Implementing New Technology
For all the dollars spent by American companies on R&D, there often remains a persistent and troubling gap between the inherent value of the technology they develop and their ability to put it to work effectively. At a time of fierce global competition, the distance between technical promise and genuine achievement is a matter of especially grave concern. Drawing on their long study of the difficulties managers have had in closing this gap, the authors identify half a dozen key challenges that managers responsible for implementing new technology must surmount: their inescapably dual role, the variety of internal markets to be served, legitimate resistance to change, the right degree of promotion, the choice of implementation site, and the need for one person to take overall responsibility.
Introducing technological change into an organization presents a
different set of challenges to management than does the work of competent
project administration. Frequently, however, the managers responsible for
shepherding a technical innovation into routine use are much better equipped by
education and experience to guide that innovation’s development than to manage
its implementation.
In the following pages, we describe some of the challenges managers must
overcome if companies are to absorb new technologies efficiently. We also
suggest strategies managers can use to address these difficulties. Although the
examples we cite are all computer related and come from the experience of large
manufacturers, the issues raised and strategies proposed apply every bit as
well to small businesses, to service operations—in fact, to any organization
where technological innovation flourishes.
A DUAL ROLE
Those who manage technological change must often serve as both technical
developers and implementers. As a rule, one organization develops the
technology and then hands it off to users, who are less technically skilled but
quite knowledgeable about their own areas of application. In practice, however,
the user organization is often not willing—or able—to take on responsibility
for the technology at the point in its evolution at which the development group
wants to hand it over. The person responsible for implementation—whether located
in the developing organization, the user organization, or in some intermediary
position—has to design the hand-off so that it is almost invisible. That is,
before the baton changes hands, the
runners should have been running in parallel for a long time. The
implementation manager has to integrate the perspectives and the needs of both
developers and users.
Marketing Perspective
That involving users in a new technology’s design phase boosts user
satisfaction is quite well known, but the proper extent, timing, and type of
user involvement will vary greatly from company to company. For example,
software developers in an electronic office equipment company established a
user design group to work with developers on a strategically important piece of
applications software when the program was still in the prototype stage.
Prospective users could try out the software on the same computer employed by
the program’s developers. The extremely tight communication loop that resulted
allowed daily feedback from users to designers on their preferences and
problems. This degree of immediacy may be unusual, but managers can almost
always get some information from potential users that will improve product
design.
A marketing perspective also helps prepare an organization to receive
new technology. Many implementation efforts fail because someone underestimated
the scope or importance of such preparation. Indeed, the organizational hills
are full of managers who believe that an innovation’s technical superiority and
strategic importance will guarantee acceptance. Therefore, they pour abundant
resources into the purchase or development of the technology but very little
into its implementation. Experience suggests, however, that successful
implementation requires not only heavy investment by developers early in the
project but also a sustained level of investment in the resources of user
organizations.
A very promising implementation effort in a large communications and
computer company went off the rails for many months because of inadequate
infrastructure in the user organization. New computerized processing control
equipment was ready for shipment to prospective users enthusiastically awaiting
its arrival, but a piece of linking software was not in place. Arguments
erupted over who should pay for this small but critical piece of the system.
Equally troubling, there were no resources for training because the developers
did not see providing these resources as part of their normal
responsibilities. No one in the user organization had prepared the way
for the innovation, so there was no one to whom developers could hand it off.
Framework for Information
Just as marketing managers carefully plan the research through which
they will gather critical product information, so implementation managers must
develop an iterative, almost accordion-like framework to guide decisions about
when and how to collect needed information from all groups affected by an
innovation. We say “accordion-like” because the process necessarily involves a
search for information, a pause to digest it, and then another active period of
search—cycle after cycle. What information is important—and who has it—may vary
at different stages of the implementation process, but someone must coordinate
the iterative work of gathering it—and that someone is the implementation
manager.
When, for example, a turbine manufacturer designed a CNC system for
shop-floor control in one of its small parts operations, project managers were
careful to:
· Observe the
current job routine. System designers visited the factory floor several times
and each time interviewed eight to ten operators about their work procedures.
· Pay special
attention to those parts of the work that required users to make decisions or
seek information about which tools or materials to use, which sequence of steps
to follow in machining, and which jobs operators ought
· Discuss with
workers what they found especially frustrating or rewarding about their work.
