Vous êtes sur la page 1sur 0

The future

of product development
roduct development is facing a fundamental challenge. Most com-
panies are under pressure to bring new products to market more and
more quickly, and companies in many sectors must work harder to ensure
that they address the needs of ever-narrower customer segments. These pres-
sures are particularly acute in fast-paced markets, such as those for high-
tech, medical, and consumer goods, and in competitive markets for complex
products that require large investments and long development times, such as
automobiles and airplanes. CEOs wonder how their product-development
organizations can meet the new challenges.
Unfortunately, at the very moment when companies need to make better
products more efficiently, previous performance innovations in product
development have hit a plateau. Over the past 15 years, most companies
have adopted standard product-development processes, with disciplined
time lines, strict design reviews, gates to decision making, and cross-
functional development teams. While these changes have made the devel-
opment of new products much more efficient, further improvements are
Richard Holman, Hans-Werner Kaas,
and David Keeling
By focusing on better information management rather than
processes, companies can dramatically boost their product-development
performance.
P
28
Q3_ProdDev_v5 6/26/03 9:58 AM Page 28
GORDON WIEBE
Q3_ProdDev_v5 6/26/03 9:58 AM Page 29
returning smaller gains. What is needed now is a way of raising product
development to a new level.
Some companies have done exactly that: they have reduced the time needed
to launch new products while dramatically raising their sales of new prod-
ucts and their market share (Exhibit 1). The key to the new approach is an
entirely different way of making product-development decisions. By improv-
ing the quality, timing, and synthesis of product and process information
throughout the development cycle, such companies have turned a linear,
sequential process into a exible one that reacts to information continually
rather than at intervals and in batches. They keep their product options open
longer than most of their rivals and can act on new information about cus-
tomers, markets, suppliers, and production capabilities later in the develop-
ment process. Decisions are made when all options have been understood
fully, not at rigidly scheduled meetings that open the gate to the next stage.
As a result, such companies make better decisions at each step and therefore
reduce the bottlenecks, information gaps, rework cycles, and wasted effort
that can add so much time and expense to the development of new products.
They create more successful products more effectively, in less time and with
fewer resources.
The new approach does have its challenges. Without skilled leadership and
some new organizational capabilities, companies may lose control of a disci-
plined process that has improved their performance signicantly. They will
have to manage their resources more exibly and monitor their teams ability
to generate the right information and to use it effectively. Yet for many com-
30 THE McKI NSEY QUARTERLY 2003 NUMBER 3
E X H I B I T 1
Better, faster, cheaper
Disguised example of automated medical device
1
Per every 1,000 hours of operation.
2
Forecast based on comparison of competing products, customer surveys, and conjoint analysis.
Quality
23
33
2
Market share,
percent
10
5
Unscheduled
maintenance calls
1
Time
50%
Time to market,
months
36
22
Cost
56
46
Product cost,
$ thousand
20%
+40% 40%
Before: fixed, process-
based approach
After: dynamic, information-
based approach
Q3_ProdDev_v5 6/26/03 9:58 AM Page 30
panies, the opportunity to change products significantly later in the develop-
ment cycle while also developing better ones more quickly is just too com-
pelling to pass up.
The information-driven approach
The changes currently required in product development resemble the lean-
manufacturing techniques that have transformed mass-production lines.
Besides optimizing the efficiency of each station on the factory oor, lean
procedures create a exible, efficient work ow intended to meet customer
demand just in time. To minimize waste and inventory and to optimize the
efficiency of the line, parts are fed
into the process as they are needed.
By contrast, the current approach to
developing new products resembles
the traditional mass-production line:
companies follow a xed sequence of
steps, moving from market research
to product concept, design specication, prototype testing, and so on. Like
a production line, this process can be improved signicantly by ensuring
a continual ow of work. To achieve the next step change in performance,
companies must therefore improve the efficiency of the entire process, from
generating ideas to launching products. Adopting lean techniques has helped
companies increase their manufacturing efficiency by 30 percent or more.
We believe they can get a similar boost in the performance of their product-
development teams by adopting an information-based approach. In our
experience, during a typical development cycle as much as one-third of the
time is spent doing unnecessary work or waiting for decisions or informa-
tion.
