Dimensions of
the Learning
Organization
by Olivier Serrat
Background
If organizational learning is still seeking a theory, there can
be no (and perhaps cannot be) agreement on the dimensions
of the learning organization. Even if the dimensions were
understood, the connection between learning (or lack thereof)
and performance remains unclear.1 However, regardless
of the disputed state of the art, a multilevel, practical but
necessarily exploratory and simple framework of common
and individual variables associated with learning and
change follows. Here as elsewhere, experimentation has an
important role to play. Individual and collective learning are
not about finding out what others already know, even if that is a useful first stage—it is
about solving problems2 by doing, reflecting,
connecting, and testing until a solution forms
part of organizational life. There is no stock
answer nor is there a single best approach.3
Figure 1 suggests concepts that can be used
individually or in association to reflect on the
overall system.
1 Most organizations know little about where they lose knowledge, so the costs of lost knowledge are largely
hidden. As a result, there is no clear ownership of the problem and little value is given to knowledge-sharing
activities.
2 Some streams of open systems theory reject problem solving as unproductive, instead preferring to work
on desirable futures and necessary actions (only “solving problems” as they become barriers to a goal). The
difference in the outlooks is significant.
3 A parallel can be found in the disparity of systems models for organizational design. Those used often in the
last 20–30 years have included McKinsey’s 7-S Model, Galbraith’s Star Model, Weisbord’s Six Box Model, Nadler
and Tushman’s Congruence Model, and Burke-Litwin’s Causal Model. Each of these shines a particular light on
an organizational system, in the way perhaps that astronomers standing on different planets would examine
different configurations of the universe. No one perspective is correct. The choice of model depends also on
how complex its user wishes it to be. In recent years, less inward-looking (closed system) models have been
developed.