In this case, it turned out that they liked some flexibility in the sequencing
of jobs, felt that the choice of materials should be theirs, and were often
frustrated by the difficulty of finding tools.
· Examine how
this manufacturing process related to others. The machine operators were
extremely dependent on materials personnel, maintenance, the tool room, and
order expediters.
From their discussions with operators, the system designers could
understand the important variables as the operators saw them and, therefore,
could design a system that solved problems the operators really faced without
creating new
ones. These discussions also facilitated a transmission of information
back to the users through education and hands-on practice sessions with the
users and their supervisors.
MULTIPLE INTERNAL MARKETS
The higher the organizational level at which managers define a problem
or a need, the greater the probability of successful implementation. At the
same time, however, the closer the definition and solution of problems or needs
are to end-users, the greater the probability of success. Implementation
managers must draw up their internal marketing plans in light of this apparent
paradox.
As these managers identify the individuals or groups whose acceptance is
essential to an innovation’s success, they must also determine whom to
approach, when, and with which arguments. Top management and ultimate users
have to buy into the innovation to make it succeed, but marketing an idea to
these two groups requires very different approaches. How, then, can an
implementation manager foster general acceptance of an innovation from such a
range of constituencies? We believe this executive must view the new technology
from the perspective of each group and plan an approach to each accordingly.
Top management, most concerned with an innovation’s likely effect on the
bottom line, is accustomed to receiving proposals that specify return on
investment and paybacks. Many of today’s computerized technologies, however, do
not lend themselves to justification in traditional financial terms, yet they
may be essential to a company’s future. Amid growing calls for the accounting
profession to provide better means to assess the value of robots, CAD, and
computer-integrated manufacturing, some companies are beginning to realize the
limitations of traditional capital budgeting models.1
When GE set up its state-of-the-art automated dishwasher plant, it
originally justified the costs on the basis of savings over time, but the plant
has experienced payoffs from the investment in unanticipated ways. The quality
of the product improved, lower manufacturing costs led to an expansion of
market share, and the plant proved able to serve as a manufacturing site for
other products. Each time managers document such nontraditional benefits, they
make it easier to justify similar investments later.
Top executives may also be swayed by strategic considerations. When
large-scale automation was introduced into GE’s large steam-turbine generator
business, the innovation was sold to top management on the basis of changing
business needs: a shift from the manufacture of large, one-of-a-kind products
to the manufacture of small parts. The new systems also helped drive the
continual quality improvements needed to keep operations competitive when the currently
sluggish market revived.
Selling top management on the case for new technology—without
simultaneous involvement of user organizations in the decision-making
process—is not enough. It is equally important for users of an innovation to
develop “ownership” of the technology. The meaning of this term depends largely
on the scope of the project. Although it is patently impossible to involve all
users in the choice and/or development of an innovation, that is no excuse not
to involve their representatives.
Perhaps even more important is to plan for the transfer of knowledge
from the old operation, in which people knew the materials and the product very
well, to the new process, which outsiders may initially design and run. The
developers of the new process (especially when it is computer software) often
know their tools very well, but rarely do they understand the materials and
processes to which their software is applied as well as the people on the plant
floor who have been working with both for years. At the very least, managers
should provide some mechanism and time for such knowledge to flow from
experienced worker to developer.
An example of well-developed ownership is the case of a marketing
organization about to switch from manual files to an electronic filing,
messaging, and data retrieval system used by both account officers and
secretaries. Managers decided to take the time to do it right the first time
instead of doing it over. The project manager set up a committee of elected
representatives from all groups affected. This committee met regularly, first
to select the right software package and then, when it became apparent that
they would have to build their own system to get all the features they wanted,
to give advice on its structure and content.
The result was an inventive, well-accepted, and widely used system.
Moreover, users regarded the minor problems that did arise as bugs to be
worked out of our system. As one manager told us, “The
users wanted it, so they built it.”
Critical to the success of this project was the choice of opinion
leaders among users for involvement. Managers who have wrought change have
known for a long time that the opinions of a few leaders profoundly influence
the speed and extent of an innovation’s diffusion. The basis for leadership
differs from organization to organization, but these leaders are not usually
hard to identify. Frequently, they occupy their place of influence as a result
of technical proficiency, not formal position.