1
By managing information ows rather than only process steps, compa-
nies can eliminate these sources of wasted time.
In practice, the new approach requires some fundamental changes in product
development (Exhibit 2, on the next page). Rather than rigidly adhering to a
standard sequence of activities, the work of teams developing new products
is organized to ensure a continual ow of high-quality information into the
development process. Companies solve problems and synthesize new infor-
mation continually instead of merely collecting bits of information from vari-
ous functions and compiling the results just before gate meetings. By using a
31 T HE F UT URE OF PRODUCT DE VE LOPME NT
1
These figures are consistent with research reported in Colin Mynott, Lean Product Development,
Northampton (Great Britain): Westfield Publishing, 2000. Mynott indicates that in one study, 25 per-
cent of the teams time was spent on value-adding necessary work (doing the right things correctly at
the right time), 10 percent on necessary work that didnt add value (such as traveling and writing
reports), 30 percent on rework (fixing errors and redesigning products), 25 percent on activities other
than work (vacations and waiting), and 10 percent on unneeded work (attending meetings and writing
reports that no one read).
Product development must now
be changed in ways resembling
the lean-manufacturing techniques
that transformed mass production
Q3_ProdDev_v5 6/26/03 9:58 AM Page 31
more exible, information-driven methodology, teams reach better solutions
more quickly.
Organize around information flows
The rst step for any team developing a new product is to determine which
attributes are critical to its success. In the information-based approach, the
team goes a step further by identifying the information required to make
each key decision along the way. Companies then recongure the develop-
ment process around these needs, ensuring that the right information is
gathered at the right time and then ows to the right people.
In this way, the sequence and concurrency of specic tasks vary for each
new-product team. If costs, say, are critical, the product team denes a work
ow built around determining and meeting cost targets instead of allowing
costs to be an uncontrolled output. As a result, the development process may
change: key suppliers and operations managers might, for example, produce
cost estimates for design features at an unusually early stage rather than try
to minimize costs only
after the development
team had decided on a
concept.
Each task is always
optimized to promote
the end goal, not the
efficiency of the task
itself. Consider, for
example, the testing of
prototypes. Under the
old approach, the pro-
totype team tries to
make the most repre-
sentative model of the
product. To do so, how-
ever, the team needs
detailed product speci-
cations, which are
available only late in
the development cycle,
and a lot of time to
construct the model. In
the information-based
approach, the rst step
32 THE McKI NSEY QUARTERLY 2003 NUMBER 3
E X H I B I T 2
A better way to develop products
. . . an information-based approach creates flexibility and improves productivity
Develop,
verify
design
Validate,
transfer
design
Select
concept
Ramp up
production,
launch
Prove
feasibility
1 2 3 4 5
Product
launch
Product
concept
Product features
Product quality
Product cost
Example of 3 work flows
Information
synthesis
Design
iteration
Management
review
Information x Decision gate
In contrast to the conventional, process-based (phase-gate) approach . . .
Q3_ProdDev_v5 6/26/03 9:58 AM Page 32
is to decide which information gaps must be lled during the testing of the
prototype and when that information will be most useful. Sometimes an ear-
lier, simpler model might yield it without additional workthereby elimi-
nating one source of delays.
Organizing product development in this way might seem simple and even
obvious, but it is a radical departure from what most companies do today.
Consider the experience of a global company that makes medical devices.
Under its old process, teams focused mostly on dening the features and
functions of new products. Market research was secondary, and no input
from operations people and suppliers was sought until prototypes had been
built. The cost and reliability of the products were outcomes of the chosen
design though not key considerations during its creation. Subteams only
33 T HE F UT URE OF PRODUCT DE VE LOPME NT
E X H I B I T 3
A medical-equipment company transforms its product-development process
1
Disguised example of automated-medical-equipment company.
