Similarly, customers did not
need to understand their suppliers'
prior experiences and capabilities,
since these were reflected in the
products they could see, touch
and experience directly.
The services transaction is
different. The exchange is
generated by both parties,
and the process of adoption
or consumption is an
integral part of the transaction.
So, the adopter or
customer is also a co-producer,
intimately involved
in defining, shaping and
integrating the service into
his or her organisation.
The supplier of the service
can extend an offer of
what is to be provided
but, as we shall see below,
it cannot entirely specify the
requirements of the service.
Instead, the supplier designs its
system to elicit this information
from its customers, and modifies
the offering in response to customers'
needs before sale. In turn,
customers select their service
provider on the basis of the capabilities
they offer, and the extent to
which the customer is able to shape
those capabilities to serve their particular
needs.
x The Nature of Knowledge
This leads to a consideration of the
nature of the two types of knowledge
that need to be considered in a services
exchange. Codified knowledge
is information that is well understood
by providers and adopters
owing, for example, to common language,
customs, media and culture.
This type of knowledge is also developed
within more technical areas,
when technical standards represent
the codification of information across
multiple entities, such as the html
and http protocols on the internet or
the digital video disk (DVD) format
for movies. These standards enable
information to transfer between
physical devices in ways that are predictable
in advance. When knowledge
is standardised in this way,
parties can exchange services with
each other even though they may be
otherwise not known to each other.
Tacit knowledge is experiential
knowledge that has not been reduced
to a codified form. A classic example
economies produce tangible
goods that are the primary
focus of exchange in the
economy. Crucially, key information
comes embedded in the products
being traded.
Services exchange is qualitatively
different. It involves a negotiated
exchange between a provider and
an adopter (supplier and customer)
for the provision of (predominately)
intangible assets.
The absence of a central product
raises an important corollary: each
party in the exchange needs the
other's knowledge in negotiating it.
On the one hand, the provider lacks
the contextual knowledge of the
customer's business and how the customer
is going to leverage the offering
to compete more effectively in the
market. At the same time, the customer
does not know the full capabilities
of the provider's technologies or
its experience from other transactions
in assessing what will work best.
When products were the main
focus, the information they contained
helped each side communicate
effectively with the other. As
products and their functions became
better understood, suppliers did not
need to understand the customer's
business to be an exchange partner.
reflected the dominant
economic shift of the
time from an agricultural economy to
a manufacturing one.
At that time, services was a residual
category used to describe activities
that did not fit into either of the
two other groupings. It was a small
sector of the economy, so shoving
"other" stuff into the category was
reasonable. As demonstrated above,
today that residual is the bulk of economic
activity and by far the fastestgrowing
part of economic activity in
the advanced economies.
The leading role of services in the
economy comes as no surprise to
many companies who were leaders
in the manufacturing sector. Today,
businesses such as GE and Xerox
find that services are the fastestgrowing
parts of their businesses.
Indeed, IBM earns the majority of its
revenues from its IBM Global Services
business, a unit that did not
exist prior to the 1990s.
x Innovation in services
What is different about innovation
in services as opposed to products?
One clear distinction is the intangible
nature of the services activity.
Both agricultural and manufacturing
ccording to a
study by the
N a t i o n a l
Academy of
Engineering,
services in 2003
represented 80
per cent of the
US's gross domestic
product. And according to the OECD,
they account for a similar percentage
of economic activity across all
advanced industrial economies. Despite
this, most analyses of innovation
tend to focus on products, not services.
It is now time to update our
curriculum for teaching and researching
innovation to address the dominant
sector of economic activity in
most advanced economies.
The National Academy of Engineering
study surveyed the contribution
of academic research to industrial
performance in seven industries. In
the five product-based sectors ­ network
systems, communications, medical
devices, equipment and aerospace academic
research was found to have
had a very significant effect. However,
in the two service industries ­ transportation
and financial services ­ it
had only a limited impact. Further,
the study concluded that "the academic
research enterprise has not
focused on or been organised to meet
the needs of service businesses."
The term "services" itself is confusing.
There is a tendency for people
to use it as a prefix or suffix to
some other knowledge domain, such
as "financial services" or "services
marketing". As a result, academics
have no shared sense of "services".
