Auburn University at Montgomery Enterprise System Discussion

If you’re not careful, the dream of information
integration can turn into a nightmare.
Putting the Enterprise into
the Enterprise System
by Thomas H. Davenport
E
NTERPRISE SYSTEMS appear to be a dream come true. These
commercial software packages promise the seamless integration
of all the information flowing through a company – flnancial and
accounting information, human resource information, supply chain
information, customer information. For managers who have struggled,
at great expense and with great frustration, with incompatible information systems and inconsistent operating practices, the promise of an
off-the-shelf solution to the problem of business integration is enticing.
It comes as no surprise, then, that companies have been beating paths
to the doors of enterprise-system developers. The sales of the largest
ARTWORK BY CURTÍS PARKER
121
ENTERPRISE SYSTEMS
vendor, Germany’s SAP, have soared from less than
$500 million in 1992 to approximately $3.3 billion
in 1997, making it the fastest-growing software
company in the world. SAP’s competitors, including such companies as Baan, Oracle, and PeopleSoft, have also seen rapid growth in demand for
their packages. It is estimated that businesses
around the world are now spending $10 billion per
year on enterprise systems-also commonly referred to as enterprise resouree planning, or ERP,
systems-and that figure probably doubles when
you add in associated consulting expenditures.
While the rise of the Internet has received most of
the media attention in recent years, the business
world’s embraee of enterprise systems may in fact
be the most important development in the eorporate use of information technology in the 1990s.
But are enterprise systems living up to companies’ expectations? The growing numher of horror
stories about failed or out-of-eontrol projects
should certainly give managers pause. FoxMeyer
Drug argues that its system helped drive it into
bankruptcy. Mobil Europe spent hundreds of millions of dollars on its system only to abandon it
when its merger partner objected. Dell Computer
found that its system would notfitits new, decentralized management model. Applied Materials
gave up on its system when it found itself overwhelmed by the organizational changes involved.
Dow Chemical spent seven years and close to half
a billion dollars implementing a mainframe-based
enterprise system; now it has decided to start over
again on a client-server version.
Some of the blame for such debacles lies with the
enormous technical challenges of rolling out enterprise systems – these systems are profoundly complex pieces of software, and installing them requires
large investments of money, time, and expertise.
But the technical challenges, however great, are
not the main reason enterprise systems fail. The
biggest problems are business problems. Companies fail to reconcile the technological imperatives
of the enterprise system with the business needs of
the enterprise itself.
An enterprise system, by its very nature, imposes
its own logic on a company’s strategy, organization,
and culture. (See the table “The Scope of an Enterprise System.”) It pushes a company toward full integration even when a certain degree of businessThomas H. Davenport is a professor at the Boston University School of Management in Boston, Massachusetts.
His most recent book, Working Knowledge: How Organizations Manage What They Know, was published in
by the Harvard Business School Press.
122
THE SCOPE OF AN
ENTERPRISE SYSTEM
An enterprise system enables a company to integrate the data used throughout its entire organization. This list shows some of the many functions
supported by SAP’s R/3 package.
Financiáis
Accounts receivable and payable
Asset accounting
Cash management and forecasting
Cost-element and cost-center accounting
Executive information system
Financial consolidation
General ledger
Product-cost accounting
Profitability analysis
Profit-center accounting
Standard and period-related costing
Human Resources
Human-resources time accounting
Payroll
Personnel planning
Travel expenses
Operations and Logistics
Inventory management
Material requirements planning
Materials management
Plant maintenance
Production planning
Project management
Purchasing
Quality management
Routing management
Shipping
Vendor evaluation
Sales and Marketing
Order management
Pricing
Sales management
Sales planning
HARVARD BUSINESS REVIEW
July-August 199Ö
ENTERPRISE SYSTEMS
unit segregation may be in its best interests. And it
pusbes a company toward generic processes even
when customized processes may be a source of competitive advantage. If a company rusbes to install
an enterprise system without first baving a clear
understanding of tbe business implications, tbe
dream of integration can quickly turn into a nightmare. Tbe logic of tbe system may conflict witb the
logic of the business, and eitber tbe implementation will fail, wasting vast sums of money and causing a great deal of disruption, or tbe system will
weaken important sources of competitive advantage, hobbling tbe company.
