Under what circumstances would a company adopt a matrix organizational structure?

• Tweet

  Nội dung chính Show

  • Why would a company consider implementing a matrix organizational structure?
  • What companies use a matrix organizational structure?
  • What is an example of a matrix organizational structure?

• Post

• Share

• Save

• Get PDF

• Buy Copies

• Print

Most discussions about matrix organizations usually quickly devolve into a debate between two sides: those who love to hate the matrix, and those who hate to love the matrix. The former claim that a matrix structure slows decision making and obfuscates accountability. The latter retort that a matrix structure is an inescapable prerequisite for lateral coordination in large complex businesses. From our two decades of experience with organization design, we tend to side with the latter. In fact, we may even belong to a third camp, those who love to love the matrix. But our love is conditional upon its sparing and wise use.

Let’s get back to first principles first. Just in case you’ve forgotten, a manager in a matrix organization has two or more upward reporting lines to bosses who each represent a different business dimension, such as product, region, customer, capability, or function. It’s often a response to, or a prophylactic against, corporate silos. Silos can form in any company, regardless of how it’s organized, whether that’s around different products, different regions, or different types of customers. When a company reorganizes, it’s often because the strategy has also changed. For example, the French global energy player ENGIE recently tilted its primary dimension from product (such as power, services, and infrastructure) toward region in order to better serve its clients in the territories in which it operates. Likewise, the British communications services company BT Group recently tilted from product toward customer (such as consumer, business, and public sector). But such reorgs don’t make silos go away — they just create new ones. So it remains crucial to ensure lateral coordination between the various units.

Lateral coordination is accomplished rather easily at the top of a company. The executives in charge of the various groups sit together naturally in the top management team. Often, they are also incentivized for the company’s well-being as a whole.

But obviously the top management team cannot afford to let all day-to-day operational coordination issues escalate upward. Its real challenge is to achieve lateral coordination also at the levels below. This can be achieved through hard-wiring or soft-wiring. A matrix structure is an example of hard-wiring, because the two bosses of a manager in a matrixed position have the joint responsibility to set his objectives, supervise his work, do his appraisal, and ensure his development. For example, the procurement manager of a business unit would report both to the business unit head and to the corporate procurement officer.

Soft-wiring relies on more informal, organic, voluntary, temporary, or one-off instruments, such as an ad-hoc multi-dimensional task force, an annual corporate planning cycle, an advisory council, a central coordination function, or a company-wide knowledge management system. For example, the U.S. oilfield services company Schlumberger builds on its knowledge management system to share technical expertise across regions and products.

Executives who are fundamentally opposed to a matrix do not argue that there’s no need for lateral coordination. They simply consider that soft-wiring can do the job all by itself. Whenever there is a real need in the field for coordination across organizational silos, the managers concerned will find each other. A matrix, they argue, needlessly complicates things. Patrick De Maeseneire, the former CEO of the Swiss HR solutions provider Adecco, formulates it as follows: “No complicated matrix structures. One man or woman. One number. In matrix structures, everybody claims a success as his and points to everybody else when things go wrong.”

We would argue that it is not an either/or issue. Provided that the hard-wired matrix is deployed sparingly and wisely, it has its place in the arsenal of management tools along with soft-wired ones. Here are five practical guidelines.

1. Adopt when purposeful. The matrix should not be the default design option. It should be used only when two conditions are met. First, when there is a major need for middle managers of different units or teams to coordinate on important business matters on a daily basis. For example, when the finance manager of a region has to coordinate intensely with the heads of the country subsidiaries in that region, it may make sense to put the business controller of each subsidiary in a matrixed position, that is, to have each of these report not only to their respective subsidiary head but also the region finance manager. The second condition for a matrix is that the required coordination cannot be achieved adequately through soft-wiring only. For example, if you are in a fairly steady business with fairly autonomous country subsidiaries selling either standard or strictly local products, something like a quarterly “country coordination council” may be sufficient.

2. Keep intrinsic conflict out. Research by Joachim Wolf and William G. Egelhoff shows that the likelihood of intra-organizational conflict in a matrix structure depends on the groups involved. For example, a matrix with region and function as organizing principles tends to have lower levels of conflict than a matrix with region and product as organizing principles. The reason is easily understood: region and function have more complementary roles and objectives than region and product have. If both region and product have P&L responsibility and are expected to manage the same factors (human resources, customer relationships, price levels, etc.) to optimize their performance, conflict is baked into the matrix. Therefore, when defining the roles and accountabilities of the two matrix dimensions, make sure there are intrinsic reasons for them to collaborate rather than to compete.

3. Limit breadth and depth. A matrix may contain more than two organizing principles (e.g., product + region + function). But since a matrix with two such principles is difficult enough to manage, stick to two unless there is an overwhelming argument to broaden beyond two, such as in a particularly complex part of the business. A trickier issue is how deep down the hierarchy to replicate the matrix. We would advocate avoiding nested matrices. Continuing the aforementioned example of the subsidiary controller, “nested” means that the risk manager reporting to the matrixed subsidiary controller would also report to the regional risk manager who in turn reports to the regional finance manager. If grasping the preceding sentence is already quite challenging, imagine how difficult it would be to make the concept work in practice.

4. Don’t pretend. The previous three guidelines are a call for the spare use of a matrix. But once you have opted for one, go all the way. Ban the oft-used distinction between a dotted and full reporting line, implying that the former carries less weight. Position the two reporting lines of a matrixed manager as fully balanced — that is, 50-50, not 70-30 or some other unequal ratio. (Don’t even use a dotted line to distinguish between different reporting lines on the org chart. Use full lines in different colors.) Likewise, don’t degrade the joint objective setting and performance appraisal of a matrixed manager into an almost perfunctory ritual whereby one of the two bosses actually takes care of it, merely soliciting some final secondary comments from the other boss. Messages or practices that appear to detract from the fundamental philosophy of the matrix are lethal because the effectiveness of a matrix depends on its credibility.

5. Escalate by exception only. A common complaint about a matrix structure is that it increases upward reporting and slows decision making. The opposite should be true in a well-functioning matrix because it pushes operational decision making down in a controlled way. Suppose product and region are the two organizing principles. When a decision is to be made about some minor regional product adaptation, the regional product manager should be able to make that judgment without escalating the issue to a global product manager and the regional business unit head, let alone to even more senior executives. It is up to the higher levels to refuse unwarranted upward escalation of trade-offs and conflicts.

Nothing in business is perfect, and all organizational structures involve trade-offs. If you’re implementing a matrix at your company, start small, learn, and fine-tune as you progress. There is not a one-size-fits-all solution. How far you implement a matrix depends on the maturity of your organization, i.e., its ability to understand that a seemingly complex and ambiguous setup can in fact improve the quality and speed of decision making. Clear communication and consistent behavior are required to dispel the matrixed manager’s anxiety about roles conflict and the boss’s fear of losing power. Ultimately, your organization will gain enough trust in the matrix to let it do its work, evolving from reluctant acceptance to full-hearted embrace, sensing the matrix is there without noticing it.

Why would a company consider implementing a matrix organizational structure?

What companies use a matrix organizational structure?

What is an example of a matrix organizational structure?