Reliability
work flow
Customer needs/
design features
work flow
Cost work flow
Flexible information-based approach
1
Internet-based
survey conduct-
ed to confirm
unmet needs
Marketing
insights com-
municated to
subteams
Alternatives ex-
plored, expanded
by marketing,
R&D, operations
Business forecast
from customer
evaluations of
e-prototypes
Features, cost /
quality trade-offs
considered daily
Market data
used to set
quality, reliability
targets
Target feature
set established
Cost of features
extrapolated
from historical
data
Targets for cost,
quality com-
municated to
subteams
Feasibility of
alternatives
quantified by
life cycle cost
Options for
features vs. cost
shared with
customers
Manufactur-
ability evaluated
by suppliers,
operations
Risk of not
meeting targets
quantified
Features prior-
itized with
repeated Web-
based conjoint
analysis
6 months
2 months
Process-based (phase-gate) approach
1
1 2
Delay Delay Delay
Initial market
research
conducted
Alternatives
defined by
R&D
Market
research
refined
Product
criteria,
targets set
Concept
redone because
of high costs
Concept
feasibility
confirmed
Gate 1: project
authorization
Gate 2: project
feasibility
Continuous synthesis of findings Management review Design iteration
Concept
feasibility
confirmed
Q3_ProdDev_v5 6/26/03 9:58 AM Page 33
loosely coordinated their development activities, and six months or even
more were needed merely to settle on a concept for a new product and to
choose its features.
A team participating in a project using the information-based approach
started by creating three simultaneous work ows: one for the needs of
customers and the design features of the product, a second for
its cost, and a third for its reliability (Exhibit 3, on the previ-
ous page). The three subteams communicated daily and
assessed all of their ndings every week. By orga-
nizing around the ow of information, the team
looked at market research through a new lens
and made an important discovery: customers
care most about the cost and reliability of prod-
ucts, not their throughput performance and other
advanced features, as the team had previously assumed.
Still more important, the new approach inspired the team
to react to this new information in a new way. Customer ser-
vice and operations were consulted early on to analyze what
had made previous models break down and to estimate the trade-offs
between design features and costs. Instead of having the marketing depart-
ment undertake the research and throw the results over the wall to R&D,
the product team and the marketing department conferred frequently. They
could therefore quickly tweak the design of the product and then determine
the implications for its cost and reliability as well as the reactions of poten-
tial customers. In this way, the team settled on the products concept and
features in two months instead of the six previously needed. In the end, it
designed a new machine that cost 20 percent less to manufacture than previ-
ous models, took 40 percent less time to develop, and had 50 percent fewer
breakdowns. Since launch, the company has nearly doubled its market share
in that product category.
Solve problems continually and share the results
Instead of using a linear approach to collect information, make a decision,
and then base other decisions on the rst one, information-based teams
solve problems continually and combine their ndings frequently. Like the
medical-device companys team, they work in a way that allows them to
converge on the best solution. This style of work resembles the daily-build
method many software companies use: the code produced by individual pro-
grammers is compiled every day so that project leaders can test it for bugs
and functions. Problems are reported immediately, and the team knows, on
a daily basis, how close it is to its ultimate goal.
34 THE McKI NSEY QUARTERLY 2003 NUMBER 3
Q3_ProdDev_v5 6/26/03 9:58 AM Page 34
By gaining the ability to delay the point when product designs must be
frozen, information-based teams keep their options open longer and can
respond to changes in the market at later stages of the development process.
Senior management imposes fewer constraints on these teams than on their
conventional counterparts, so they can consider a broader set of solutions;
concepts may take longer to develop, but the company saves time later on.
Toyota Motor, which uses some aspects of the information-based approach,
routinely develops new car models in at least 20 percent less time and with
30 percent fewer resources than its competitors, even though it evaluates
more design alternatives and keeps its options open longer.
2
The reason is
that its developers understand design trade-offs better than conventional
developers do, avoid costly rework, and respond more quickly to problems
arising later in the process. Moreover, by amassing deeper knowledge about
any one project, Toyotas teams can often develop other, related products
more quickly.
Effective teams that can solve problems, synthesize ndings, and orchestrate
information ows are vital to the new process. Subteams meet daily to
review their progress and to ensure that new results are fed
immediately to different branches of the team and, where
necessary, to the broader business. Team leaders bring
together people with the right mix of skills to solve
problems as soon as they arise and eliminate barri-
ers to progress. Meetings with senior executives
are used not to make routine decisions but to
review the progress of development projects and
to discuss product strategy. Teams can slow down
or speed up their work as needed and adjust the
process in midstream as new information becomes available.