It is certainly true that enterprise systems can deliver great rewards, but the risks they carry are
equally great. Wben considering and implementing
an enterprise system, managers need to be careful
tbat their entbusiasm about tbe benefits does not
blind tbem to tbe bazards.
The Allure of Enterprise Systems
In order to understand tbe attraction of enterprise
systems, as well as their potential dangers, you first
need to understand tbe problem they’re designed to
solve: the fragmentation of information in large
business organizations. Every big company collects, generates, and stores vast quantities of data.
In most companies, tbougb, the data are not kept in
a single repository. Rather, tbe information is
spread across dozens or even bundreds of separate
computer systems, eacb boused in an individual
function, business unit, region, factory, or office.
Eacb of tbese so-called legacy systems may provide
invaluable support for a particular business activity. But in combination, they represent one of the beaviest drags on business productivity and performance
now in existence.
Maintaining many different computer systems leads to enormous
costs-for storing and rationalizing
redundant data, for rekeying and reformatting data from one system for
use in another, for updating and debugging obsolete software code, for
programming communication links between systems to automate tbe transfer of data. But even more
important than tbe direct costs are tbe indirect
ones. If a company’s sales and ordering systems cannot talk witb its production-scheduling systems,
tben its manufacturing productivity and customer
responsiveness suffer. If its sales and marketing systems are incompatible witb its financial-reporting
systems, tben management is left to make impor-
tant decisions by instinct ratber than according to
a detailed understanding of product and customer
profitability. To put it bluntly: if a company’s systems are fragmented, its business is fragmented.
Enter tbe enterprise system. A good ES is a tecbnotogical tour de force. At its core is a single comprebcnsive database. Tbe database collects data
from and feeds data into modular applications supporting virtually all of a company’s business activities-across functions, across business units, across
tbe world. (See tbe cbart “Anatomy of an Enterprise
System.”) Wben new information is entered in one
place, related information is automatically updated.
Let’s say, for example, tbat a Paris-based sales
representative for a U.S. computer manufacturer
prepares a quote for a customer using an ES. Tbe
salesperson enters some basic information about
tbe customer’s requirements into his laptop computer, and tbe ES automatically produces a formal
contract, in Frencb, specifying tbe product’s configuration, price, and delivery date. When the customer accepts tbe quote, the sales rep hits a key; tbe
system, after verifying tbe customer’s credit limit,
records tbe order. Tbe system scbedules tbe sbipment; identifies tbe best routing; and tben, working
backward from tbe delivery date, reserves tbe necessary materials from inventory; orders needed
parts from suppliers; and scbedules assembly in tbe
company’s factory in Taiwan.
Tbe sales and production forecasts are immediately updated, and a material-requirements-planning list and bill of materials are created. Tbe sales
rep’s payroll account is credited witb tbe correct
commission, in French francs, and bis travel account
is credited with tbe expense of the sales call. Tbe
The growing number
of horror stories about failed
or out-of-control projects should
certainly give managers pause.
HARVARD BUSINESS REVIEW
|uly-AugU8t 1998
actual product cost and profitability are calculated,
in U.S. dollars, and tbe divisional and corporate balance sbeets, tbe accounts-payable and accountsreceivable ledgers, tbe cost-center accounts, and
tbe corporate casb levels are all automatically updated. The system performs nearly every information transaction resulting from the sale.
An ES streamlines a company’s dataflowsand
provides management witb direct access to a wealth
123
ENTERPRISE SYSTEMS
ANATOMY OF AN ENTERPRISE SYSTEM
At the heart of an enterprise system is a central database that draws data
from and feeds data into a series of applications supporting diverse company functions. Using a single database dramatically streamlines the flow
of information throughout a business.
Managersand
stakeholders
Reporting
applications
Sales and
delivery
applications
Customers
V
Sales force
and customer
service reps
Manufacturing
^applications
Back-ofñce
administrators
and workers
Suppliers
V
Human
resource
management
applications
Employees
of real-time operating information. For many companies, these benefits have translated into dramatic
gains in productivity and speed.
