When structure doesn’t help (and when it does)

What is the point of structure?

As more and more organizations are moving to matrix structures (even if they often don’t realise it) they are effectively challenging basic assumptions about organizational structure.

This is because at one extreme the matrix means highly complex structures (with multiple reporting lines of both dotted and solid varieties) and at the other extreme the matrix can be a way of working that is largely without structure (in which case it is – sometimes referred to as a globally networked company).  So organizations need to be clear about their attitude to structure so that they implement the right type of matrix.

So which is right type of Matrix?

Nowadays there are many examples of organizations that achieve notable success with hardly any, or even no formal structure.  Of these, open source software organizations such as Linux or Mozilla Firefox are probably the best known.

Does this mean that those who want structure are stuck with in an old world view?  Are these examples going to become increasingly the norm?  Is it a case of finally shaking off the ideas of the military and of mass production line management or would ‘structureless’ organizations be a big mistake?

With this in mind, I have asked the question many times of participants on my courses “What is the point of Structure?”, and for the most part their answers tend to be “because you need to control things” or “it would be chaos without it”.  To my mind, these responses, don’t really answer the question at all.

On probing this question further it increasingly seems to me that leaders like structure because they like to ‘own’ things. They like to know and to say “500 people work in my organization”, or “I’m responsible for £2m turnover” etc.  Equally individuals working within companies want to know where they fit in and which part of the organization they work for, which is why they like some form of structure.

But even though people have these reasons for preferring some structure, it still doesn’t necessarily make it right for the business.  So we come back to the central question….

“When does structure help?” (And when does it get in the way?)

This week I read an article summarising some research that actually spoke to this question directly.  The authors’ work is based on game theory and, within this context, they examined how people would communicate with others within a simulated organization.  They looked at whether communication patterns naturally gravitated to the most efficient situation (the so called Nash Equlibrium) or whether a structure was needed to make sure the most efficient communication was achieved.

It should be emphasised that their studies were conducted in a situation of their creation, where one function was more popular than others (they chose marketing for this role in their simulations).  So they have studied communication where there is potential for a specific communication bottleneck.

Within this world they looked at the impact of a range of factors on communication.

Their basic conclusion is that: – structure helps when at least two of the following factors are high: –

Diversity – the number of independent skill types possessed within the network

Differentiation – the contrast in skill level

Interdependence – the amount that people need each other

Load – the demand for work relative to resource

Urgency – the rate at which pending work becomes useless if not completed

This raises some fascinating questions….

For example, how many of these factors are high because we make them high? And by contrast how many are naturally high because of the specific world we work in and the demands it places upon us?

Would it be better to ensure the right resources are in the right place rather than turn to structure to deal with overload?

How interdependent do we really need to be?  What level of interdependence does the work demand?

Clearly this research tells us that the more we move to ‘star group’ working, the less we will need or benefit from structure.  Equally, the more the team has similar skills the less we will benefit from structure.

In addition

Two other highly important things also flowed from this work for me.

Firstly, when structure was useful, the maximum benefit gained was typically less 10%. This is significant and important, but it is not huge and it raises another question – does the imposition of structure have any downsides (for example in motivation) that would (more than?) offset this gain.

Secondly, in their simulations the main benefit of structure was to prevent or block communication with the (most) popular function and force it elsewhere so structure in this sense is restrictive in that it prevents things.

This article does not provide a definitive answer, but it does raise some very interesting questions and give us some useful pointers.

Reference

Interaction Value Analysis: When Structured Communication Benefits Organizations, by Walid Nasrallah, Raymond Levitt and Peter Glynn, Organization Science Volume 14 No 5, September to October 2003.