Companies using the new approach sometimes appoint senior man-
agers from marketing, R&D, and operations to lead cross-functional
teams in hopes of getting the three key disciplines to work together from
the start. Typically, one of them is rst among equals and ultimately leads the
project.
Product development based on this approach is relatively exible but doesnt
want for discipline. Companies set their performance targets and conduct
management-review meetings, much as they do in the conventional process.
The difference is that they reach decision gates when criteria are met, not
when a given amount of time has passed; a product concept might, for
example, be made nal when certain technical functions have been validated
35 T HE F UT URE OF PRODUCT DE VE LOPME NT
2
For more on Toyotas set-based design approach, see Durward K. Sobek II, Allen C. Ward, and Jeffrey
K. Liker, Toyotas principles of set-based concurrent engineering, MIT Sloan Management Review,
Winter 1999, Volume 40, Number 2, pp. 6783.
Q3_ProdDev_v5 6/26/03 9:58 AM Page 35
and the products cost falls within a particular range. Project leaders may
estimate in advance how long it might take to realize these goals, but since
the team strives to do so as quickly as possible, waiting time and the infor-
mation gaps created by premature decisions are eliminated (Exhibit 4). The
result is better decision making, more efficient development cycles, and
fresher products that better t the markets needs.
The challenge
Moving from todays less exible, process-oriented approach to a dynamic
and exible one demands new organizational capabilities and skills. Without
the right project leaders, the ability to allocate resources as needed, and new
ways of measuring the performance of project teams, companies wont cap-
ture the potential efficiency improvements of the new approach and, worse,
may nd that it lacks discipline and forfeits many of the past decades gains.
Several companies have found ways of combining the two approaches. They
have learned a few lessons along the way.
Let leaders lead
Strong project leadersthe key to any good product-development organiza-
tionare even more important in the information-based approach. In con-
ventional product development, process discipline keeps the organization
in check and thus limits a good managers ability to make the process more
efficient and effective, though it also controls the potential negative impact
of a poor project manager. With the new approach, project leaders drive per-
36 THE McKI NSEY QUARTERLY 2003 NUMBER 3
E X H I B I T 4
Just-in-time decision making
1
Sufficient information to confirm or predict market success, product performance, and technical and business risks.
Process-based approach Information-based approach
Information
threshold
1
I
n
f
o
r
m
a
t
i
o
n
Time
Scenario A
Scenario B
Scenario C
Just-in-time decisionmade when
all necessary information is available
Information
threshold
1
Fixed decision
gate
I
n
f
o
r
m
a
t
i
o
n
Time
Scenario A
Scenario B
Time wasteddecision
could be made sooner
Information gap
increased risk of re-
work downstream
and disappointing
market results
Q3_ProdDev_v5 6/26/03 9:58 AM Page 36
formance by focusing on the highest-value activities, skipping unnecessary
ones, and shaping the team and the work ow in response to new informa-
tion. Successful project leaders must therefore be both inspired people man-
agers and skilled problem solvers. They must bring the right individuals and
information together to develop the best solutions for problems, coach sub-
teams to perform at a higher level, and have a working knowledge of all
areas of their projects, including the technical side, marketing, operations,
and supply chain management. Above all, senior executives must trust these
leaders to make sound, fact-based decisions about the direction of their proj-
ects without always seeking input from above.
Do such people exist? The good news is that they do, but only the very best
project leaders can now perform at this level. In our experience, the potential
of the best leaders is stied by the current inexible approach. Giving them
enough exibility and authority to make decisions will unleash their poten-
tial and raise the performance of their teams a few notches. Meanwhile, most
project managers will have to upgrade some combination of their leadership
ability, their problem-solving skills, or their cross-functional expertise
hardly surprising, since many of them are engineers promoted to manage-
ment without training or even, in some cases, natural aptitude. Organizations
now have the task of creating processes to spot project leaders with strong
potential and to develop their skills.
To address the shortage of project leaders, one high-tech company with sales
of more than $4 billion identied the 20 most promising ones it had and
then put them through a training workshop (later rolled out to all project
managers) on problem solving, coaching techniques, and leadership skills.