Autodesk, a leading maker of computer-aided
design software, used to take an average of two
weeks to deliver an order to a customer. Now, having installed an ES, it ships 98% of its orders within
24 hours. IBM’s Storage Systems division reduced
the time required to reprice all of its products from
5 days to 5 minutes, the time to ship a replacement
part from 22 days to 3 days, and the time to complete a credit check from 20 minutes to 3 seconds.
Fujitsu Microelectronics reduced the cycle time for
124
filling orders from 18 days to a day and a half and
cut the time required to close its financial books
from 8 days to 4 days.
When System and Strategy Clash
clearly, enterprise systems offer the potential of big
benefits. But the very quality of the systems that
makes those benefits possible-their almost universal applicahility-also presents a danger. When
developing information systems in the past, companies would first decide how they wanted to do
business and then choose a software package that
HARVARD BUSINESS REVIEW
July-August 1998
ENTERPRISE SYSTEMS
would support their proprietary processes. They
often rewrote large portions of the software code
to ensure a tight ñt. With enterprise systems, however, the sequence is reversed. The husiness often
must he modifled tofitthe system.
An enterprise system is, after all, a generic solution. Its design reflects a series of assumptions
about the way companies operate in general. Vendors try to structure the systems to reflect best
practices, but it is the vendor, not the customer,
that is defining what “best” means. In many cases,
the system will enable a company to operate more
efficiently than it did before. In some cases, though,
the system’s assumptions will run counter to a
company’s best interests.
Some degree of ES customization is possible. Because the systems are modular, for instance, companies can install only those modules that are most
appropriate to their business. However, the system’s
complexity makes major modifications impracticable. (See the insert “Configuring an Enterprise
System.”) As a result, most companies installing
enterprise systems will need to adapt or even completely rework their processes tofitthe requirements of the system. An executive of one company
that has adopted SAP’s system sums it up by saying,
“SAP isn’t a software package; it’s a way of doing
business.” The question is. Is it the hest way of doing husiness? Do the system’s technical imperatives
coincide or conflict with the company’s business
imperatives?
Imagine, for example, an industrial products
manufacturer that has built its strategy around its
ability to provide extraordinary customer service in
filling orders for spare parts. Because it is ahle to
consistently deliver parts to customers 25% faster
than its competitors-often by circumventing formal processes and systems-it has gained a large
and loyal clientele who are happy to pay a premium
price for its products. If, after installing an ES, the
company has to follow a more rational hut less flexible process for filling orders, its core source of
CONFIGURING AN ENTERPRISE SYSTEM
Configuring an enterprise system is largely a matter
of making compromises, of balancing the way you
want to work with the way the system lets you work.
You begin by deciding which modules to install.
Then, for each module, you adjust the system using
configuration tables to achieve the best possible fit
with your company’s processes. Let’s look more closely
at these two configuration mechanisms:
• Modules. Most enterprise systems are modular, enabling a company to implement the system for some
functions but not for others. Some modules, such as
those for finance and accounting, are adopted by almost all companies that install an ES, whereas others,
such as one for human resource management, are
adopted by only some companies. Sometimes a company simply doesn’t need a module. A service business, for example, is unlikely to require the module
for manufacturing. In other cases, companies choose
not to implement a module because they already have
a serviceable system for that particular function or
they have a proprietary system that they believe provides unique benefits. In general, the greater the number of modules selected, the greater the integration
benefits, but also the greater the costs, risks, and
changes involved.
• Configuration tables. A configuration table enables
a company to tailor a particular aspect of the system
to the way it chooses to do business. An organization
can select, for example, what kind of inventory
accounting-FIFO or LIFO-it will employ or whether
HARVARD BUSINESS REVIEW
|uly-August 1998
it wants to recognize product revenue by geographical
unit, product line, or distribution channel. SAP’s R/3,
one of the more comprehensive and complex ES offerings, has more than 3,000 configuration tables. Going
through all of them can take a long time. Dell Computer, for example, spent more than a year on the task.
Although modules and configuration tables let you
customize the system to some degree, your options
will be limited. If you have an idiosyncratic way of doing business, you will likely find that it is not supported by an ES. One company, for example, had long
had a practice of giving preferential treatment to its
most important customers by occasionally shipping
them products that had already been allocated to other
accounts. It found that its ES did not allow it the flexibility required to expedite orders in this way. Another
company had always kept track of revenues by both
product and geography, but its ES would allow it to
track revenue in only one way.