Opinion leaders, however, are not necessarily the most skilled
operators. Behavioral science studies have shown that people commonly seek two
kinds of credibility in such leaders: “safety” credibility (this person is
enough like me for his opinions to be trusted) and “technical” credibility
(this person knows what she is talking about). Someone whose technical skills
are so superior that followers can have no hope of emulation may fall too far
outside the norms of a group to be a real opinion leader.
In the marketing organization just described, one senior account manager
refused to use the new electronic system. The system implementers were at first
alarmed but then realized that this individual was not an opinion leader. Their
efforts flowed around him, unimpeded by his opposition. Six months after
everyone else went on the system, he capitulated, convinced at last of its
utility.
PROMOTION VS. HYPE
Many a technology developer will confess bewilderment that innovations
do not win automatic acceptance. It may be overly optimistic to believe that an
innovation will sell itself, but it is equally dangerous to oversell the new
system. Novel and exotic technologies are especially vulnerable to hype.
Articles in the media about robots and artificial intelligence, for
example, have raised expectations far higher than the actual performance of
current technologies warrants. Potential users quickly grow disillusioned when
much touted innovations perform below expectations. When one computer maker
developed artificial intelligence software to be used in manufacturing, the
outside world thought it was a finished product long before it was out of the
“vaporware” stage. Months before they had their hands on the software,
intended users faced questions from their customers about how they liked it.
The gap between perception and reality was traceable to the energetic
efforts of one project manager early on. Knowing the importance of selling the
concept to management, this enthusiast had extended his campaign to virtually
anyone who would listen. Since it was a sexy topic, the new artificial
intelligence system received wide attention in the media as well as in
organizational newsletters. This oversell presented a problem to implementation
managers, who had to fight the perception that their project was way behind
schedule and that their product delivered less than promised.
RISKY SITE, SAFE INNOVATION
There are two reasons for conducting a pilot operation before
introducing an innovation across the board in a large organization: first, to
serve as an experiment and prove technical feasibility to top management and,
second, to serve as a credible demonstration model for other units in the
organization. These two purposes are not always compatible.
If the innovation must succeed at the pilot site in order to survive
politically, the implementation manager may choose a site that poses virtually
no risk but that neither offers real benefit to the organization nor
establishes a model for other units. At the same time, however, if the trial is
to be a credible test, it cannot take place among the most innovative people in
the corporation. Success at this kind of site is vulnerable to the criticism
that these users are far from typical.
Testing the new technology at the worst performing unit, even though it
may be where the innovation is most needed and would show the most spectacular
results, is no better a choice. If the project fizzles, the implementation
manager will not know how much of the failure was caused by extraordinary
problems with the site and how much by the inherent properties of the
technology. If the project succeeds, critics will be quick to note that
anything would have helped operations at that site.
The solution, therefore, is to be clear about the purpose of the
test—experimental or demonstration—and then to choose the site that best
matches the need. The customized end of one large computer manufacturer’s
business
suffered from a problem. If customers canceled orders, the partially
built systems were either totally scrapped—that is, broken down into components
and sent back to the warehouse—or matched with incoming orders to determine if
the fit was close enough to warrant retrofitting. When this matching process,
which had been done manually, was computerized, the first applications site was
an operation with an enthusiastic champion, but it was to be phased out in a
matter of months. The site was politically risk free but not useful for a
demonstration. Although the first application was successful, the operation
closed down before the site could serve as a demonstration for other plants,
and the implementation manager in charge of the next site had to start all
over.
Consider a different example: a paper maker that chose one of its
high-visibility mills as the first site for an expensive, large-scale
computerized control system. Although the system was needed to boost sagging
profit margins, the mill was neither the company’s best nor worst operation in
financial terms. Local management was determined to see the system succeed for
the sake of the mill; corporate management viewed it as an experiment. The site
was promising but not risk free.
Even if managers realize that the trial of a new technology is a
critical demonstration, they do not always ask the next question: a
demonstration for whom? The physical and organizational position of the first
site will heavily influence who the next wave of users will be.