The company also created an apprenticeship program. Average project man-
agers became deputies to one of the top leaders for a year; after active men-
toring and on-the-job training, they were sent back to manage new teams
on their own. Unusually, the company also offered large bonuses to project
leaders and team members if their products met or exceeded specic targets,
such as a certain level of sales after six months.
Allocate resources dynamically
The ability to allocate resources exibly to each new-product team is vital
as well. Indeed, exibility is the key to information-based product develop-
ment: the testing of prototypes might take place sooner than planned, for
example, or certain developers might be needed longer. To accommodate the
uncertainty, it is necessary to be able to allocate resources dynamically.
The rst step is to organize product development around project or product
teams, not functional areas. New-product teams that have to beg resources
37 T HE F UT URE OF PRODUCT DE VE LOPME NT
Q3_ProdDev_v5 6/26/03 9:58 AM Page 37
from functional silos inevitably lack the required exibility and focus.
Second, most product-development organizations will nd that they need to
hire or train staff members capable of wearing more than one hat at a time
for instance, developers who can play a variety of design roles, engineers
who have training in several disciplines and can undertake both series and
advanced engineering, and operations people who know how to build proto-
types and manage suppliers. In general, the most valuable employees have
experience in several technical functions as well as some business training.
Finally, companies must build up their reserves for dealing with unexpected
demands. They can do so by making sure that some key multitalented
employees are always at work on
projects with exible time con-
straints (such as long-term ones) so
that these people can be moved to
higher-priority projects in emergen-
cies. (The same applies to market-
research, prototype-testing, and
operations units.) Companies can also create outsourcing options for part of
their development efforts (such as market research) or rank projects in order
of priority and identify those that can surrender resources when necessary.
We have found that the productivity gains generated by the information-
based approach free up resources, thereby making it easier to create reserves
and develop multitalented employees. Like a lean-manufacturing environ-
ment, in which multitalented workers are the key, this approach improves
employee morale.
Measure performance
Most companies track the costs of their projects, their ability to complete
projects on time, and the resulting sales to judge how good they are at devel-
oping new products. Companies that take the information-based approach
also need metrics for the quality of the information they gather and the
efficiency with which they process it. In short, teams must be sure they are
gathering the best information possible and spending the right amount of
time analyzing it.
Just as wastage is a key metric in lean manufacturing, information wastage
is important in gauging the efficiency of information-driven product devel-
opment. This metric should show the amount of time development teams
wastefor example, by undertaking unsuccessful projects, waiting for
information, doing avoidable rework, or taking extra time to ramp up pro-
duction to the specied quality and scale.
38 THE McKI NSEY QUARTERLY 2003 NUMBER 3
Product-development units must
hire or train staff members who
can wear more than one hat
Q3_ProdDev_v5 6/26/03 9:58 AM Page 38
Measuring the quality of information processing is equally important.
How, for example, did initial market forecasts or cost estimates compare
with actual results? Did information gaps become apparent during the
process and, if so, why? Were customer complaints fully addressed while
the concept was under development and, if not, why not?
For the past couple of decades, product developers have improved their
performance largely by making the process more disciplined and rigorous.
Such improvements can no longer satisfy the increasing demand for better
products launched more frequently and aimed at ever-narrower customer
segments. Companies must now turn their attention to building a more
nimble and exible product-development organization. To do so, they will
have to focus on information ows within the development team, coordinate
the efforts of dozens of subteams more successfully, learn to solve problems
and synthesize results on an ongoing basis, and give more decision-making
authority to project leaders.
Those companies that meet the challenge will have an undeniable edge over
their competitors because they will succeed not only in responding to cus-
tomer and market changes right up to a fairly late stage in the process but
also in bringing their products to market more quickly and efficiently.
Rick Holman is a consultant in McKinseys New Jersey office; Hans-Werner Kaas is a principal in
the Detroit office; David Keeling is a principal in the Chicago office. Copyright 2003 McKinsey
& Company. All rights reserved.
39 T HE F UT URE OF PRODUCT DE VE LOPME NT
Q3_ProdDev_v5 6/26/03 9:58 AM Page 39

Vous aimerez peut-être aussi