What happens when the options allowed by the system just aren’t good enough? A company has two
choices, neither of them ideal. It can actually rewrite
some of the ES’s code, or it can continue to use an existing system and build interfaces between it and the
ES. Both of these routes add time and cost to the implementation effort. Moreover, they can dilute the
ES’s integration benefits. The more customized an
enterprise system becomes, the less able it will be to
communicate seamlessly with the systems of suppliers and customers.
125
ENTERPRISE SYSTEMS
advantage may be at risk. The company may Integrate its data and improve its processes only to lose
its service edge and, in turn, its customers.
This danger becomes all the more pressing in
light of the increasing ubiquity of enterprise systems. It is now common for a single ES package to
be used by virtually every company in an industry.
For example, SAP’s R/3 package is being implemented by almost every company in the personal
It realized, in particular, that in a huild-to-order
environment an important advantage would accrue
to any company with superior capabilities for forecasting demand and processing orders. Compaq
therefore decided to invest in writing its own proprietary applications to support its forecasting and
order-management processes. To ensure that those
applications would be compatible with its ES,
Compaq wrote them in the computer language
used by its ES vendor.
Compaq’s course was not the obvious one. It cost the company considerably more to develop the proprietary
application modules than it would
have to use the modules offered by the
ES vendor. And using customized applications meant forgoing some of the
integration heneflts of a pure enterprise system. But Compaq saw the decision as a strategic necessity: it was
the only way to protect a potentially critical source
of advantage.
For companies that compete on cost rather than
on distinctive products or superior customer service, enterprise systems raise different strategic issues. The huge investment required to implement
an ES at large companies – typically ranging from
$50 million to more than $500 million-need to be
weighed carefully against the eventual savings the
system will produce. In some cases, companies
may ñnd that by forgoing an ES they can actually
gain a cost advantage over competitors that are embracing the systems. They may not have the most
elegant computer system or the most integrated information flows and processes, but if customers are
concerned only with price, that may not matter.
Air Products and Chemicals, for example, saw
that many of its competitors were installing large,
complex enterprise systems. After a thorough evaluation, it decided not to follow their lead. Its managers reasoned that the cost of an ES might force the
company to raise its prices, leading to lost sales in
some of the commodity gas markets in which it
competes. The company’s existing systems, while
not state-of-the-art, were adequate to meet its
needs. And since the company had no plans to exchange information electronically with competitors, it didn’t worry about being the odd man out in
its industry.
Of course, the long-term productivity and connectivity gains created by enterprise systems are
often so compelling that not adopting one is out of
the question. In the petrochemicals industry, for
example, enterprise systems have improved the
flow of information through the supply chain to
A speedy implementation of
an enterprise system may be a
wise business move, but a rash
implementation is not.
computer, semiconductor, petrochemical, and to a
slightly lesser degree, consumer goods industries.
[R/3 is the client-server version of SAP’s software;
R/2 is the mainframe version.) Such convergence
around a single software package should raise a
sobering question in the minds of chief executives:
How similar can our information flows and our
processes be to those of our competitors before we
begin to undermine our own sources of differentiation in the market?
This question will be moot if a company’s competitive advantage derives primarily from the distinctiveness of its products. Apple Computer, for
example, has many problems, but the loss of competitive differentiation because of its ES is not one
of them. With a strong brand and a unique operating
system, its computers still differ dramatically from
competing offerings. But Apple is an unusual case.
Among most makers of personal computers, differentiation is based more on service and price than on
product. For those companies, there is a very real
risk that an enterprise system could dissolve their
sources of advantage.
Compaq Computer is a good example of a company that carefully thought through the strategic
implications of implementing an enterprise system. Like many personal-computer companies,
Compaq had decided to shift from a build-to-stock
to a build-to-order business model. Because the success of a build-to-order model hinges on the speed
with which information flows through a company,
Compaq believed that a fully integrated enterprise
system was essential. At the same time, however,
Compaq saw the danger in adopting processes indistinguishable from those of its competitors.
