matrix orgg right size

Three Misconceptions about the Matrix Organization

What is a matrix organization?

Are matrix structures inevitable in complex organizations?

Unfortunately, there is a lot of confusion about this.

One reason is that textbooks and articles fail to provide good definitions.

A book I just read proclaims that the matrix “is an excellent choice” for higher education institutions. But it doesn’t describe in any detail what it is.

Consultants and “thought leaders” who recommend this solution aren’t very precise, either.

This is one reason why some misconceptions have taken hold.

Let me address the three most common ones.

Misconception #1: We are talking about the same thing when we use the term “matrix organization”.

Consider a fairly typical re-design process in an organization. A project team or working group is considering alternative organizational models. They first look at a product based model:

…they may then compare it to a second alternative, such as a geographical model:

…but then, somebody may ask why they can’t combine them, and they may produce a drawing like this one:

It might be a good idea to combine two different dimensions.

But this type of illustration is not very helpful in itself. The reason is that it leaves a lot of questions open to interpretation.

What IS it supposed to show?

  • Does it mean that the units that are connected by the lines are supposed to collaborate, and if so, about what and in what manner?
  • Does it mean that people in these units will participate in joint meetings, teams or projects?
  • Does it mean that these units will share the same goals and KPIs?
  • Does it mean that people, somewhere (usually not shown on this type of chart) will report to two or more managers? (If so, who does this apply to – will everybody have two bosses, or just some people at a specific level of the organization? )

Or all of the above, or none of the above?

It wasn’t always like this.

When the matrix solution was first introduced, it was relatively clear what it entailed. But it has gradually become a very ambiguous concept.

There are also multiple variations of it, and these variations are usually not described or discussed.

As the author Stephen Haeckel put it, ambiguity is good for many activities in life.

It’s a good thing if you are writing a poem or creating an abstract painting.

But it is the mortal enemy of good systems design.

So the kind of drawing that you see above is not very helpful – unless we specify what we mean.

To do that, we need to use a different format, we need different types of illustrations – and a more precise vocabulary.

Only then will we be able to explore what the actual design options are and identify the risks and limitations of the different alternatives. Only then will we be able to create a common understanding among those who are involved in making the decision about a new organizational model.

Misconception #2: “The main challenge of the matrix is that people have to work for more than one boss”

It is fairly common to hear that the main problem with the matrix is that it’s difficult to work for – or report to – multiple managers.

However, there are few organizations today, matrix or not, where you only relate to one manager.

In most organizations, regardless of reporting structure, you work for multiple managers, participate in teams and projects, contribute to cross-functional improvement initiatives, and so on.

This can be time-consuming for sure, but it is not necessarily the main problem.

What is problematical, however, is reporting to two or more managers if the goals and KPIs that these managers have been assigned are in conflict with each other.

This may be the case if the units that the managers belong to have conflicting mandates or overlapping jurisdictions, something that happens relatively frequently in large and complex organizations.

And installing matrix reporting lines doesn’t do much to resolve this problem. On the contrary, it will probably make it worse.

(If you want want more detail, there’s a piece of research that confirms this point. You can find it here.)

The implication is: If you have a matrix structure that doesn’t work, you don’t need to abandon it completely. With a few modifications (aimed at clarifying roles and removing overlaps in unit mandates) it may become a lot more effective…

….which leads me to my third point:

Misconception #3: “The choice is either/or”

Some proponents of the matrix solution basically say: “Either you have a modern organization with a matrix, or you are old-fashioned and have a silo organization”.

In my opinion, this is a false dichotomy.

It is based on a confusion between two different things – the formal reporting structure, or who should be the line manager, and the extent of information sharing and collaboration across units.

It’s false dichotomy because you can of course have extensive collaboration across units even in an organization where every employee has only one line manager.

The matrix is one type of multi-dimensional organization, where you have different units such as product units and geographical units at the same level in the organization.

However, as I describe in my book, there are other organizational models that also have this feature and that look similar on paper, but that do not assume that employees report to two bosses and do not require overlapping goals and KPIs (see a brief summary of the line of reasoning in this blog post).

The matrix organization is also quite rare (if you define it as I do – “a two-boss reporting structure”).

During the last 7 or 8 years, I have interviewed managers in about 40 different firms.

Only 2 or 3 of these firms had a matrix with a two boss reporting structure, and in these cases, it applied to a relatively small group of people, typically middle managers at one specific level or in one part of the organization.

So at least in my experience, a formal matrix structure is not very common.

However, what is quite common is something I call a hidden matrix.

By that I mean governance processes and authority relationships that cross the formal units, but which are not shown on the official organization chart (and that are usually not deliberate).

This is an important issue, which I will discuss in my next blog post.

Digital-twin-www.organizationdesign.net_-768x524

Creating a Digital Twin of your Organization

The concept of a digital twin has been around for at least two decades. It refers to having a computerized representation of a real-world physical system, such as an engine or a building.  

In the initial phase, the digital twin may be more like a “digital shadow”, that is, it may be based on data received once.

But over time, it may be extended to be a living simulation model that updates and changes as the system is undergoing changes.

In other words, a digital twin is more than a dashboard. Dashboards provide visualization of current processes. Digital twins enable simulation and prediction.[1]

Because organizations are also systems, the question becomes: Is it possible to create digital twins of organizations?  

This is in fact the vision that we are now working toward in Reconfig. In the current version of the tool, you are able to visualize the work processes that people participate in and the interactions between them. This is then mapped onto the current organizational structure.

But more importantly, you can alter the structure. For example, you can move roles from one unit to another, or combine two units. When you do this, you can see the effect in real-time by means of five indicators (a sixth will be added soon).

The key purpose is to allow decision-makers to visualize and test the effects of potential organizational changes before they are introduced. See the image below for an example.

I am aware that other people, particularly in the business process management community, are also working on similar concepts.

However, it seems that their goal is somewhat different. Typically, these efforts are based on data that are extracted from existing IT systems, such as event logs, and from what I gather, the purpose is mainly to support operational efficiency.

In contrast, in Reconfig, we are more oriented toward the overall, organizational model or the “architecture” of the firm, if you will.

We collect data by using a survey questionnaire distributed to employees of the organization. We have been experimenting with other data sources, and in future, we may, for example, import email traffic data into Reconfig.

But our experience so far is that it is easier to get valid data directly from the employees in the organization. It means that we can get their views on the activities and working relationships that they believe are most important. It also has the benefit of involving them in the process, and prepare them for the analysis that follows.

The creation of digital twins of organizations will change organization design.

I have always insisted that we should use a data-driven approach. But this approach goes further. To borrow terms from two well-known contributors in my field,[2] it not only allows us to describe “what-is” but to explore how our organization “might-be” or “should-be” in the future.

 

  

[1] Becker, M. C. & Pentland, B. T. (2022). Digital twin of an organization. Are you serious? In A. Marrella & B. Weber (Eds.), BPM 2021 Workshops, LNBIP, pp. 243-254.

[2] Burton, R. M., & Obel, B. (2011). Computational modeling for what-is, what-might-be, and what-should-be studies—and triangulation. Organization Science, 22(5), 1195–1202.

Image by Freepik