Organizational navel-gazing

Quoting Ionesco seems to be a good start, because, frankly, I don’t have answers. But I think it is a topic worth looking at.

What is the topic? The area you might call organizational architecture. And the question is: To which extent should an organization be occupied with itself? And to which extent should an organization interact with other parts of the company? Especially a support function? Or put it to a more provocative statement: How much are we occupied with organizational navel gazing?

What is the purpose of an organization structure in a company? Not a new topic, so we are using Wikipedia to summarize the key points. The theory of network analysis and complete graphs gives us a first grip on what to expect. Then finally we look at the data from a company on the highest organizational level.

“An?organizational structure?defines how activities such as?task allocation, coordination, and supervision are directed toward the achievement of organizational aims.” “Organizations need to be efficient, flexible, innovative and caring in order to achieve a sustainable competitive advantage.” (Organizational Structure, Wikipedia) In my words an organization optimizes cooperation & collaboration for value creation.

Henry Mintzberg (a guest of ours in the 4D’s sessions) defines:

And “Support staff?(helping other members of organization to perform their function)” (Organizational Structure, Wikipedia). Thus, a Support Function is specifically defined by the value that it is creating for other entities of the company.? Obviously if a Support Function would be 100% occupied with itself, it has lost its purpose. Thus, it is a relevant question to ask about the sweet spot of interaction with others and self-interaction.

What can we expect in reality? The example of 100% self-occupation is helping to understand more complex scenarios from its simplistic extreme. Obviously, it is not very realistic.

We approximate reality from another unrealistic model: The complete network. ?“In the?mathematical?field of?graph theory, a?complete graph?is a?simple?undirected graph?in which every pair of distinct?vertices?is connected by a unique?edge.” (Complete graph, Wikipedia)

For simplicity we look only at the interaction within the company, the picture can with the same logic be extent to customers and suppliers and a whole ecosystem. There you have for example the very relevant question, to which extent a sales organization should be occupied with the customer and to what extent with internal interaction.

Please note that in the following, we are not counting the interaction (edges), but counting and calculating the proportions of nodes, with which interaction has happened, in the organization versus outside. We look at a chosen organization and it interaction within and with employees in other entities. Then we count the employees interacting with each other in the organization (self-interaction), and we count the employees outside of the organizations (interaction with others) and look at the proportions.

Applying the concept of a complete graph to a company, it would mean everyone is interacting with everyone. Unrealistic, yes, but simple: For a fixed size of a company, the proportion of nodes of interaction with others and self-interaction is a linear function of the organizational size within the company. If my “organization” consists only of myself, then we are at 0% self-interaction (all interaction is with outside). If my “organization” is the full company than every interaction is self-interaction of the organization, self-interaction reaches 100% (the interaction with others is 0%).

?

?

?

What kind of company is this? A company, where everyone interacts with everyone.

Now remember what was said earlier about the purpose of an organizational structure. An organizational structure is fostering collaboration between those that should collaborate and should create boundaries between those that have nothing to do with each other, not distracting.

If we have a company, where everyone interacts with everyone, this means, there is not organization in the company.

Still it is an unrealistic model for the 100.000 employees of a large company, but we can derive some consequences:

The proportions of self-interaction and interaction with others are a function of organizational size. This is formalizing our intuition: if our unit is larger, of course we expect more interaction within the unit.

Any organization, where the self-interaction is below the line of the complete network (no organization) is obvious dysfunctional. It would mean that you are interacting with your own organization less than with the outside.

An organization is also dysfunctional, if only occupied with self-interaction, which means approaching 100% self-interaction.

The sweet spot, respectively the area of optimal proportions of self-interaction and interaction with others is somewhere between the line of the complete network and the line of total self-interaction. You see a drawing of that area in the figure below. Please note that it is schematic, nothing is known about the concrete functional relation apart form the two end points.

?

So much of theory, how does it look like in reality??

For this we have looked at the organizations on the highest organization layer (Business Areas/Market Areas/Group Functions, marked as blue dots)

?

Indeed, we see that the organizations are above the line of complete networks, we see the increase of the self-interaction proportion with size and we can imagine them in the organizational sweet spot area schematically.

Obviously, there are no organizations above the 30.000 employees, thus we can zoom in to

?

And since there are only the two organizational units close to 30.000 employees even more to

To fit the data better, a quadratic regression was used.

Is there any practical use of this?

In my opinion there is. In my opinion, organizational leaders should ask themselves: Is this where I want to be with my organization? Is our business so complex that it justifies the proportion of self-interaction, or are we more engaged in navel gazing than understanding our customers and delivering value to the customer.

This consideration is even more relevant for Support Functions, because as cited above, there is no self-purpose of a support function. They should be more paranoid about organizational navel gazing, as they can be seen as a commodity. Commodities are prey for outsourcing, as they are usually only defined and distinguished via the price.

As you realize from the formulation, the data is bringing the question: What is the right proportion of self-interaction and interaction with others. Are we for our specific organization in the organizational sweet spot or are we enjoying organizational navel gazing. What is driving the amount of self-interaction in our organization? Is it the complexity of what we deliver to other entities in the company (e.g. IT dealing with newest model of Artificial Intelligence is undoubtedly an area of some complexity)? Or are cumbersome and un-streamlined processes to blame for the amount of self-interaction? All these are relevant questions. Relevant business questions. And thus, the answer to these questions needs to come from the business. Sorry, guys, don’t blame the nerd. If you are not understanding your business, well ... . But at least the data insights have one or two helping hands:

If there are several organizations of comparable size, what justifies the differences in the self-interaction proportion? Benchmarking against other organizations in the company can be an analysis tool. Imagine possible that this kind of Human Network Analysis was mainstream, then a benchmarking of various support functions with other companies would be possible: Why is Ericsson’s IT self-interaction at 55% for 2000 employees, while Amazon IT self-interaction for 2000 employees is at xyz %.

If my organization is above the regression line, what justifies the “above average” self-interaction. Again, imagine possible that this kind of Human Network Analysis was mainstream, then a benchmarking of various company regression lines would be possible. Why is Ericsson’s regression line different from the regression line of Google.

In my opinion, there are enough enlightening questions out there.

?

Appendix:

The term “organizational navel gazing” has obviously been chosen to provoke the thinking about a wanted position. It is only justified if the percentage of self-interaction leaves the sweet spot area and approaches the area of “pure” self-interaction. The term “organizational brain power” was proposed, which is a more just description in the organizational sweet spot area.