Continuous improvement can certainly he a subsetof mass customization. The autonomous operatingunits within a mass customizer can andshould strive to continuously improve their processes.
But as Toyota, for one, seems to bave finallyrealized, mass customization generally cannot bea subset of continuous improvement.
One of tbe main causes of Toyota's recent prohlemswas that it had been pursuing mass customization
but had retained tbe structures and systemsof continuous-improvement organizations.
By doing this, Toyota ended up not succeeding atmass customization and, at tbe same time, underminingits continuous-improvement efforts.
For example, Toyota assumed that its work forcehad attained the skills needed to handle productionof its rapidly growing range of product offerings. But
when the frequently cbanging tasksbutted up against tbe limit oi workers'capabilities, managers did notrealize tbat the problems stemmedfrom a failure to transform the organization.
Ratber than developing tbeloose network necessary to make amass-customization organizationwork, Toyota managers turned tomachines. Over time, this ended up weakening the
skills of the workers and thus violated an essentialtenet of continuous improvement. It also caused internalfriction.
One action Toyota took was to invest heavily in
robots. But as one Toyota manager later commented,
"Robots don't make suggestions." Toyota also
installed monitors at some stations along the assemblyline tbat told workers bow to put togethera particular car. And the company installed computer-#p#分頁標題#e#
controlled spotlights illuminating the binscontaining tbe rigbt components. These measures
deprived employees of opportunities to learn andthink about the processes and, tberefore, reducedtheir ability to improve them.
Another big problem at Toyota was tbat productproliferation took on a life of its own. Like mindlesscontinuous improvers, engineers created teebnically
elegant features regardless of whether customerswanted the additional choices. In mass customization,
customer demand drives model varieties.
A third problem arose wben Toyota's management,in its pursuit of low-cost customization,
pushed product development teams to use morecommon components across its models. At Toyota,project leaders have overall responsibility for the
development of a given model, but separate teamsdevelop individual components, such as brake systemsor transmissions, which ideally will be usedin several models. Project leaders felt that the intensifyingpressure to share components was forcingthem to compromise their models, and they
began to resist. Eventually, the company couldn'tachieve targeted levels for sharing design expertise,components, and production processes, and overallproduct development costs rose.
Other companies bave also heen attempting toachieve mass customization with less than optimalresults. Some of their experiences highlight the potentialpitfalls companies can encounter in tryingto make this leap.
• Nissan, Mitsubishi, and Mazda have run intomany of the same problems that hurt Toyota. Nissan,for example, reportedly bad 87 different varietiesof steering wheels, most of which were greatengineering feats. But customers did not want
Toyota created technicallyelegant features regardless ofwhether customers wanted them.many of tbcm and disliked having to choose fromso many options.
D Amdahl, which built its business on a low-cost
strategy hut never made the move to continuous
improvement, adopted a goal similar to Toyota's:
deliver a custom-built mainframe in a week. However,
Amdahl did not achieve its objective througb
flexible process capabilities, a dynamic network, or
anything else resembling mass customization. It
stocked inventory for every possible combination
110 HARVARD BUSINESS REVIEW September-October 1993
that customers could order, an approach that ended
up saddling it with hundreds of millions of dollars
in excess inventory.
• Dow Jones, tbrougb the Wall Street fournal and
its other news-gathering resources, has a storehouse
of information that it can customize and
then deliver in a number of ways, including newswires,
faxes, and on-line computer systems. Dow
Nissan, Mitsubishi, and Mazda
ran into many of the same
problems that hurt Toyota.
Jones, however, has not yet found the right formula
for packaging services at a low price that
would allow it to increase its share of the market.#p#分頁標題#e#
We suspect that two factors are responsible. Dow
Jones seems to he trying to push a somewhat customized
product out the door ratber than first determining
what customers truly need and how they
want it delivered. The company also doesn't appear
to have developed the organizational capabilities
that would enable it to lower its costs enough to
expand the still emerging market for customized
information.
Despite the fact so many companies are struggling,
scores of otbers are joining the quest. Tbe appeal
is understandable. Mass customization offers
a solution to a basic dilemma that has plagued
generations of executives.
Breaking New Ground
Until the widespread adoption of continuous improvement
began about 15 years ago, either/or dichotomies
dictated most managerial choices. A
company could pursue a strategy of providing large
volumes of standardized goods or services at a low
cost, or it could decide to make customized or highly
differentiated products in smaller volumes at a
high cost. In other words, companies bad to choose
between being efficient mass producers and being
innovative specialty businesses. Quality and low
cost and customization and low cost were assumed
to he trade-offs.
