I work within a consulting organization which carries out project work for clients. We have a matrix structure - one (weak) structure is staff management, and the (more meaningful) structure is client and project management. There are many aspects of the matrix structure that don't work, but what would be better?
To go into more detail, the company I work for has a number of practices, each led by a practice leader, each with about 25 people. They each have about 6 DRs, some of whom are managers, who in turn have up to about 6 DRs. The managers are responsible for the hiring, firing, reviews and salaries of their DRs, but the manager and their DRs don't actually make up a team that owns any work. In reality the practice leader retains most of the responsibility and it's difficult to delegate it out efficiently.
Overlaid on this is the project structure. Each project has a project manager, who often has a team of perhaps 5 people working for them, mostly full-time, and mostly for several months. The PM is responsible for the day-to-day work of each of their team on a very real basis. This structure has much more actual meaning for most people, although organizationally has less official power. However, it's a fluid structure, as projects start and finish, and as people move between projects.
I'm both a manager and a project manager. From the perspective of allowing me to manage people effectively, being a project manager is great. The roles and objectives of me and the project team are clear, and I can do lots of good MT trinity things. Being a staff manager is much more challenging. It frankly feels much more like a mentorship relationship. I still do O3s, and some of the rest, but it's rather insubstantial.
My main question here is that clearly the matrix structure is creating some challenges, but what would be better? Should we have 25 people reporting directly to the practice leader? We could in theory have managers and teams grouped around particularly responsibilities, but I don't see any natural groupings that would provide much more 'meat' to the roles than now. Should we give more formal power to the PMs? But if so, how do we handle the fluidity of that?
Many thanks,


12string's picture
Training Badge

It seems that using this matrix organization, each employee ends up with two bosses.  That means for every employee there are three relationships to manage (PM to DR, Mgr to DR, and PM to Mgr).  If the two managers have different personalities (novel concept), then the employee will cling to one and push away from the other. This is an unhealthy situation for employees.

With two management structures, it would seem that one of them adds very little benefit to the organization as a whole.  I ask my employees to do tasks that they weren't technically hired to do.  I don't see why managers can't carry the load regarding day-to-day tasks, and managing employees structurally.

I guess if I were to have 25 DRs, I would raise up 3-5 'lead' personnel to manage, and mentor them in managing the rest of the crew.

Just my 1.5 cents.


bflynn's picture

 My instinct on reading the subject line was to think "Better than Matrix?  Ummm...anything?"  I don't like matrix organization but I've had enough experience with them to know that they're not going away anytime soon.  

I've also worked in the specific situation of having a Project Manager and an Admin Manager who were sometimes the same person.  The dedication is toward the PM because that is the day to day work.  Our admin managers handled the admin tasks such as gathering information for performance reviews, promotions and some light talent development work.

A variation on a full matrix that might be better would be a teamed matrix (or mini-matrix).  Teamed matrix splits the large group into several smaller matrix groups managed at a smaller level.  

One of the big problems with a matrix is resource allocation; it takes a lot of effort to keep track of what so many people are doing.  The advantage is that a single person can manage the workload for 20 or 30 employees across 10 projects in the manager's portfolio while an entire team was needed to manage the resource allocation for hundreds of people with dozens of projects in the company's portfolio.  The mini-matrix managers can trade among themselves when special needs arise, but the mini-matrix should be mostly robust enough to handle the projects within it.

Advantages:  The matrix is smaller, so the same people wind up working on multiple teams and you get some of the advantages of solid teams.  You also get most of the load balancing that matrix organizations are meant to solve.

Dis-Ads:  You don't get the full efficiency of the matrix.  There is some wasted time in the team matrix.  Also, employees are still stuck in the matrix, so without good mentoring or a dedicated talent development manager, matrix teams tend to grow stale.

Initial thoughts - actual mileage may vary.





stringl's picture

Brian, 12string,

thanks for your thoughts on this. From reading your comments, it sounds like you agree that the matrix structure we have isn't perfect, but that actually there really isn't anything better. And in fact the fairly small-scale matrix we have (groups of ~25 people) is the best way to go, if there is a matrix structure.

