New product development 327
beta site test, voice of the customer systems,
concept test and customer site visits. Moreover,
Tzokas (2000) reported that better-performing
companies make more frequent use of all the
market research tools studied by the research,
including customer site visits, depth interviews
with customers, beta site testing, field test
markets, creativity sessions for idea generation,
concept tests, lead user analysis, concept engi-
neering, conjoint analysis focus group inter-
views, information acceleration and test market
simulations.
The models can provide a useful frame-
work on which to build a complete picture of
the development, particularly with regard to
the potential advantage of the product viewed
from the customers’ perspectives. They have
been evolving for years and now the BAH
model of the 1980s may still serve as the core of
what is the Stage–Gate™model developed by
Cooper, but even that has evolved further into
what is referred to as ‘third generation’ pro-
cesses. These so-called third generation process
models, which are characterized by ‘fuzzy and
flexible gates’, have evolved due to a number of
limitations on the earlier versions. These are
described briefly below.
Related strands of development
These related strands of development refer to
marketing, technical (design) and production
tasks or decisions that occur as the process
unwinds. Each strand of development gives
rise to problems and opportunities within the
other two. For example, if, at the product
development stage, production people have a
problem which pushes production costs up,
this could affect market potential. The market-
ing and technical assumptions need to be
reworked in the light of this new information.
A new design may be considered, or a new
approach to the marketplace may be attempted.
Whatever the nature of the final solution, it has
to be based on the interplay of technical,
marketing and manufacturing development
issues, meaning that product development
activity is iterative, not only between stages,
but also within stages. Indeed, since many of
the linear models do not adequately commu-
nicate the horizontal dimensions of the NPD
process, several ‘new generation’ models have
been developed. Although Cooper’s Stage–
Gate™model is one of these (see Figure 12.2),
others include the ‘fifth generation innovation
process’ of Rothwell (1994) and the ‘blocked’
model of Saren (1994). These new generations
of process models touch on the idea of ‘parallel
processing’, which acknowledges the iterations
between and within stages, categorizing them
along functional configurations. The idea of
parallel processing is prescriptive: it advises
that major functions should be involved from
the early stages of the NPD process to its
conclusion. This, it is claimed, allows problems
to be detected and solved much earlier than in
the classic task-by-task, function-by function
models. In turn, the entire process is much
speedier, which is now recognized to be an
important element in new product success.
Hence, these are presented as solutions to the
dilemma raised earlier, that the execution of the
activities is required but not at the expense of
speed in the process. Other related techniques
which allow firms to accelerate their NPD
include quality function deployment (QFD),
increasing rewards for R&D performance, rely-
ing on external sources of technology and
improving the interface among the relevant
functional areas (Calantone et al., 1997). So
important is the last of these methods that we
revisit the substantial amount of research it has
attracted towards the end of this chapter.Idiosyncrasy
The NPD process is idiosyncratic to each
individual firm and to the new product project
in question. Its shape and sequence depends on
the type of new product being developed and
its relationship with the firm’s current activities
(Cooper, 1988; Johne and Snelson, 1988). In
addition to the need to adapt the process to
individual instances, in real situations there is