When people talk about an organization’s structure, they often mean the org chart. Who reports to whom, how many layers there are, which teams sit where. That’s part of it, but it’s not the whole story.
In organizational behavior, organizational structure is how job tasks are formally divided, grouped and coordinated. The design decisions behind that arrangement usually come down to six elements: work specialization, departmentalization, chain of command, span of control, centralization (versus decentralization), and formalization. Get those choices right and people know what they own, who they answer to, and how much discretion they have. Get them wrong and you get duplication, bottlenecks, and a lot of “that’s not my job.”
These six elements are a standard teaching framework in management and OB courses. They also sit underneath the bigger structural forms discussed elsewhere on the site, such as Mintzberg’s organizational configurations. In this article we’ll walk through each element, then come back to what they mean for a manager who’s building or reshaping a team.
A Running Example: A Startup That Outgrows Its Flat Chart
This is a hypothetical example.
Imagine a software product company called Northloop. Two founders and six staff work out of a shared office. Everyone does a bit of everything. Decisions happen in a group chat. There are almost no written procedures. That works while the product is small and the customers are few.
Three years later Northloop has 85 people, three product lines, and clients in three countries. The founders still want to approve every hire and every pricing exception. Support tickets sit unanswered because nobody owns them. Two people rewrite the same feature. New staff keep asking, “Who do I go to with this?”
Northloop hasn’t suddenly become a bad company. It has outgrown a structure that was never deliberately designed. The six elements give us a way to think about what needs to change.
1. Work Specialization
Work specialization (sometimes called division of labor) is the degree to which activities are broken into separate jobs. At one extreme, one person designs, builds, sells and supports the whole product. At the other extreme, each person does a narrow, repeated task.
The classic case for specialization is efficiency. Frederick Taylor’s scientific management and Henry Ford’s assembly line both pushed the idea that people get faster when they focus on a few tasks, tools stay where they’re needed, and skilled work isn’t wasted on simple tasks. By the mid-20th century most manufacturers had taken specialization a long way.
But specialization has a cost. When jobs get too narrow, people get bored, quality can slip, and absenteeism and turnover rise. From the 1960s onwards many firms responded by widening jobs again, through job enlargement and job enrichment.
At early Northloop, specialization is low. Priya writes code, talks to customers, and drafts the help docs. That keeps people flexible. At 85 staff, low specialization starts to hurt. Priya is still jumping between coding and support, so neither gets done well. A sensible move is to create clearer roles: product engineers, a small support team, a sales role. Specialization rises, but not so far that people never see the bigger picture.
The practical question for a manager is: how finely should we divide the work? Too little and people are overloaded and unfocused. Too much and the work becomes dull and coordination gets harder.
2. Departmentalization
Once jobs are split up, they need to be grouped. Departmentalization is the basis on which jobs are clustered. Common bases include:
- Function: marketing, finance, HR, operations. People with similar skills sit together. Easy to build expertise, but functions can become silos.
- Product or service: a separate unit for each product line. Accountability for that product is clearer.
- Geography: regions or countries. Useful when customers and regulations differ by location.
- Process: stages of a workflow (casting, finishing, packing; or intake, assessment, discharge in a hospital).
- Customer: consumer, business, government. Useful when different customers need different treatment.
Large organizations often mix these. Manufacturing might be organized by process, sales by geography, and head office by function.
Northloop starts with no real departments. As it grows, the founders group people by function: engineering, product, sales, and customer support. That helps specialists talk to each other. The risk is that engineering and sales stop talking. Later, if the three product lines pull in different directions, Northloop might switch some grouping to product teams instead.
There’s no single “best” basis. It depends on what the organization needs to coordinate most closely, and on how uncertain the external environment is.
3. Chain of Command
The chain of command is the unbroken line of authority from the top of the organization to the bottom. It answers a simple question: who reports to whom?
Two related ideas usually travel with it. Authority is the right that comes with a managerial position to give directions and expect them to be followed. Unity of command, which Henri Fayol treated as a core principle of administration, says a person should report to only one boss. When someone has two bosses giving conflicting instructions, priorities get messy fast.
Modern workplaces have softened this. Matrix structures, project teams and remote work mean many people effectively answer to more than one person. Technology also lets information skip layers. Still, when things go wrong, someone needs a clear line of accountability. The chain of command hasn’t disappeared. It’s just less rigid than Fayol imagined.
At Northloop the early “everyone reports to the founders” model collapses once there are dozens of staff. The founders become a bottleneck. Creating team leads and a head of engineering rebuilds a usable chain. The trade-off is more layers between a front-line engineer and the CEO, which can slow messages up and down.
This also connects to what managers actually do day to day. Clear reporting lines shape how managers plan, coordinate and control work, which is part of the broader checklist of what managers do.
4. Span of Control
Span of control is how many people a manager can effectively supervise. A wide span means one manager oversees many people, so the organization tends to be flatter and cheaper on management payroll. A narrow span means more managers and more layers, so the organization is taller.
Classical writers worried that spans could get too wide. In 1933, V. A. Graicunas published a paper (later reprinted in Gulick and Urwick’s Papers on the Science of Administration) arguing that the relationships a manager must manage grow much faster than the headcount. It isn’t only the link between the manager and each person. There are also the links among those people, and the combinations they form in meetings. Graicunas suggested that for interlocking work, four or five direct reports was often a practical ceiling. Lyndall Urwick later defended limited spans in a 1956 Harvard Business Review article.