126
HARVARD BUSINESS REVIEW
[uly-August 1998
ENTERPRISE SYSTEMS
such a degree that they have become a de facto operating standard. Because participants in the industry
routinely share information electronically, it
vv’ould today be hard for a company to survive in the
business without an ES. Still, the cost of implementation should be a primary concern. It v^^ill often be
in a company’s interest to go ahead and rework its
processes tofitthe system requirements. The alternative-customizing the system to fit the processes
or writing proprietary application modules – will
simply be too expensive to justify. As the CEO of
one large chemical firm says, “Competitive advantage in this industry might just come from doing
the best and cheapest job at implementing SAP.”
The Impact on an Organization
In addition to having important strategic implications, enterprise systems also have a direct, and
often paradoxical, impact on a company’s organization and culture. On the one hand, by providing
universal, real-time access to operating and financial data, the systems allow companies to streamline their management structures, creating flatter,
more flexible, and more democratic organizations.
On the other hand, they also involve the centralization of control over information and the standardization of processes, which are qualities more consistent with hierarchical, command-and-control
organizations with uniform cultures. In fact, it can
be argued that the reason enterprise systems first
emerged in Europe is that European companies
tend to have more rigid, centralized organizational
structures than their U.S. counterparts.
Some executives, particularly those in fast-growing high-tech companies, have used enterprise systems to inject more discipline into their organizations. They see the systems as a lever
for exerting more management control and imposing more-uniform processes on freewheeling, highly entre,
preneurial cultures. An executive at
July-August 1998
Owens Corning, for example, adopted an ES to replace 211 legacy systems. For the company to grow
An enterprise system imposes its
O W n lOiïlC O n a C npany’s strategy,
^
culture, and organization.
a semiconductor company, for exampie, says, “We plan to use SAP as a
battering ram to make our culture less
autonomous.” The manager of the ES
implementation at a computer company expresses
a similar thought: “We’ve had a renegade culture in
the past, but our new system’s going to make everybody fall into line.”
But some companies have the opposite goal.
They want to use their enterprise systems to break
down hierarchical structures, freeing their people
to he more innovative and more flexible. Take
Union Carbide. Like most companies implementHARVARD BUSINESS REVIEW
ing enterprise systems. Union Carbide is standardizing its basic business transactions. Unlike many
other companies, however, the leaders of its ES
project are already thinking in depth about how
the company will be managed differently wben the
project is completed. They plan to give low-level
managers, workers, and even customers and suppliers much hroader access to operating information.
Standardizing transactions will make Union Carbide more efficient; sharing real-time information
will make it more creative.
For a multinational corporation, enterprise systems raise another important organizational question: How much uniformity should exist in the way
it does business in different regions or eountries?
Some big companies have used their enterprise systems to introduce more consistent operating practices across their geographically dispersed units.
Dow Chemical, for instance, became an early convert to enterprise systems because it saw them as a
way to cut costs by streamlining global financial
and administrative processes. (A good idea in principle, although it became much more expensive to
achieve than Dow had anticipated.) Some large
manufacturers have been even more ambitious,
using the systems as the basis for introducing a
global lean-production model. By imposing common operating processes on all units, they are able
to achieve tight coordination throughout their
businesses. They can rapidly shift sourcing, manufacturing, and distribution functions worldwide in
response to changing patterns of supply and demand.
This capability allows them to minimize excess
manufacturing capacity and reduce both component andfinished-goodsinventory.
internationally, its chief executive, Glen Hiner, felt
it was critical to coordinate order-management,
financial-reporting, and supply chain processes
across the world. Having implemented the system
and established a new^ global-procurement organization, the company is now able to enter into larger,
more advantageous international contracts for supplies. Finished-goods inventory can be tracked daily,
both in company warehouses and in the distribution
127
ENTERPRISE SYSTEMS
cbannel, and spare-parts inventory bas been reduced
by 50%. Tbe company expects to save $65 million by
tbe end of 1998 as a result of its adoption of tbese
globally coordinated processes.