Tbis old competitive dictum was grounded in tbe
seemingly well-substantiated notion that the two
strategies required very different ways of managing,
and, therefore, two distinct organizational forms.
The mechanistic organization, so named because
of tbe management emphasis on automating tasks
and treating workers like machines, consists of
a bureaucratic structure of functionally defined,
highly compartmentalized jobs. Managers and industrial
engineers study and define tasks, and
workers execute them. Employees learn their jobs
by following rigid rules under tight supervision.
In contrast, tbe organic organization, so named
because of its fluid and ever-changing nature, is
characterized by an adaptable structure
of loosely defined jobs. These
are typically held by highly skilled
craftsmen. They learn through apprenticeships
and experience, are
governed by personal or professional
standards, and arc motivated by a
desire to create a unique or breakthrough
product.
The mechanistic organization, whether in a manufacturing
or a service setting, gives managers the
control and predictability required to achieve high
levels of efficiency. The organic organization yields
the craftsmanship needed to pursue a differentiation
or niche strategy. Each of tbese organizational
forms bas innate limitations, however, which in
tbe past have forced managers to choose one or the
other. Almost all change is anathema to the mechanistic
organization. And the artistry and informality#p#分頁標題#e#
at the heart of the organic organization defy efforts
to regulate and control.
The development of tbe continuous-improvement
and tbe mass-customization models sbow
that companies can overcome the traditional tradeoffs.
In other words, companies can have it all.
Continuous improvement has enahled thousands
of companies to realize lower costs than traditional
mass producers and still achieve the distinctive
quality of craft producers. But mass customization
has enabled its adherents, which are as varied as
Motorola, Bell Atlantic, the diversified insurer
United Services Automobile Association (USAA),
TWA Getaway Vacations, and Hallmark, to go a
step further. These companies are achieving low
costs, high quality, and the ahility to make highly
varied, often individually customized products.
Is Your Company Ready for Mass
Customization?
Since achieving mass customization requires
nothing less than a transformation of the husiness,
managers must assess whether their companies
must and wbetber in fact they can make
tbe transformation.
Not all markets are appropriate for mass customization.
Customers of commodity products
HARVARD BUSINESS REVIEW September October 1993 111
MASS CUSTOMIZATION
like oil, gas, and wheat, for example, do not demand
differentiation. In other markets, like public utilities
and government services, regulation often bars
customization. In some markets, the possible variations
in services or products simply are of little
value to customers. Also, variety in and of itself is
not necessarily customization, and it can be dangerously
expensive. Some consumer
electronics retailers and supermarkets
today are experiencing a backlash
from customers confused by too
hroad a range of choiees.
Continuous improvement will
continue to be a very viable strategy
for companies whose markets are
relatively stable and predictable. But
tbose companies whose markets are highly turbulent
hecause of factors like changing customer
needs, technological advances, and diminishing
product life cycles are ripe for mass customization.
To have even a chance of successfully becoming
a mass customizer, though, eompanies must first
achieve high levels of quality and skills and low
cost. For this reason, it seems impossible for mass
producers to make the leap without first going
througb continuous improvement.
Westpac, the Australian financial services giant,
is a ease in point. It spent huge sums attempting to
become a mass customizer hy automating both the
creation and delivery of its products. It wanted to
install software building blocks that would allow it
to create new financial products like mortgages and
securities more quickly. Strategically, the move#p#分頁標題#e#
made sense. Deregulation had spawned a dizzying
Companies must break apart
long-lasting, cross-functional
teams and relationships and
form dynamic networks.
array of new products and services, and intensifying
competition bad caused significant downward pressure
on prices.
Westpac tried to leapfrog continuous improvement
by going from mass production directly to
mass customization. Tbe challenges of automating
inflexible processes, building on ossified products,
and trying to create a fluid network within a hierarchical
organization-particularly at a time wben
the company was in poor financial condition due to
intensifying competition in depressed marketsproved
too difficult. Westpac has had to scale back
significantly its ambitious dreams of becoming a
tailored-product factory.
As we bave stressed, even a company that has
mastered continuous improvement must change
radically the way it is run to become a successful
Variety in and of itself is not
customization-and it can be
dangerously expensive.
mass customizer. A company must break apart the
long-lasting, cross-functional teams and strong relationships
built up for continuous improvement to
form dynamic networks. It must cbange the focus
of employee learning from incremental process improvement
to generating ever-increasing capabilities.