Modern practice is more flexible. Routine work, experienced staff, good information systems and clear goals all support wider spans. Complex, interdependent work usually needs narrower ones. Many organizations have widened spans to cut cost and push decisions downward. That can work. It can also leave managers with too many people and too little time for coaching.
At Northloop, one founder currently “manages” 20 engineers. That’s too wide for complex product work. Splitting into teams of six or seven under team leads is more realistic. The payroll for managers goes up a little. The quality of supervision usually goes up a lot.
5. Centralization and Decentralization
Centralization is the degree to which decision making sits at a single point, usually near the top. Decentralization pushes decision rights downward, so lower-level managers and staff can act without waiting for headquarters.
Neither is automatically better. Centralization helps consistency, control and coordination across units. Decentralization helps speed, local adaptation and employee involvement. Harvard Business School research on decision rights makes the trade-off plain: you want decisions made where the relevant information is, but you also have to manage the risk that local decision makers pursue local interests. Over-centralizing is a common mistake. So is leaving it unclear who actually has the right to decide.
Think of Northloop’s pricing. Early on, the founders approve every discount. Clients wait days for an answer, and a competitor that can quote in 24 hours wins the deal. Decentralizing routine discounts to sales, within clear limits, speeds things up. Big, unusual deals can still go upstairs. That’s not “all or nothing.” Most real organizations mix the two.
A related point: centralization and formalization are different dials. A firm can be decentralized (local managers decide) and still highly formalized (they must follow a detailed rulebook). Or the reverse.
6. Formalization
Formalization is the degree to which jobs are standardized by rules, procedures and written job descriptions. In a highly formalized job, the person has little say over what to do, when to do it, or how. In a low-formalization job, people improvise within broader goals.
High formalization brings consistency. An accounts payable clerk processing invoices the same way every time reduces errors. An airline checklist reduces risk. The downside is rigidity. Staff faced with an unusual case may freeze, escalate everything, or work around the rules quietly.
Low formalization suits work that needs judgment and creativity. It also depends more on hiring capable people and trusting them, which links to corporate culture and how groups actually coordinate day to day in groups and teams.
Northloop’s early days are almost unformalized. That helps speed. At 85 people, the absence of basic procedures means every support case is reinvented. Writing a short playbook for common tickets raises formalization in one area without turning the whole company into a bureaucracy.
How Do the Six Elements Fit Together?
It’s tempting to treat each element as a separate checklist item. In practice they move together.
A tall hierarchy with narrow spans, high specialization, high formalization and centralized decisions is the classic mechanistic pattern: efficient in a stable setting, slow when the environment shifts. A flatter structure with wider spans, broader jobs, lower formalization and more decentralized decisions is more organic: better for adaptation, harder to control at scale. Mintzberg’s configurations are one way of describing the resulting forms. Handy’s culture types are another lens on how those choices feel to people inside the firm (see Handy’s model of organizational culture).
For Northloop, the shift isn’t from “no structure” to “full bureaucracy.” It’s a set of deliberate moves: more specialization in a few roles, functional departments with some cross-team rituals, a clearer chain of command, spans that managers can actually handle, selective decentralization of routine decisions, and light formalization where consistency matters.
What Should Managers Watch For?
A few practical checks help when you’re redesigning work:
- Where are the bottlenecks? If one person must approve everything, centralization or the chain of command is probably too tight.
- Where is work duplicated or falling through gaps? Specialization and departmentalization may be unclear.
- Are managers drowning in direct reports? Span of control may be too wide for the complexity of the work.
- Are people following a thick rulebook for tasks that need judgment? Formalization may be too high.
- Are customers or front-line staff waiting for decisions that local people could make safely? Consider decentralizing those decisions with clear limits.
Also watch the human side. Structure shapes identity and politics. A departmental reorg can feel like a demotion even when the job title stays the same. Explain the “why,” involve people where you can, and don’t change all six dials at once unless you have to.
Key Points to Take Away
- Organizational structure is how tasks are divided, grouped and coordinated. Six common design elements are work specialization, departmentalization, chain of command, span of control, centralization/decentralization, and formalization.
- Specialization raises efficiency up to a point, then boredom and coordination costs rise.
- Departmentalization can be by function, product, geography, process or customer, and large firms often mix bases.
- Chain of command and unity of command clarify authority, though matrix and project work soften them in practice.
- Span of control shapes how tall or flat the organization is; Graicunas showed why relationships multiply faster than headcount.
- Centralization and formalization are separate choices. Match decision rights to where the information sits, and use rules where consistency matters without crushing judgment.
Sources
- Lumen Learning / LibreTexts. What is Organizational Structure? (CC BY).
- Nickols, F. Span of Control: The Formulas of V. A. Graicunas (summary of Graicunas 1933 / Gulick & Urwick 1937, and Urwick 1956 HBR).
- Harvard Business School Working Knowledge. Decision Rights: Who Gives the Green Light?
- OpenStax. Principles of Management, 10.1 Organizational Structures and Design.