For most companies, bowever, differences in regional markets remain so profound tbat strict
process uniformity would be counterproductive. If
companies in sucb circumstances don’t allow tbeir
Hewlett-Packard. Monsanto’s managers knew tbat
different operating requirements would preclude
tbe complete standardization of data across its
agrocbemical, biotechnology, and pbarmaceuticals
businesses. Nevertheless, they placed a high priority on achieving tbe greatest possible degree of
commonality. After studying tbe data requirements of eacb business unit, Monsanto’s managers
were able to standardize fully 85% of
tbe data used in tbe ES. Tbe company
went from using 24 coding schemes
for suppliers to using just one, and it
standardized all data about materials
using a new set of substance identification codes. Wbile customer and factory data bave not been fully standardized – differences among the units’
customers
and
manufacturing processes are too
regional units to tailor tbeir operations to local cusgreat
to
accommodate
common data-Monsanto
tomer requirements and regulatory strictures, tbey
bas
achieved
a
remarkable
degree of commonality
risk sacrificing key markets to moreflexiblecomacross
a
diverse
set
of
global
businesses.
petitors. To preserve local autonomy wbile maintaining a degree of corporate control-what migbt
At Hewlett-Packard, a company witb a strong
be called a federalist operating model – a very differtradition of business-unit autonomy, management
ent approacb to enterprise systems needs to be has not pusbed for commonality across the several
taken. Rather than implement a single, global ES,
large divisions tbat arc implementing SAP’s enterthese companies need to roll out different versions
prise system. Except for a small amount of common
of tbe same system in eacb regional unit, tailored to
financial data necessary to roll up results for corposupport local operating practices. Tbis approacb
rate reporting, HP’s federalist approacb gives all the
has been taken by a number of large companies, inpower to tbe “states” where ES decisions are concluding Hewlett-Packard, Monsanto, and Nestlé.
cerned. This approacb fits the HP culture well, but
Tbey establisb a core of common informationit’s very expensive. Eacb divisional ES bas bad to be
financial, say-that all units share, but tbey allow
implemented separately, with little sbaring of reotber information-on customers, say-to be colsources. Managers estimate that well over a billion
lected, stored, and controlled locally. Tbis method
dollars will be spent across the corporation before
of implementation trades off some of tbe purity and
tbe various projects are completed.
simplicity of tbe enterprise system for greater market responsiveness.
Enterprise systems can deliver
great rewards, but the risks they
carry are equally great.
Tbe federalist model raises wbat is perhaps tbe
most difficult challenge for a manager implementing an ES: determining what sbould be common
tbrougbout the organization and what sbould be
allowed to vary. Corporate and business-unit managers will need to sit down together-well before
system implementation begins-to think through
eacb major type of information and eacb major
process in tbe company. Difficult questions need
to be raised: How important is it for us to process
orders in a consistent manner worldwide? Does tbe
term “customer” mean tbe same tbing in every
business unit? Answering sucb questions is essential to making an ES successful.
Different companies will, of course, reach very
different decisions about the right balance between
commonality and variability. Consider the starkly
different approaches taken by Monsanto and
128
Doing It Right at Elf Atochem
Considering an ES’s far-reacbing strategic and organizational implications, tbe worst tbing a company
can do is to make decisions about a system based on
tecbnical criteria alone. In fact, having now studied
more tban 50 businesses witb enterprise systems,
I can say witb some confidence tbat the companies
deriving the greatest benefits from tbeir systems
are tbose tbat, from tbe start, viewed them primarily in strategic and organizational terms. They
stressed the enterprise, not tbe system.
Elf Atocbem Nortb America, a $2 billion regional
chemicals subsidiary of tbe French company Elf
Aquitaine, is a good case in point. Following a series of mergers in the early 1990s, Elf Atocbem
found itself hampered by tbe fragmentation of critical information systems among its 12 business
units. Ordering systems were not integrated witb
HARVARD BUSINESS REVIEW
fuly-August 1998
ENTERPRISE SYSTEMS
production systems. Sales forecasts were not tied to
budgeting systems or to performance-measurement systems. Each unit was tracking and reporting its financial data independently. As a result of
the many incompatible systems, operating data
were not flowing smoothly through the organization, and top management was not getting the information it needed to make sound and timely
business decisions.