Over the years, many studies have shown a strong inverse relationship
between proximity to facilities and use of them. This result is not surprising
if the distance is measured in miles. What is surprising is that out of
sight—no matter by how much—generally means out of mind. The difference in the
use of a library by engineers on a college campus depended on how many more
feet, not miles, nonusers were from the library than users. Similarly, new
computer terminals in a large oil company were used first by people with
adjoining offices and only reluctantly by people even a few more feet down the
hall. Distance is a relative, not absolute, measure to be weighed against
current routine rather than against any objective standard.2
Obviously, it is not always possible to site new equipment for
everyone’s convenience. Even so, the placement of an innovation frequently
determines who uses the new technology first and most. If the equipment is
located
farther away from older or more reluctant potential users, they have a
ready excuse for avoiding it. Consequently, managers who do not consider
physical layout in their implementation strategies may, by default, select as
first users people with little or no influence in the organization.
As noted earlier, involving opinion leaders in the planning process
helps smooth the path of implementation. If the first users of a new technology
are credible role models (neither extraordinarily adept nor very poorly
skilled), their demonstration has heightened meaning for a wide audience.
Sometimes these opinion leaders strongly resist the technology, and getting
even one of them to use it can create the necessary crack in the dam. Getting
them to try the innovation may require nothing more elaborate than a well-paced
and tactfully presented training session.
Often, however, an implementation manager has to create new role models
by siting the innovation where the workers most open to change can demystify
the technology for others by using it themselves. Although it is definitely a
mistake to correlate resistance with age per se, it remains true that people
with a long-term investment in certain routines and skills often hesitate to
give up the security of those habits. Again, it is best to avoid extremes and
to site new technology near workers who are fairly open to change but not so
different from those whose resistance makes them poor models.
When a large warehouse installed a materials handling system, it relied
on its so-called “hippy” crane operators instead of workers on the loading
platform. Once the crane operators had worked out the wrinkles, management
could progressively install the system throughout the plant. The crane
operators were not opinion leaders at first because of their relative youth and
different backgrounds, but they were both receptive to innovation and not so
very different as to be unacceptable role models.
THE MANY & THE ONE
If an innovation is to succeed, the implementation team must include (1)
a sponsor, usually a fairly high-level person who makes sure that the project
receives financial and manpower resources and who is wise about the politics of
the organization; (2) a champion, who is salesperson, diplomat, and problem
solver for the innovation; (3) a project manager, who oversees administrative
details; and (4) an integrator, who manages conflicting
priorities and molds the group through communication skills. Since these
are roles, not people, more than one person can fulfill a given function, and
one individual can take on more than a single role.
Even if all these roles are filled, however, the project can still stall
if the organization does not vest sufficient authority in one person to make
things happen. One of these individuals—usually the sponsor or the
champion—must have enough organizational power to mobilize the necessary
resources, and that power base must encompass both technology developers and
users.
There are, of course, many ways to mobilize supplies and people. By
encouraging ownership of an innovation in a user organization, for example,
skillful advocates can create a power base to pull (rather than push) the
innovation along. But enthusiasm for a new technology is not enough. New
technology usually requires a supportive infrastructure and the allocation of
scarce resources for preparing the implementation site. A champion based in the
development group with no authority among the receivers must rely on
time-consuming individual persuasion to garner the necessary resources.
Further, even if prospective users believe in an innovation’s worth, they may
have to convince their superiors to free up those resources.
A short case will illustrate the point. A manufacturer of engineering
test equipment was in trouble because many orders for its customized products
reached the plant floor missing vital components. Technical experts were able
to catch omissions and incorrect selection of parts before the orders went into
production, but the mechanics of checking orders and cycling them back through
the purchase-order process cost enormous amounts of time, money, and customer
goodwill. Customers were angry at the delay of orders for weeks when manufacturing
bounced them back to the initial salespeople and were even more dismayed when
price quotations had to be revised upward because of a part forgotten in the
first go-around.
An internally developed technology offered a partial solution: a
computer program could automatically check the orders before salespeople issued
quotations. Although the people who placed the orders were enthusiastic about
the concept, the work of implementing the system was fraught with problems. No
sales manager was willing to function as either sponsor or champion for the
innovation. Although a user group funded its development, the appointed
champion in that organization was too low in the hierarchy to
control the resources necessary to install the system. Moreover, he
lacked a clear endorsement for the project from his superiors and had mixed
feelings toward the innovation. He believed in its purpose but was not certain
it was being developed correctly and was afraid to stand behind it
wholeheartedly lest it fail in the field. He was, therefore, slow to seek the
resources and upper management support that would have moved the project
forward quickly. Ultimately, an innovation has to be one person’s
responsibility.
