Designing Structure: Horizontal Differentiation - Principles of Management

As we have seen, vertical differentiation is concerned with the location of decision-makingresponsibilities within an organization.

Horizontal Differentiation in Organizational Structure

In contrast, horizontal differentiation is concerned with how to divide the organization into subunits. We look at four different types of structure here: functional, multidivisional, geographic, and matrix.


Most firms begin with no formal structure and are run by a single entrepreneur or a small team of individuals. As they grow, the demands of management become too great for an individual or small team to handle. At this point the organization is split into functions that typically represent different aspects of the firm’s value chain.

In other words, in a functional structure the structure of the organization follows the obvious division of labor within the firm, with different functions focusing on different tasks. Thus there might be a production function, an R&D function, a marketing function, a sales function, and so on . These functions are typically overseen by a top manager, such as the CEO, or a small top management team.

Functions themselves can be and often are subdivided into subunits. Further horizontal differentiation within functions is typically on the basis of similar tasks and processes. Within the manufacturing facilities of Toyota, for example, the workforce is grouped into teams, and each team is responsible for a discrete activity or task, such as the production of a major component that goes into an automobile.

At Nucor too the workforce is grouped into teams, with each team taking on responsibility for a particular step in the steelmaking process. These teams may have significant decision-making responsibility, be held accountable for their performance, and have their pay and bonuses tied to team wide goals.

A functional structure can work well for a firm that is active in a single line of business and focuses on a single geographic area. But problems can develop once the firm expands into different businesses or geographies. Consider first what happens when a firm expands into different business lines.

The Dutch multinational Philips NV, for example, began making electric lights, but diversification took the company into consumer electronics (visual and audio equipment), industrial electronics (integrated circuits and other electronic components), and medical systems (MRI scanners and ultrasound systems). In such circumstances a functional structure can be clumsy. Problems of coordination and control arise when different business areas are managed within the framework of a functional structure.

It becomes difficult to identify the profitability of each distinct business—and thus to assess whether a business is performing well or poorly—when the activities of businesses are scattered across various functions. Moreover, because no individual or management team is responsible for the performance of each business, there is a lack of accountability within the organization, and this too can result in poor control.

Activities of businesses are scattered across various functions

As for coordination, when the different activities that constitute a business are embedded in different functions, such as production and marketing, that are simultaneously managing other businesses, it can be difficult to achieve the tight coordination between functions needed to effectively run a business. Moreover, it is difficult to run a functional department that is supervising the value creation activities of several business areas.


The problems we have just discussed were first recognized in the 1920s by one of the pioneers of American management thinking, Alfred Sloan, who at the time was CEO of General Motors, then the largest company in the world. Under Sloan GM had diversified into several businesses. In addition to making cars under several distinct brands, it made trucks, airplane engines, and refrigerators.

After struggling to run these different businesses within the frameworkof a functional structure, Sloan realized that a fundamentally different structure wasrequired. His solution, which has since become the classic way to organize a multi business enterprise, was to adopt a multidivisional structure. In a multidivisional structure the firm is divided into different product divisions, each of which is responsible for a distinct business area.

Thus Philips created product divisions for lighting, consumer electronics, industrial electronics, and medical systems. Each division is set up as a self-contained, largely autonomous entity with its own functions. Responsibility for operating decisions and business-level strategy is typically decentralized to the divisions, which are then held accountable for their performance.

Headquarters is responsible for the overall strategic development of the firm (corporate-level strategy), for the control of the various divisions, for allocating capital between divisions, for supervising and coaching the managers who run each division, and for transferring valuable skills between divisions.

The product divisions are generally left alone to run their daily operations so long as they hit performance targets, which are typically negotiated annually between the head office and divisional management. Head office management, however, will often help divisional managers think through their strategies. Thus while Jack Welch at GE did not develop strategy for the various businesses within GE’s portfolio (that was decentralized to divisional managers), he did probe the thinking of divisional managers about their strategies.


In addition, Welch devoted a lot of effort to getting managers to share best practices across divisions. For example, Welch was a driving force in getting different divisions at GE to adopt the six sigma process improvement methodology.