The accidental history of the term
helps to explain this uncertainty. It
emerged in the early 20th century, when
a taxonomy of the economy defined
the major economic sectors as agriculture,
manufacturing and services. This
6
HENRY CHESBROUGH
Services dominate economic activity
in developed economies, and yet
understanding of innovation in
this sector remains very limited
Henry Chesbrough is
executive director of the
Center for Technology,
Strategy and Management
at the Haas School of
Business, University of
California, Berkeley.
chesbrou@haas.berkeley.edu
A JOHN McFAUL
A failing grade for the
innovation academy
of use has been enormously valuable,
not least to Apple's shareholders.
x The role of service
innovators
To build new and useful systems,
services innovators need to do two
things. First, they must learn how
to deconstruct complex knowledge
and how to integrate, recombine
and reuse it from one instance to
another. Innovators are now building
information technology tools to help
with this process.
One such tool is called "business
process modelling", and is used by
service providers such as IBM Global
Services, Accenture and Infosys. This
tool analyses a business process by
breaking it down into individual
activities, and noting the links that
connect the activities together to
deliver a particular function. Such
mapping is an attempt to codify the
knowledge in this process ­ how the
different activities in a business
process connect to one another.
Ideally, elements that recur frequently
in a variety of customer
processes can be recombined in
a range of ways to serve different
potential purposes. These recurrences
might be overlooked if the activities
weren't codified, forfeiting a chance
to reuse knowledge, instead of creating
it from scratch.
Understanding the customer's
business process is necessary but not
sufficient to the challenge of innovating
in services. As noted above,
the customer must interact with the
supplier at various points in the services
process. Thus, a second necessity
to business process mapping
is the idea of experience points.
These are points of contact between
customers and suppliers in the
exchange of services, where each
entity's respective processes must
interact with the other's in order to
fulfil the exchange. At these experience
points, customers select paths
from sets of choices constructed by
suppliers, and the exchanges branch
into different domains depending on
the choice made by the customer.
Even customers within the same
industry will not necessarily share
the same experience points.
Because each party brings its
unique knowledge and experience to
the exchange, and because part of
this knowledge is tacit in character,
no two experience points between
different suppliers and customers
are exactly alike. Thus, while there
are important economies to be
obtained from the recombination
and reuse of earlier elements, there
will always be some degree of difference
that must be accommodated in
the context.
x Why Now?
If we have gone this long without a
conception of innovation in services,
why do we require one now? The critical
enabling technologies in communications
and information technology
are a crucial part of the answer.
As these technologies have
advanced according to Moore's Law,
which predicts that the density of
semiconductor circuits will double
every 18 to 24 months, the ability to
codify and transmit knowledge ­ and
to reuse and recombine that knowledge
­ has grown exponentially. This
has resulted in sophisticated technologies
being embedded in a wide
variety of devices, and a robust network
of interconnections has fused
them in useful ways.
For example, today a car is a
highly sophisticated system that contains
dozens of chips that collectively
help power, steer, brake and monitor
the car while giving the user control
over the interior climate, audio,
visual displays and often telecommunications
portions of the vehicle.
More subtly, these advances have
enabled information that had previously
accompanied the production of
agricultural or manufactured products
to be separated from those artefacts.
Today, this information can move at a
much higher velocity from the end
user all the way back up the value
chain to the supplier of raw materials.
What is more, the information is
not consumed in the exchange, but
remains available for additional use
or reuse by others. This was not true
in earlier eras, when the consumption
of a product meant that others
could not consume it.
x An agenda for research on
innovation in services
In essence, innovation in services
boils down to improving workforce
productivity. It promises greater job
creation and higher living standards
in the west, and will bolster western
businesses competing with companies
in China, India and other parts of
the rapidly developing world.
Moreover, innovations can also
help to close the wide productivity
gap described in the work of Eric
Brynjolffson of MIT and Robert Gordon
of Northwestern respectively. In
their separate studies, each has
demonstrated that there is tremendous
variation in productivity
among different industries, much of
which can be explained by the ways
in which companies have (or have
not) incorporated new communications
and information technologies
into their businesses.
At this early stage, academic
research about innovation in services
is not well defined. (There is
even an active debate over the
proper definition of the term "services".)
Any useful understanding of
the opportunities and risks that are
unique to services innovation will
invariably involve business process
modelling, business models, systems
integration and design. More deeply,
questions of complexity in systems
design, cognitive processing of information,
and the role of codified and
tacit knowledge will also be
involved. The design of choice sets
and experience points in facilitating
interaction with customers will also
be a rich vein of inquiry.