I guess I'll just have to continue to make the best of it.



cruss's picture
Licensee Badge

I guess I missed how this structure offers any benefits over a more traditional structure. What are the main justifications for using Matrix in an org?

At my company we have a very traditional hierarchy of VPs, directors, managers, and directs. We also have a PMO (project management org) that functions as it's own department. Each project gets a PM and they are assigned or request resources from other managers or teams to complete their projects. They use the relationship power they have coupled with the role power from management buy in on the project to get it done.

It sounds like your Matrix is taking Managers and PMs from the same group. While you are PMing one set of people you are "responsible" for another group. It would be difficult to create the necessary relationships to really be effective. I guess that's why most of Mark and Mike's comments on Matrix orgs was about relationship building.

Canyon R

JBormel's picture

No discussion of Matrix Management would be adequate without the perspective from "In Praise of Hierarchy" by Elliott Jaques, in Harvard Business Review, 1990.  There's currently a freely available copy associated with that link.

Jaques' argument is that, in most cases, it's premature to say that traditional management hierarchy doesn't work.  You cannot, in Jaques view, conclude that something other than traditional management hierarchy is necessary, until you've assured that the capabilities of the people in the hierarchy are consistent with their role obligations.  That usually corresponds to "felt-fair salaries", length of the time horizons people can think in, and people's abilities for abstract thinking.  There are only a few "stratum" that adequately capture where most of us are at any point in our lives.  And, we do get "smarter" as we age, but at different rates and only to a point. 


Like Mark and Mike, Elliott Jaques was a humanist.  Relationships are foundational and imply accountability to the trinity (O3, FB, C, and D).  I suspect that he was a High D and High I, but I've only read about him.  He probably was perceived as an anti-S.

Unlike anything Mark has said publicly, Jaques articulated that it was critical that a superior had the capacity to consistently add value to their directs work.  And, to draw a concrete example, it is statistically unlikely that a twenty year old superior has the relevant thinking ability to consistently add value to a 60 year old direct (subordinates)'s work.  The ages are not the point, but the capacity to add value, relative to the directs capacities is critical.  Cognitive capacity does tend to improve with age, as both Mark and Mike have discussed, as recently as in the High Cs Downfall cast.  They haven't gone as far as Jaques to address cognitive capacity as requisite (i.e. required) to staffing the effective management trinity.

To give the guys credit, Jaques "theory" as articulated in the HBR article was only met with relatively private praise and rewards. Publicly, Jaques' Requisite Organization Theory is a non-starter.


There are lots of common reasons why people are promoted up in a hierarchy that cannot add value to their subordinates work.  Although some reasons don't seem appropriate (i.e. The Peter Principle, rising to a level of incompetence), others are highly appropriate (such as strong organizational know-how with weak professional know-how, during a company or project life-cycle phase where that blend is appropriate.)  That does not, however, demand a non-hierarchical (e.g. matrix) approach to structuring accountability. 


To Jaques, matrices usually diffuse accountability in unprofessional and non-productive ways.


When a hierarchy fails, declaring the management structure is the problem is an invalid conclusion.  Jaques articulated that often, matrix management structures evolved exactly from that failure to adequately staff a hierarchy correctly (that is, attentive to capabilities of managers to add value to the work of their directs).


The modern management philosopher, Scott Adams has described this failure to assess competency relative to accountability as follows:


Frame one: 

  Wally:  I don't understand how the new reorganization will help us "focus on our core business"

Frame two: 

  Wally:  Did our core business change?  Or are you saying the _every_ reorg prior to this was a misdirected failure?

Frame three (of three):

  Boss:  Wally, when a car gets a flat tire, what do you do?

  Wally:  Well, if I'm you, I rotate the tires and drive home.



That is matrix management.  It's what you do when you don't or can't recognize a flat tire.




stringl's picture

Canyon, JBormel, thanks for your input on this. That HBR article is very interesting, and helps me clarify my thoughts on this. Let me try and explain my scenario a little more.

Technically perhaps we don't have a matrix. Officially everyone has one manager, who does their annual review and other things. This is the management structure that our HR team would recognize. Project managers then have a team of people to complete their project - similar to the structure Canyon describes.

