hayes, S.C. Wheelwright, Restoring our Competitive
related to the capacity level relative sales, should                   Edge } Competing Through Manufacturing, Wiley, New
be interpreted in the following manner. A lead                         York, 1984.
                                J. Olhager et al. / Int. J. Production Economics 69 (2001) 215}225                              225

[2] G.K. Leong, D.L. Snyder, P.T. Ward, Research in the                    enterprises, Production Planning and Control 6 (5) (1995)
    process and content of manufacturing strategy, Omega                   461}468.
    18 (2) (1990) 109}122.                                           [8]   M.I. Kamien, L. Li, Subcontracting, co-ordination, #exib-
[3] R.C. Ling, W.E. Goddard, Orchestrating Success } Im-                   ility, and production smoothing in aggregate planning,
    prove Control of the Business with Sales and Operations                Management Science 36 (11) (1990) 1352}1363.
    Planning, Wiley, New York, 1988.                                 [9]   L. Pan, B.H. Kleiner, Aggregate planning today, Work
[4] R.A. Stahl, Sales and operations planning (S&OP) } a fun-              Study 44 (3) (1995) 4}7.
    damental that still works, Proceedings of APICS 38th            [10]   S.J. Nam, R. Logendran, Aggregate production plann-
    International Conference, Orlando, 1995, pp. 215}218.                  ing } a survey of models and methodologies, European
[5] L.J. Krajewski, L.P. Ritzman, Operations Management                    Journal of Operational Research 61 (3) (1992)
    } Strategy and Analysis, 4th Edition, Addison-Wesley,                  255}272.
    Reading, MA, 1996.                                              [11]   H.W. Oden, G.A. Langenwalter, R.A. Lucier, Handbook
[6] G. Buxey, The myth of aggregate planning, Production                   of Material and Capacity Requirements Planning,
    Planning and Control 1 (4) (1990) 222}234.                             McGraw-Hill, New York, 1993.
[7] J. Johansen, J.O. Riis, Managing seasonal #uctuations           [12]   T. Hill, Manufacturing Strategy } Text and Cases, Mac-
    in demand: Practice and experience of selected industrial              Millan, London, 1995.

To top