Paul's concerns about "no one looking out for the whole" are valid in
the experience of many of us. More dangerous than that by far is
that each may compete against each other, possibly in covert ways,
and actually be more or less intentionally destroying the whole.
However, I do want to offer another possible interpretation and a
design possibility for a move which will distribute the function.
I'm using "distribute" to point out the inferences already in the
choice of language used by Paul - namely "fragmenting".
It is posisible that the strategic design accountability will be more
effective when more distributed. Even without a single person
concerned for the whole. Surely there is still some other person or
group who is concerned for the whole. The possible advantage arises
from the condition of the space of possibility being too large to
fully explore and too complex to evaluate optimimums. That means,
that any single choice can easily get stuck in an apparently good
strategy which is merely a local peak which is also fairly low
compared to others not discovered. Distribution can support the
exploration of more of the space of possibility.
The other design consideration is that strategy be a matter of the
intelligence, awareness, thinking of the total organisation and that
the strategic leaders are accountable for gathering, interpreting and
providing feedback to the organisation about strategy rather than
taking an expert position on somehow independently deciding strategy.
If the expert position is being taken, then the split will still have
the added exploration possibility but miss the largest potential. If
the position is one related to the distributed intelligence of the
organisation, then the dangers are minimal and the design may even be
optimal.
The concerns we have are generally founded in unexamined bureacratic,
hierarchical and linear views of how organisations work. These are
reflected in the location of individuals as a most important factor
rather than, from a complex adaptive system (or intelligent system)
point of view, merely incidental.
Instead of judging and assessing "what's best", how about working to
make whatever the structure is fit the purpose or empowering the
existing structures whatever they are (if you have nothing to say
about their design).
-- Michael McMaster Michael@kbddean.demon.co.uk