Measuring Organizational Learning©
Arthur M. Schneiderman
Twenty years ago, Peter Senge and I were co-workers in the MIT System Dynamics Group; he was working on his PhD while I was completing an MBA and subsequently serving as a Research Associate working on the System Dynamics national model. Our professional relationship resumed when I went to Analog Devices where my boss, Ray Stata, had become one of Peter's patrons in his work on organizational learning. So when my interest in measurement led me to the question of measuring organizational learning, Peter was the obvious person to contact. Although he readily acknowledged that measurement was a growing issue in the area of organizational learning, he observed that to his knowledge little progress had been made. This was compounded by the absence even of an agreed upon definition of organizational learning. So, throwing caution to the wind, here's my obviously biased view on the subject.
I would like to offer the following definition:
Let me explain. The process view sees the organization as a collection of interconnected processes. In general, these processes are not being executed at their ultimate potential; there are many gaps, even in those core processes that can affect competitive position. Organizations try to close the gaps that are most important from their stakeholders' perspective. But, the improvement methodology that they use determines the maximum rate at which this can be accomplished. If their improvement process is mature and held constant, than the learning that goes on within the organization can be called "process learning."
However, what if the organization decides to pursue a new improvement paradigm that has the potential of accelerating their rate of process learning? I call the acquisition, application and mastery of this new capability "organizational learning." Organizational learning does not in itself lead to any changes in an organization. What it does is enable faster process leaning. So like the incremental improvement/redesign tide that occurs at the process level, the organizational/process learning tide occurs for the entire system of processes. Organizations continuously move between periods when they are learning and periods when they are applying their newly acquired knowledge. If I understand Chris Argyris' distinction between single- and double-loop learning correctly, then process learning is an example of the former, and organizational learning of the latter as I've defined them.
Perhaps you can guess where I'm going. We have a measure of individual process learning in the form of the improvement half-life. We can in principle create an organization-specific version of my half-life/complexity matrix by populating the cells with the appropriate organization-wide averages. If we wanted, we could collapse this matrix into a single number by simple or weighted averaging. Either the matrix or the grand average would then be a metric for process learning.
In practice, this metric decreases over time as it approaches the capability of the process improvement methodology that's being used. This decline in the process learning metric is evidence of organizational learning as defined above. The rate of decline is initially high but approaches zero as the methodology is diffused and mastered by the entire organization.
We can put this in mathematical terms (for individual processes or for the organization as a whole:)
The process learning metric, Lp is identically equal to the process improvement half-life, while the organizational learning metric, Lo is the rate of change of the process improvement half-life at constant process complexity (or complexity mix, as appropriate). The negative sign is there to assure that Lo is positive if the half-life is declining (note that this is a change from my previously published version of these equations).
It is very tempting to go one step further and to calculate the half-life associated with Lo. This single metric characterizes the organizations ability to learn. If valid, it would represent the amount of time required for it to internalize 50% of a new improvement paradigm. Organizations who repeatedly learn new paradigms faster than their competitors should achieve a significant advantage, assuming they choose their paradigms judiciously. But I'll resist the temptation for now, since it is likely that this organizational learning half-life is measured in decades rather than months or years, and few Western organizations are sufficiently stable long enough for it to be of much predictive value.
So until a better metric comes along, I'll think of the half-life as a metric of process learning and the rate of change of an aggregate process half-life as a metric of organizational learning.
Last modified: August 13, 2006