And leaders must replace a vision of "being the
best" in an industry with an ideology of satisfying
whatever customers want, when they want it.
Tbe traditional mechanistic organization, aimed
at achieving low-cost mass production, is segmented
into very narrow compartments, often called
functional or vertical silos, each of which performs
an isolated task. Information is passed up, and decisions
are handed down. Compensation of employees,
who are viewed as mere cogs in the wheel, is
generally based on standardized, narrowly defined
job levels or categories.
In continuous-improvement organizations,
tbe control system is
much more, although never completely,
horizontal. Inereasingly,
teams have not only responsibility
for but also authority over a problem
or task area. Such organizations are
moving to much more generalized
and overlapping job descriptions as
well as to team-based compensation.
Wben mass customization is the
objective, organizations structured around crossfunctional
teams can create horizontal silos just as
isolated and ultimately damaging to tbe long-term
health of the organization as vertical silos have
been. When Toyota expanded dramatically its variety,
for example, it found that tightly linked teams
did not share easily across their boundaries to improve
the general capabilities of the company. As a#p#分頁標題#e#
result, the costs of increasing variety rapidly outstripped
any value it was creating for customers.
114 HARVARD BUSINESS REVIEW September-October 1993
To achieve successful mass customization, managers
need first to turn their processes into modules.
Second, they need to create an architecture for
linking them that will permit them to integrate
rapidly in the best combination or sequence required
to tailor products or services. The coordination
of the overall dynamic network is often centralized,
while each module retains operational
authority for its particular process. Job descriptions
become increasingly hroad and may even disappear.
And compensation for eacb module, wbether it's a
team or an individual, is based on the uniqueness
and value of the contrihutions it makes toward producing
the product.
Making Mass Customization Work
The key to coordinating tbe process modules is a
linkage system with four key attributes.
1. Instantaneous. Processes must be able to be
linked together as quickly as possible. First, the
product or service each customer wants must be defined
rapidly, preferably in collaboration witb the
customer. Mass customizers like Dell Computer,
Hewlett-Packard, AT&T, and LSI Logic use special
software that records customer desires and translates
them into a design of the needed components.
Then the design is quickly translated into a set of
processes, which are integrated rapidly to create the
product or service.
2. Costless. Beyond the initial investment required
to create it, the linkage system must add as
little as possible to tbe cost of making the product
or service. Many service businesses bave databases
tbat make availahle all the information they know
about their customers and their requirements to all
tbe modules, so nothing new needs to be regenerated.
USAA, for example, uses image technology that
can scan and electronically store paperwork and a
companywide database, so every representative
who comes into contact with a customer knows everything
about him or her.
3. Seamless. An IBM executive
once commented, correctly, "You always
ship your organization." What
be meant was, if you have seams in
your organization, you are going to
have seams in your product, such
as programs that do not work well
together in a computer system. Since
a dynamic network is essentially constructing a
new, instant team to deal with every customer interaction,
the occasions for "showing the seams"
are many indeed. The recent adoption of case workers
or case managers is one way service companies
like USAA and IBM Credit Corporation avoid this.
These people are responsible for the company's relationship#p#分頁標題#e#
with the customer and for coordinating
the creation of tbe customized product or service.
They ensure that no seams appear.
4. Frictionless. Companies tbat are still predominantly
continuous improvers may have the most
trouble attaining this attribute. The need to create
instant teams for every customer in a dynamic
network leaves no time for the kind of extensive
team building that goes on in continuous improvement
organizations. The instant teams must be
frictionless from the moment of their creation, so
information and communications technologies are
mandatory for achieving this attribute. These technologies
are necessary to find the right people, to
define and create boundaries for tbeir collective
task, and to allow them to work togetber immediately
without the benefit of ever having met.
Using Technology. In mechanistic organizations,
the primary use of technology is to automate tasks,
replacing human labor witb mechanical or digital
machines. People are sources of variation and are
relatively costly, so mass producers often try to
automate tbeir companies as much as possible.
This has the natural effect of reducing the numbers
and skills of tbe work force.
In continuous-improvement eompanies, wbere
workers are not only allowed but also encouraged
to think about their jobs and how processes can he
improved, technology is primarily used to augment
workers' knowledge and skills. Measurement and
analysis programs, computerized decision-support
systems, videoconferencing, and even machine
tools are aids, not people replacements.