The company’s executives saw that an enterprise
system would be the best way to integrate the data
fiows, and they decided to go with SAP’s R/3 system, which was rapidly becoming the standard in
the industry. But they never labeled the ES project
as simply a technology initiative. Rather, they
viewed it as an opportunity to take a fresh look at
the company’s strategy and organization.
Looking beyond the technology, the executives
saw that the real source of Elf Atochem’s difficulties was not the fragmentation of its systems but
the fragmentation of its organization. Although the
12 business units shared many of the same customers, each unit was managed autonomously.
From the customer’s perspective, the lack of continuity among units made doing husiness witb the
company a trial. To place a single
order, a customer would frequently
have to make many different phone
calls to many different units. And to
pay for the order, the customer would
have to process a series of invoices.
Inside the company, things were
equally confused. It took four daysand seven handoffs between departments-to process an order, even though only four
hours of actual work were involved. Because each
unit managed inventory and scheduled production
independently, the company was unable to consolidate inventory or coordinate manufacturing at the
corporate level. More than $6 million in inventory
was written off every year, and plants had to be shut
down frequently for unplanned production-line
changes. And because ordering and production systems were not linked, sales representatives couldn’t
promise firm delivery dates, which translated into
lost customers.
Management knew that in the petrochemicals
business, where many products are commodities,
the company that can offer the hest customer service often wins the order. So it structured the implementation of its ES in a way tbat would enable
it to radically improve its service levels. Its goal was
to transform itself from an industry laggard into an
industry leader. Even though many competitors
were also adopting the R/3 package. Elf Atochem
knew that if it could achieve a tighter, smoother fit
between its business processes and the system, it
could gain and maintain a service advantage.
The company decided to focus its efforts on four
key processes: materials management, production
planning, order management, and financial reporting. These cross-unit processes were the ones most
distorted by the fragmented organizational structure. Moreover, they had the greatest impact on the
company’s ability to manage its customer relationships in a way that would both enhance customer
satisfaction and improve corporate profitahility.
Each of the processes was redesigned to take full advantage of the new system’s capabilities, in particular its ability to simplify the flow of information.
Layers of information middlemen-once necessary
for transferring information across incompatible
unit and corporate systems-were eliminated in
order to speed theflowof work and reduce the likelihood of errors.
To maintain its focus on the customer, the company chose to install only those R/3 modules required to support the four targeted processes. It did
not, for example, install the modules for human resource management or plant maintenance. Those
Those companies that stressed
the enterprise, not the system,
gained the greatest benefits.
HARVARD BUSINESS REVIEW
July-August 1990
functions did not have a direct impact on customers, and the existing information systems that
supported them were considered adequate.
Elf Atochem also made fundamental changes to
its organizational structure. In the financial area,
for example, all the company’s accounts-receivable
and credit departments were combined into a single
corporate function. This change enahled the company to consolidate all of a customer’s orders into a
single account and issue a single invoice. It also allowed the company to monitor and manage overall
customer profitability-something that had been
impossible to do when orders were fragmented
across units. In addition. Elf Atochem combined all
of its units’ customer-service departments into one
department, providing each customer with a single
point of contact for checking on orders and resolving problems.
Perhaps most important, the system gave Elf
Atochem the real-time information it needed to
connect sales and production planning-demand
129
ENTERPRISE SYSTEMS
and supply-for tbe first time. As orders are entered
or changed, the system automatically updates forecasts and factory schedules, which enables the
company to quickly alter its production runs in response to customers needs. Only one other company in the industry had this capability, which
meant that Elf Atochem gained an important edge
over most competitors.
The company understood, however, that just
having the data doesn’t necessarily mean the data
processing, supplier management, and financial
reporting. The unit-by-unit process ensures that
the effort is manageable, and it also helps the team
refine the system and the processes as it proceeds.
For example, the second unit to implement the system found that it didn’t adequately support bulk
shipments, which are the main way the unit gets its
products to customers. [The first unit uses package
shipping for all its orders.) The system was then
modified to support bulk as well as package shipping, and the new configuration became the new standard.
Using the large and broadly representative implementation team, together with the unit-by-unit rollout.
Elf Atochem has been able to staff the
effort mainly with its own people. It
has had to engage only nine outside
consultants to assist in the projectfar fewer than is usually the case. The
reliance on internal resources not
only reduces the cost of the implementation, it also
helps ensure that Elf Atochem’s employees will
understand how the system works after the consultants leave.