LEGITIMATE RESISTANCE TO CHANGE
Overt resistance to an innovation often grows out of mistakes or
overlooked issues in an implementation plan. Tacit resistance does not
disappear but ferments, grows into sabotage, or surfaces later when resources
are depleted. Because the advocates of change have such a clear view of an
innovation’s benefits, resistance often catches them by surprise. The worst
thing a manager can do is shrug such resistance aside on the dual assumption
that it is an irrational clinging to the status quo and that there is nothing
to be done about it. Clinging to the status quo it may indeed be—but
irrational, rarely. And managers can do something about it.
Thus the beginning of wisdom is to anticipate opposition. An innovation
needs a champion to nurture it, and any new technology capable of inspiring strong
advocacy will also provoke opposition. Where there are product champions, there
will also be innovation assassins. Assassins, moreover, can fell a project with
just one well-aimed bullet, but champions need to marshal forces and nurture
support to implement new technology in the face of resistance. The most common
reasons for opposition to a new technology are fear of the loss of skills or
power and absence of an apparent personal benefit.
Fear of Loss
As talk about the deskilling potential of new computerized technologies
has grown, unions are seeking retraining for their members whom automation
would otherwise displace. Many companies are upgrading the status of their
workers who are forced to trade hard-earned manual skills for the often dreary
routine of button pushing. Although the problem is far from being resolved, it
has at least merited recognition.
There is, however, another aspect of deskilling that has been much less
obvious to implementers: the simple necessity of extending concern about
deskilling to foremen and supervisors. They do not, of course, actually have to
run the new machinery or to possess the intimate knowledge of the system that
daily operators do. Even so, giving subordinates knowledge that supervisors and
foremen do not have undermines their credibility. If the foremen or supervisors
worked their way up through the ranks, they will know the old machinery well.
They served as problem solvers when it broke down and derived no small part of
their authority from their experience with it. To train their subordinates and
leave them out is to invite hostility.
When a pulp mill introduced a new computerized control room, vendor
representatives trained the operators and their assistants. No similar effort
was made for the foremen, who thought (with some justification) that they had
lost control over the mill’s operation. Some of the operators relinquished
their novel power by tactfully educating their foremen, but others felt they
had earned the right to more autonomy because the foremen’s knowledge was
obsolete.
One way to deal with this kind of situation is to teach supervisors how
to instruct hourly workers about the new technology. These sessions should
transmit details of the information hourly workers require, instructions on how
best to present it, guides to practice sessions, and audiovisual aids.
Another reason for resistance is fear that the innovation will be
politically enfeebling and that supervisors and even operators will lose some
control by adopting it. A good implementation plan should try to identify where
a loss of power may occur so that managers can anticipate and possibly avert
any problems arising from that loss.
In one large manufacturing plant, corporate research developed a
computerized system for scheduling the production—in small batches—of
customized health care products. Although the manufacturing manager outwardly
supported the idea, he never made any of the decisions or appointments
necessary to put the new technology into effect. The implementation team
finally realized what he had seen at the outset: using the software removed
from his hands control over a key piece of his operation. The programmers
working on the project reported to management information systems (MIS), not
manufacturing. The manager never voiced his opposition since there was little
rational basis for it, but his resistance effectively stalled the project. In
this case, the programmers were quite willing, as was MIS, to report for the
duration of the project to manufacturing, a change that allowed the project to
swing into place.
Personal Benefit
An innovation must offer an obvious advantage over whatever it replaces,
or potential users will have little incentive to use it. The more visible the
costs of an innovation (financial, convenience, the need to learn new skills),
the greater the importance of making potential benefits and rewards apparent.
These benefits include expanded influence over work (stopping a production
line), increased value of work (no in-process inventory), greater recognition
(being part of a valued implementation team), solution of a longstanding
problem (a shop-floor control system that gives up-to-the-minute production
reports), and preservation of jobs.