While our understanding is at an
early stage, we can already anticipate
some initial academic insights about
innovation in services. Numerous
studies of innovation in products
show a life cycle pattern, with different
challenges confronted at different
stages of a technology's phase in the
cycle. It is likely to be the case that
this pattern exists in services, too.
For example, an unexplored area
of research opportunity is the termination
of services engagements.
We know very little about how these
projects terminate, and how posttermination
issues are addressed.
Yet companies should understand
the endgame as they consider the
choice of whether to enter into a services
arrangement.
Systems integration is another
area ripe with implications for services.
As the number of potentially
reusable bits of codified knowledge
expands, the wider the scope of
potential services that can be produced
from those reusable bits.
However, we also know that the
same increase in these knowledge
elements expands exponentially the
number of possible ways that these
elements can be combined. This
means that the gains in scope could
be outweighed (at least in theory) by
the even greater increase in complexity.
Academic research is needed
to weigh the costs, as well as the
benefits, of expanding the menu of
reusable knowledge.
Furthermore, there is the tradeoff
between standardisation and customisation:
when should a module
be reused, instead of employing a
custom-engineered piece of knowledge,
to serve a customer need?
The former will cost less to develop,
since it has already been created.
The latter will be more tightly connected
to the context of the customer's
problem. How should we
compare the two?
Another insight could come from
understanding sources of rigidity.
Academic studies into business
processes have noted the benefits of
complementary activities among
business process steps. But these
studies have also identified rigidities
that emerge from these same complementarities
that impair the ability
to adapt that business process to
new challenges. So, targeting a solution
in a customised way to serve a
current need might impose a hidden
cost in being less flexible to adapt to
a future need.
Services innovators will also
require a different set of measurements
to manage and advance their
efforts. Time and cost have unusual
characteristics in the services world,
with many services having a high
initial development cost and a very
low replication and distribution cost.
Most companies' current measures
of innovation ­ developed for a world
of products ­ do not take adequate
account of these features in the services
context.
This agenda may sound daunting
but the effort is certainly worth it. We
need to do a better job of understanding
the process of innovation in services.
Only through continued
innovation in what is now the supermajority
of economic activity in
advanced industrial economies can
we uncover ways to increase highvalue,
high-wage employment that
will support an abundant and rising
standard of living.
is learning to ride a bicycle ­ knowledge
that is difficult to transfer when
parties do not know each other. This
difficulty of transmission greatly
complicates the services exchange. It
limits the ability of each party to
fully comprehend the needs and abilities
of the other.
These constraints have been well
understood in international business
for decades, where the very cultural
and linguistic identities that facilitate
communication within a culture
become part of the challenge of managing
a transnational organisation.
Even in technical domains, however,
tacit knowledge is vital. Professional
associations, school ties,
convention gatherings and the like
provide face-to-face experiences that
help to transfer tacit knowledge.
Consider the challenge of transferring
a technology from one company
to another. Some aspects of
it may be well understood, due
to patent publications, technical
blueprints, bills of material and
product manuals.
But putting the technology into
practice invariably requires more
knowledge than has been written
down. The patent may describe how
a product or process works, but anyone
who has licensed a patent will
testify that the know-how required
to make use of the technology is not
included. Many licensing transactions,
therefore, explicitly include
consulting services to supply this
know-how.
These knowledge transfer challenges
are heightened by increasing
complexity in the systems in which
the knowledge is exchanged. In complex
systems, which are defined as
being composed of two or more technologies,
processes must be combined
in order to deliver value. The
possible ways in which these technologies
might work together grows
exponentially as the number of additional
technologies within the overall
system increases. Companies
struggle early on in the technology
life cycle to identify a feasible system
out of this complexity that does
something valuable.
While the problem of complexity is
not unique to services, the lack of
products heightens the challenge of
the exchange. Today, services
exchanges involve many complex
combinations of both codified and
tacit knowledge. Thus, identifying
feasible solutions ("systems integration")
that actually work before others
have done so is an increasingly
important source of economic value.
To see this, think of Apple's iPod.
None of its component parts are particularly
new but the speed and ease
8
In essence, innovation in services
boils down to improving workforce
productivity. It promises greater job
creation and higher living standards