Elf Atochem’s ES is now more than 75%
complete-9 of the 12 business units are up and
running on the new system-and the rollout is
ahead of schedule and under budget. Customer satisfaction levels have already increased, and the
company is well on the way to its goal of confirming 95% of all orders with one call, a dramatic improvement over the previous average offivecalls. In
addition to the service enhancements, the company
is operating more efficiently. Inventory levels, receivables, and labor and distribution expenditures
have all been cut, and the company expects the system will ultimately reduce annual operating eosts
hy tens of millions of dollars.
Only a general manager is
equipped to act as a mediator
between the imperatives ofthe
technology and ofthe business.
will he used well. Computer systems alone don’t
change organizational behavior. It therefore established a new position-demand manager-to be the
focal point for the integrated sales and productionplanning process. Drawing on the enterprise system, the demand manager creates the initial sales
forecast, updates it with each new order, assesses
plant capacity and account profitability, and develops detailed production plans. The demand manager is able to schedule a customer’s order-and
promise a delivery date-up to six weeks ahead of
production. Previously, production could he allocated to individual orders no more than a week in
advance. Now central to the company’s operation,
the role of demand manager could not even have existed in the past because the information needed to
perform it was scattered all over the company.
The way Elf Atochem is managing the implementation effort also refiects the breadth of its
goals. The project is being led by a 6o-person core
implementation team, which reports to a member
of the company’s executive committee. The team
includes both business analysts and information
technologists, and is assisted by a set of so-called
super users, representing the business units and
corporate functions. These super users help ensure
that decisions about the system’s configuration are
made with the broadest possible understanding of
the business. They also play a crucial role in explaining the new system to their respective departments and training people in its use.
The team is installing the ES one business unit at
a time, with each unit implementing the same system configuration and set of procedures for order
130
The Role of Management
Every company that installs an ES struggles with
its cost and complexity. But the companies that
have the biggest problems-the kind of problems
that can lead to an outright disaster-arc those that
install an ES without thinking through its full business implications.
Managers may well have good reasons to move
fast. They may, for example, have struggled for
years with incompatible information systems and
may view an ES as a silver bullet. They may be
looking for a quick fix to the Year 2000 problem
(enterprise systems arc not infected with the muchfeared millennium bug). Or they may be trying to
HARVARD BUSINESS REVIEW
July-August 1998
ENTERPRISE SYSTEMS
keep pace with a competitor that has already implemented an ES. The danger is that while an enterprise system may help them meet their immediate
challenge, the very act of implementing it may create even larger problems. A speedy implementation
of an ES may be a wise business move; a rash implementation is not.
A numher of questions should be answered before any decisions are made. How might an ES
strengthen our competitive advantages? How
might it erode them? What will be the system’s
effect on our organization and culture? Do we need
to extend the system across all our functions, or
should we implement only certain modules?
Would it be better to roll the system out globally or
to restrict it to certain regional units? Are there
other alternatives for information management
that might actually suit us better than an ES?
The experience of Elf Atoehem and other successful adopters of enterprise systems underscores
the need for careful deliberation. It also highlights
the importance of having top management directly
involved in planning and implementing an ES. Not
only is Elf Atochem’s executive committee overseeing its ES project, but its entire board reviewed
and approved the plans. At Compaq, the decision to
go with an ES was also made at the board level, and
the senior management team was involved with
the implementation every step of the way.
Many chief executives, however, continue to
view the installation of an ES as primarily a technological challenge. They push responsibility for it
down to their information technology departments. Because of an ES’s profound business implications-and, in particular, the risk that the technology itself might undermine a company’s
strategy-off-loading responsibility to technologists is particularly dangerous. Only a general manager is equipped to act as the mediator between the
imperatives of the technology and the imperatives
of the business. If the development of an enterprise
system is not carefully controlled by management,
management may soon find itself under the control
of the system.
^
Reprint 98401
To order reprints, see the last page of this issue.
T
“It appears the Ethics Committee has gone over the edge.
HARVARD BUSINESS REVIEW
fuIy-August 1998
131

Purchase answer to see full
attachment