One of the great virtues claimed for the multidivisional structure is that it creates an internal environment that gets divisional managers to focus on efficiency. 17 Because each division is a self-contained entity, its performance is highly visible. The high level of responsibility and accountability implies that divisional managers have few alibis for poor performance. This motivates them to focus on improving efficiency.

Base pay, bonuses, and promotional opportunities for divisional managers can be tied to how well the divisions do. Capital is also allocated by top management between the competing divisions depending on how effectively top managers think the division managers can invest that capital. The desire for capital to grow their businesses, and for pay increases and bonuses, creates further incentives for divisional managers to focus on improving the competitive positions of the businesses under their control.

On the other hand, if the head office puts too much pressure on divisional managers to improve performance, this can result in some of the worst practices of management. These can include cutting necessary investments in plant, equipment, and R&D to boost short-term performance, even though such actions can damage the long-term competitive position of the enterprise.

To guard against this possibility, head office managers need to develop a good understanding of each division, set performance goals that are attainable, and have staff who can regularly audit the accounts and operations of divisions to ensure that each division is not being managed for short-term results or in a way that destroys its long-term competitiveness .


Some firms first grow not by expanding into different businesses through diversification, but by expanding into other geographic regions—either within their home countries or (increasingly in today’s global economy) into other national markets. For firms that are active in multiple regions with a single business, the structural solution to managing growth is to adopt a geographic structure.

In a geographic structure the main subunits of the organization are geographic areas, such as regions within a country, countries, or multicountry regions. Figure illustrates a form of geographic structure found in some international businesses.

Under this structure the firm is divided into geographic areas. An area may be a country (if the market is large enough) or a group of countries. Each area division tends to be a self-contained, largely autonomous entity. Each may have its own set of functions (such as its own production, marketing, R&D, and human resource functions). Operations authority and strategic decisions relating to each of these activities may be decentralized to each area, with headquarters retaining authority for the overall strategic direction of the firm and financial control.

This structure facilitates responsiveness to local market conditions. Because decisionmaking responsibilities are decentralized, each area can customize product offerings, marketing strategy, and business strategy to the local conditions. As we earlier, there are significant differences in cultures and business systems across nations, so this approach has some appeal. On the other hand, this structure also encourages fragmentation of the organization into highly autonomous entities, which can make it difficult to transfer core skills between areas.

Moreover, because each geographic region has its own production facilities, duplication inhibits the realization of economies of scale that could be gained if the firm served the entire world market from a single favorable location. The duplication of functions across regions also implies that the firm is not placing different functions where they can be performed most efficiently.


To solve these problems, many international businesses operate with a hybrid geographic– functional structure, similar to that illustrated in Figure. In this structure functions likeR&D, purchasing, and production are centralized at the optimal locations. The world is then divided into geographic regions for local marketing and sales. A geographic structure, or a hybrid geographic–functional structure, can become unwieldywhen a firm is engaged in several different businesses.

Under such circumstances each geographicarea might itself be divided into different product divisions—one for each business—with functions appearing underneath the divisions. This was the structure Unilever used tooperate with. Although this structure had the advantage of allowing Unilever to customize itsproduct offering and marketing strategy from country to country, the duplication ofmanufacturing facilities drove up costs.

hybrid geographic– functional structure

By the late 1990s this structure was no longer tenable,and in 2000 Unilever established two worldwide product divisions: one to manage its foodsbusiness (packaged foods) and one to manage its home products business (shampoos, detergents).The division heads were responsible for the worldwide profitability of the businessesunder their control.

worldwide profitability

Under this new structure Unilever consolidated its manufacturing in fewer facilities, each located where costs were favorable. Those facilities served regional or global markets, letting the company realize economies of scale and drive down costs. R&D was also centralized at the marketing strategy, as well as product packaging, to account for country differences in tastes,preferences, distribution systems, and the like.

To solve this dilemma, Unilever has kept elementsof its geographic structure in place, with the head of each geographic area maintainingresponsibility for profitability in the area under his or her control, and country managers withineach region being given responsibility for local marketing and sales (and held accountable forperformance in that country).