In the dynamic networks of mass customizers,
technology still automates tasks where that makes
sense. Certainly, technology must augment people's
knowledge and skills, but the elements of
mass customization require that technology must
also automate the links between modules and en-
Technology helps managers tap
the capabilities of employees
to service customers.
sure that tbe people and tbe tools necessary to
perform them are brought together instantly. Communication
networks, shared databases that let
everyone view the customer information simultaneously,
computer-integrated manufacturing.
HARVARD BUSINESS REVIEW Septemher October 1993 115
MASS CUSTOMIZATION
Understanding the Differences
Mass Production
The traditional
mass-productian company is
bureaucratic and hierarchical.
Under close supervision,
workers repeat narrowly
defined, repetitious tosks.
Result: low<ost, stondord
goods and services.
Continuous Improvement
In continuous-improvement
settings, empowered,
cross-functional teams strive
constantly to improve
processes. Managers are#p#分頁標題#e#
coaches, cheering on
communicotions and
unceasing efforts to improve.
Result: low-cost, high-quality,
standard goods and services.
workflow software, and tools like groupware (such
as Lotus Notes) can automate the links so that a
company can summon exactly the right resources
to service a customer's unique desires and needs.
Many managers still view tbe promises of advanced
technologies through tbe lens of mass production.
But for mass customizers, the promise of
technology is not the lights-out factory or the fully
automated back office. It is used as a tool to tap
more effectively all the diverse capabilities of employees
to service customers.
While automating the links between modules is
crucial, often some modules themselves can be automated
by adopting, for example, a flexible manufacturing
system that can choose instantly any
product component within its wide envelope of variety.
Motorola's Bravo pager factory in Boynton
Beach, Florida, for example, can produce pagersthanks
to hardware and software modularity - in lot
sizes as small as one within hours of an order arriving
from a customer. Tbe pager business is also a
good example of how a mass eustomizer can automate
links hetween modules. At Motorola, a sales
rep and a customer design together, on a rep's laptop
computer, the set of pagers (out of 29 million possihle
comhinations) that exactly meets that customer's
needs. Then tbe almost fully automated dynamic
network takes over. The rep plugs the laptop
into a phone and transmits one or more designs to
the factory. Within minutes, a bar code is created
with all the steps tbat a flexible manufacturing system
needs to produce the pager.
116 DRAWINGS BY GARISON WEILAND
Mass Customization
Mass customization colls for
flexibility and quick
responsiveness. In an
ever-changing environment,
people, processes, units, and
technology reconfigure to give
customers exactly what they
want. Managers coordinate
independent, capoble
individuals, and an efficient
linkage system is crucial.
Result: low-cost, high-quality,
customized goods and
services.
As wonderful as these technological miracles
sound, it is important to realize that technology is
also potentially harmful. Mass customizers must
periodically overhaul the linkages that tbey bave
adopted because as the market, the nature of tbeir
businesses, and tbe competition cbange, and as
technology advances, any linkage system inevitably
will become obsolete.
Anotber caveat: in this age when automated systems
are handling daily millions of customer orders
and inquiries placed via phones or computer systems,
mass customizers must constantly he on#p#分頁標題#e#
their guard against eliminating their opportunities
to learn what their customers like or dislike. Companies
must always make it possible for their customers
to "drop out" of the automated system so
tbey can talk to a real person who is committed to
helping tbem.
Learning from Failure. In the mechanistic organization,
learning how to do something better is the
prerogative of management and its collection of industrial
engineers and supervisors. Workers only
need to learn to do what is assigned to them; they
don't have to think about it as well. The breakthrough
of continuous improvement was tbe acknowledgment
that workers' experience and knowhow
can help managers solve production prohlems
and contribute toward tightening variances and reducing
errors.
The differences hetween organizational learning
in continuous-improvement and in mass-customization
companies are most visible when you see
I lARVARD BUSINESS REVIEW September-October 1993 117
MASS CUSTOMIZATION
bow tbe two treat defects, Continuous-improvement
organizations look at them as process failures,
which tbe Japanese consider "treasures" because
they provide the knowledge to fix problems and
to ensure that failure never recurs.
In the dynamic networks of mass-eustomization
organizations, defects are considered capability
ers witb the ability to capture valuable new knowledge.