I think the main difference is that the projects are typically a few months long, people work on them full-time, and the team is often isolated out on a client's site. Most of the time, the project manager has the three critical features that Jaques describes in the article jbormel mentioned. These are, "First, and most critical, every manager must be held accountable not only for the work of subordinates but also for adding value to their work. Second, every manager must be held accountable for sustaining a team of subordinates capable of doing this work. Third, every manager must be held accountable for setting direction and getting subordinates to follow willingly, indeed enthusiastically. In brief, every manager is accountable for work and leadership."

The official manager has none of these things. They perhaps have some shared accountability for them, but in reality the Practice Leader (who the managers report to) has it in full. And as Canyon says, this does indeed make it difficult to create the necessary relationships to really be effective.

This brings me back to my question: what would be better? We used to have a structure where all 25 people reported to the practice leader. This was less artificial than the structure we have now, but it meant that insufficient attention was paid to career development, people got lost somewhat, and the practice leader was overstretched. But perhaps this would be better, and instead we find alternative ways of solving the original problems? Or perhaps there is some other option I'm missing?

Any thoughts?



JBormel's picture


Sounds like you have a good handle on the situation and the org structure is not your primary challenge.

I agree with M-T that attention to DISC and the mgmt trinity (all four pieces of the trinity) are required as a necessary basis for management (ie doing things right).

A secondary method to test the effectiveness of your structure is looking at the typical determinants of what you want to avoid, the passive-aggressive organizational failure mode.  The four tests are:


1) who has the decision rights involved in your business?  That's more than ownership.  It's ownership that those who need to be consulted and informed of decisions are in the loop.  That's real work.  If your periodic status reports (say monthly) don't have "communication plan" identified, that's a sign of a probable problem.  If you have well-run, daily scrum standup meetings, reliable face contact (both at your project sites and with the mother ship,) or equivalent, you may be okay.


2) how does information flow through to everyone who needs awareness, acceptance, and/or commitment to the messages?  Obviously, O3s that happen 80% of weeks of the year, if done by M-T certified manager ensures this.


3) Has there been adequate attention to what motivates each of your employees?  Maybe career growth.  Maybe simply inclusion.  Maybe work opportunities.  Maybe work/home life balance issues.  Inattention to motivators is energy depleting for directs.  If you have someone "role playing" for people not in the room, you've guaranteed all kinds of problems, including demotivation of impacted (non) participants.  This is very common.


4) Lastly, the organizational structure, matrix, hierarchical, wheel-and-spoke, or full contextual structure matters, if it's a legacy to coexisting historical cultures within your organization that are creating a us-versus-them dynamic internally.  This is nearly ubiquitious  in organizations that have grown through acquisition.  If that exists, the organizational structure is a secondary issue.


I hope that helps.

dunnmark's picture

Interesting thread, although it seems that the matrix organization is here to stay. I don't see too many examples of large companies these days that don't employ a matrix structure. I'm interested to hear from the forum of any large companies that are NOT matrix based. I can start the thread by mentioning Danaher - a 50,000 person industrial company. I know people there who say it's a collection of business with GMs, all of whom have all the people in the business reporting directly into them (including HR, Finance, Manufacturing, Product Development...). Impressive - others?


JBormel's picture

 - Most companies do use a primary hierarchy for performance reviews and basic communication functions, not matrices.


- One of the core problems with matrix management is when a direct has two bosses who are each asking (i.e. telling) the direct to make different things their top priority (and "don't do that other things until you complete my thing").  This aspect of the matrix puts the direct in a no-win bind.


- It's extremely helpful to have, over that direct, two accountable executives, one to be accountable to make the decisions (or ensure that they're made correctly according to a professional competency), and a second one to drive the decisions (finishing the project on time and on budget, i.e. the project management role, with special attention to DISC downfalls, interpersonal conflicts, and individual capacity demonstration during the project in the project's context ).  


If those two functions, making and driving decisions, are handled by different people, there are often different workable team structures.  All too often, I've seen the project management role unstaffed, or, a very effective project manager laid off in response to a budget crunch.


Dunnmark, is your experience substantially different?