17. How to combine exploration and
exploitation?
When highly systemic, the
system has many densely connected elements, with narrow tolerances in
interfaces. Here, deviations in elements, for the sake of experimentation,
would jeopardize the integrity of the system. An example is a refinery. Then,
there is a high cost, while the promise of results is highly uncertain.
In stand-alone systems,
by contrast, elements are highly autonomous in their exploitation. If, in
addition, exploitation itself already requires highly differentiated products
for different customers, there is more scope to combine exploitation and
exploration. An example is a consultancy firm.
An intermediate case is
that of a modular system, where elements are many and mutually connected but
self-contained and replaceable, provided that they satisfy the constraints in
the system.
In systemic
arhitectures, there are several ways of dealing with the problem of combining
exploration and exploitation.
Logically,
one option is to separate them in time, focusing first on exploration, then on
exploitation, and back again. That goes back to the old idea of some ‘life
cycle’. Life cycles of products and technologies have shortened, and different
products in a portfolio will be in different stages of their cycles. In case of
products in the same stage of a cycle, it would require something like a
pulsating organization, with alternating large and small cognitive distance,
order and chaos. It is hard to do this other than by engaging and disengaging
people with different skills and mentalities, which is problematic. And then,
with such discontinuity of staff, how would novelty be rooted in experience
with what went on before?
A second
possibility is to separate them in place. One classic form is the one indicated
earlier, in a split between production and development, with the problem of
combining the two in one structure and culture. To make exploitation receptive
to exploration, and to root exploration in exploitation, one may rotate staff
between the two functions, and hire people who have the ability to cope with
that.
It would
help to engage generalists, i.e. people with a wide scope of knowledge and
capability, and a large absorptive capacity, so that they can maintain
understanding between them even under changing logics or frameworks, and can
quickly shift between those.
Another
option is to specialize in one or the other, in different organizations, with a
linkage between them in an alliance or other form of collaboration. But here
also there would need to be some interchange or rotation of staff between the
two, or versatile bounday spanners to mediate between the two.
There would
also need to be investment in mutual understanding, cognitively and morally
(concerning modes of conduct). Such investments are largely relation-specific,
in the sense discussed in item 7, which would require a certain stability and
duration of the relationship.
A classic
example is that of small biotech firms that engage in the exploration of novel
active substances, diagnostics or production technologies, whose results are
hived over to large pharmaceutical companies with the stability, size and
resources needed for exploitation, in carrying a new medicine through the
lengthy process of regulatory approval, and large scale production, promotion
and delivery in markets.
The need to
engage in specific investments for mutual understanding and cooperation raises
issues of dependence and power. In an asymmetric relationship between small
firms for exploration and large firms for exploitation the smaller firm would
have to beware of the risks of one-sided dependence. This may be, and often is,
compensated by the small firm being uniquely capable in exploration, to provide
countervailing power. I will expand on problems of collaboration in later items
of this blog.
[1] Langlois,
R.N. 1998. ‘Personal capitalism as charismatic authority: The organizational
economics of a Weberian concept’, Industrial and Corporate Change, 7/1,
195-214.