It is easy for managers to forget that benefits buried in the system,
which they can see because of their position, may be totally invisible to the
operators on whom the success of the innovation depends. A new technology may
pay off for an organization as a whole but not for individuals in any form they
can recognize. That is why it is so important to make these benefits visible
through encouragement from supervisors as well as through explicit and timely
feedback on how the innovation is affecting workers’ output. In general, the
faster the positive feedback to users, the more visible the benefits will be.
A very large natural resources company ran into difficulties with
introduction of a methodology for constructing software. This approach required
programmer-analysts to sit down with their clients and, following a regimented
procedure with standardized notations, analyze the client’s business. A
structured approach was expected to identify more potential problems at the
design stage and facilitate communication between client and designer. Moreover,
the company hoped that a standardized notation would facilitate the transfer of
project work between programmers and cut the time spent on program maintenance.
In retrospect, it is clear that all the benefits of the new technology
accrued to the organization, not to the individuals who used it. In fact, many
potential users thought they would be penalized for using the new methodology,
since managers judged their performance on speed and low cost, not on the
quality of their output. The organization’s rhetoric supported, indeed
mandated, use of the new technology, but the reward structure militated against
it.
Now, contrast this situation with one in which managers gave some
thought to the challenge of translating organizational benefits into individual
rewards. Before installing a shop-floor control system, a major appliance
manufacturer conducted informal research into the problems of the hourly work
force. They discovered that the current voucher system never permitted workers
to know how much their pay would be in a given week. A small modification of
the control system’s design made it possible for employees to receive a report
on cumulative salary with each job they entered. Although this piece of
information was not central to the needs of the organization, adding it to the
system’s design was a low-cost way to boost the innovation’s benefits to
workers. This small feature more than compensated them for the pain of
developing new skills and habits, and the advantage of the new system over the
old was apparent every time they used it.
A WORD ABOUT HEDGERS
Besides the champions and assassins in an organization, there will
always be some “hedgers,” individuals who refuse to take a stand against an
innovation so that others can address their objections but who also refuse to
support the new technology. They straddle the fence, ready to leap down on
either side to declare that they had foreseen the value of the innovation all
along or that they had known it would fail from the start. These risk-averse
managers can affect
the future of a new technology when they are a key link in the
implementation plan. Because these hedgers are usually waiting for signals to
tell them which way to leap, astute implementation managers will see to it that
they receive the appropriate signals from those higher up in the organization.
Like product assassins, hedgers can be found at any level in an
organization, and dealing with them effectively requires a sequence of actions.
The first, and the easiest, is to persuade top management to take some kind of
quick symbolic action in support of the innovation. Whether the action takes
the form of a memo, a speech, or a minor policy change, it must send a signal
that top management will stand behind this technology even in a budget crisis.
The second step, which is harder, is to help managers at all levels send
out the right signals. If, for instance, the first step was an announcement of
a new drive for quality, the second should be to increase the emphasis on
quality throughout the company. If workers hear an announcement about a new
quality program but continue with impunity to ship products that they know are
inferior, the initial symbolic gesture loses potency. Worse, all future
gestures lose credibility too.
The third step is the hardest—and the most necessary. Managers must
bring the criteria used to judge the performance of innovation users into
conformance with the demands of the new technology. New technologies often
require new measures. If, for example, a new, structured software technique
requires more time than did the old, managers must evaluate programmer-analysts
less on the basis of the quantity of output than on the basis of its quality.
Further, because productivity commonly declines whenever a new
technology is introduced, more accurate measurements of productivity in the old
sense may lead supervisors to fear that their performance will look worse—not
least because, with a fully automated system, direct labor drops but indirect
labor grows.
Other adjustments might include a phase-in period for the new technology
during which the usual output measurements do not apply. It might also make
sense to reward people for preventing rather than just solving problems and for
developing work behavior identified with the new technology. Although operators
do not respond well when they view technological systems as controlling their
behavior, they respond quite well when a system gives them
feedback on their performance and the performance of their machines.
Information increases the amount of control people have over their environment.
Converting hedgers into believers is not a simple task, but it is one
more of the inescapable challenges managers face as they try to implement new
technology. Indeed, as the competitive environment changes and as the
systematic effects of new technologies become even more pronounced, the work of
implementing those technologies will increasingly pose for managers a
distinctive set of challenges—not least, the task of creating organizations
flexible enough to adjust, adapt, and learn continuously.