Motorola's and USAA's systems are good examples
of this, as is Bally Engineered Structures
Inc.'s (see tbe insert, "Overcoming the Hurdles at
Bally"). This is very different from what goes on in
both mass-production and continuous-improvement
organizations. Typically, in those settings,
there are almost no individual cus-
•|- . » ^ I tomer interactions that generate
It IS heresy to try to know what to new knowledge.
produce in the future. Customer
orders will determine that.
failures: the inability to satisfy the needs of some
specific customer or market. Tbey are still valuable
treasures; but rather than sparking a spate of process-
improvement activities, these defects call on
the organization to renew itself hy enhancing the
flexihility within its processes, joining with anotber
organization tbat has the needed capability, or
even creating completely new process capabilities -
whatever it takes to ensure tbat tbe customer is satisfied
and, tberefore, tbat capability failure doesn't
happen again.
Capturing customer feetlhack on capability failures
is crucial to sustaining any advantage that
mass customization yields. A company that does
this well is USAA, which targets its financial services
and consumer goods to events in a customer's
life, such as huying a house or car, getting married,#p#分頁標題#e#
or having a baby. Its information system allows
sales reps to get customer feedback
quickly on the phone and route it instantly
to the appropriate department
for analysis and action.
At Computer Products, inc., a
manufacturer of power supplies,
marketing managers and engineers
cold-call customers every day not to
make a sale hut to understand their
prohlems and needs and to discuss
product ideas. They then enter tbe
information into a database that serves as an invaluable
reference throughout the product-development
cycle. Applied Digital Data Systems, a unit
of AT&.T's NCR suhsidiary, uses a database system
to store all its production information, including
workers' comments and suggestions, and tben
regularly analyzes it to improve hoth its processes
and products.
The capahility to codesign and even coproduce
products with customers provides mass customiz-
Creating a Vision. In addition to
different attitudes ahout customer
interactions, leaders of continuousimprovement
companies and mass
customizers foster very different approaches
to the future. The former
think they know wbat the organization needs to do
to succeed in the future, whereas the latter helieve
that it's impossible to know and heresy to try hecause
the future should he shaped hy each successive
customer order.
Leaders of continuous-improvement organizations
provide a vision of not just what is to he done
today hut also what needs to he realized tomorrow,
and this can work, provided that the market is relatively
stahle. Their vision is often expressed in
terms of some competitive ideal of customer satisfaction.
Allstate's "To he the hest," Federal Express's
"Never let tbe best get in the way of better,"
and Steelcase's "To provide the world's hest office
environment products, services, systems, and intelligence"
are good examples. Tbe common vision
provides everyone in the company with the motivation,
direction, and control necessary to continue
USAA sales reps get customer
feedback quickly on the phone
and route it instantly to
appropriate departments.
improving all the time. Without a sustained vision,
a company's attempts at process improvement can
hecome lost in "program-of-the-month" fads or lip
service to quality.
The highly turhulent marketplace of the mass
customizer, with its ever-cbanging demand for innovation
and tailored products and services,
doesn't result in a clear, shared vision of tbat
market. A standard product or market vision isn't
just insufficient; it simply doesn't make sense. In
118 HARVARD BUSINESS KEVIEW September-October 1993
a true mass-customization environment, no one#p#分頁標題#e#
knows exactly what the next customer will want,
and, therefore, no one knows exactly what product
the company will he creating next. No one knows
what market-opportunity windows will open, and,
therefore, no one can create a long-term vision of
certain produets to service those markets. But everyone
does know that the next customer will want
something and the next market opportunity is out
tbere somewhere.
Many companies are articulating this scenario by
using words like "anything," "anywhere," and"anytime." Peter Kann, chief executive of Dow
Jones, describes his organization's strategic goal asproviding "husiness and financial news and information
however, wherever, and whenever customers
want to receive it." Nissan's vision for theyear 2000 is the "Five A's": any volume, anytime,anyhody, anywhere, and anything. Motorola's
pager group has a TV ad that asks, "How do you useyour Motorola pager?" Various people answer withphrases like, "Anytime," "For anything," and"Anywhere I want."
No matter what they are called, such ideologiessay two things ahout an organization: one, we don'tknow exactly what we'll have to provide to whom,and two, within our growing envelope of capabilities,
we do know that we have or can acquire the capahilitiesto give customers what they want.
留學生dissertation網Leaders who can articulate such an ideology and
create the dynamic network that can make it bappenwill succeed in moving tbeir organizations far
beyond continuous improvement to the new competitivearena of mass customization.
Authors' note: IBM Consulting Group partners and consultantscontributed significantly to the development ofthe ideas in thisarticle. The research was sponsored in part by the IBM Consulting
Group and the IBM Advanced Business Institute.
Reprint 93509

CARTOON BY NICK DOWNES 119