Unilever’s structure as of 2005, which is illustrated in Figure ,is an attempt to solve conflicting demands on the organization while maintaining the best featuresof a multidivisional structure and a geographic structure. Unilever is hardly alone isstruggling with such a dilemma; many other firms do as well. One solution to such organizationaldilemmas is to adopt a matrix structure, which is in effect how Unilever is now operating.As we will see next, however, matrix structures also have problems.


The matrix structure is sometimes adopted when no single structural design seems to solve all of a firm’s problems. With a matrix structure, managers try to combine two different organizing philosophies in a single design. Unilever, for example, saw benefits to being organized both on the basis of divisions (enabling the company to consolidate manufacturing facilities and realize economies of scales) and on the basis of geographic areas (enabling the company to respond to different national and regional markets).

The firm’s senior managers want all of these benefits, so they have adopted a matrix structure with two overlapping hierarchies: one on the basis of business and one on the basis of area. This means that a lower-level manager might have two bosses—divisional and regional.

In addition to diversified multinational firms like Unilever, high-technology firms based in rapidly changing environments sometimes adopt a matrix structure. In such cases the need for a matrix is driven by the desire for tight coordination between different functions, particularly R&D, production, and marketing. Tight coordination is required so that R&D designs products that can be manufactured efficiently and are designed with customer needs in mind—both of which increase the probability of successful product commercialization.


Tight coordination between R&D, manufacturing, and marketing has also been shown to result in faster product development, which can help a firm gain an advantage over its rivals. As illustrated in Figure, in such an organization an employee may belong to two subunits within the firm. For example, a manager might be a member of both the manufacturing function and a product development team.

A matrix structure looks nice on paper, but the reality can be different. Unless this structureis managed carefully it may not work well. 24 In practice a matrix can be clumsy and bureaucratic. It can require so many meetings that it is difficult to get any work done.

The dual hierarchy structure can lead to conflict and perpetual power struggles between the different sides of the hierarchy. In one high-technology firm, for example, the manufacturing manager was reluctant to staff a product development team with his best people because he felt that would distract them from their functional work. As a result, the product development teams did not work as well as they might.

To make matters worse, it can prove difficult to ascertain accountability in a matrix structure. When all critical decisions are the product of negotiation between different hierarchies, one side can always blame the other when things go wrong. As a manager in one high-tech matrix structure said to the author when reflecting on a failed product launch, “Had the engineering
(R&D) group provided our development team with decent resources, we would have gotten that product out on time, and it would have been successful.”

For his part, the head of the engineering group stated, “We did everything we could to help them succeed, but the project was not well managed. They kept changing their requests for engineering skills, which was very disruptive.” Such finger-pointing can compromise accountability, enhance conflict, and make senior management lose control over the organization.

However, there is also evidence that properly managed matrix structures can work. Among other things, making a matrix work requires clear lines of responsibility. Normallythis means that one side of the matrix must be given the primary role while the other is given a support role. In a high-tech firm, for example, the product development teams might be given the primary role because getting good products to market as quickly as possible is keyto competitive success.

In a diversified multinational firm like Unilever, the divisions might be given the primary role because they are responsible for manufacturing costs and product development, while the geographic areas take on the support role. Clear goals should be well prioritized so that when conflicts occur, which is inevitable, the goals help to indicate what is most important. In Unilever’s case, for example, driving down costs is an important goal, andthe need to first satisfy this goal can be used as a decision rule to help solve any conflicts between the division and area hierarchies in the matrix.

Despite such steps, managing within a matrix structure is difficult. In light of these problems, managers have sometimes tried to build “flexible” matrix structures based more on enterprisewide management knowledge networks, and a shared culture and vision, than on a rigid hierarchical arrangement. Within such companies the informal structure plays a greater role than the formal structure. We discuss this issue when we consider informal integrating mechanisms in the next section.

All rights reserved © 2018 Wisdom IT Services India Pvt. Ltd Protection Status

Principles of Management Topics