Hierarchy is dead. Long live hierarchy.

Like Tweet Pin it Share Share Email

It’s fashionable these days to rail against the hierarchical organization. Business pundits frequently claim the age of hierarchy is over. They claim to have developed exciting alternatives, which they’ll sell you for a small fee.

Uh … no. Far from being dead, hierarchical organizational design continues to be inevitable for all but the smallest organizations, and for a very simple reason: There’s no other way to sort out who ought to be responsible for what when coordinating the efforts of a lot of people.

(Please forgive the clumsy construction. The alternative was, “An organizational chart is the only way to organize an organization,” which is even worse.)

Hierarchy is the only way to organize an organization because of what a hierarchy is: An outline — a successive decomposition of how to delegate responsibility for what a business has to do to be successful. I’ll believe hierarchical organizational designs are obsolete when someone convinces me an outline is obsolete as a way to organize thoughts.

This doesn’t mean there’s only one way to organize a business into a hierarchy. Far from it — how to organize the hierarchy is a tough challenge.

At the large enterprise level, CEOs generally make the first level of the hierarchy autonomous business units, and leave organizing each autonomous business unit to the person running it (otherwise the business units wouldn’t be autonomous).

CEOs of less-than-giant businesses … or of the autonomous business units that comprise most giants … commonly use these as organizing principles when delegating responsibilities: Business functions (some of which are better characterized as shared services), product categories, customer segments, sales and distribution channels, or geographical regions.

The tricky part, as anyone knows who has had to design an organization, is that at most you can make use of two organizing principles when figuring out the top level of the company, one of which must be shared services (unless you want multiple, independent and incompatible general ledgers).

Even with just one organizing principle, clarity isn’t absolute, because whoever is responsible for, say, the IT organization (let’s call this person you) has a fuzzy boundary with whoever is responsible for Human Resources with respect to questions like recruiting, employee policies, and training. Every fuzzy boundary is a challenge because on the one hand, you want to tailor everything to fit IT’s specific requirements while on the other, the head of HR will focus on the importance of uniform standards.

With two organizing principles it gets interesting. If, for example, you delegate some responsibilities based on the principle of shared services and others based on that of customer segments, determining who is responsible for (for example) teenage consumer customers (customer segment) needs the authority to make branding decisions, but so does whoever is responsible for marketing (shared service).

Now imagine a company organized into shared services, customer segments, and geographic regions. Figuring out who does what when positioning the company for its European teenage consumer customers would be unnerving. In effect, everyone is responsible for everything. Executives, even with the best of intentions, will have little recourse but to fight over turf so as to control the factors that affect their ability to succeed.

So stick with shared services plus one other organizational scheme. Everything else will be handled at lower levels of the hierarchy. So far, nobody has developed a superior alternative for organizational design.

For decision-making? That’s another story.

View the organizational hierarchy as software designers view an object hierarchy. Each organizational unit is a black-box. Every other organizational unit interacts with it exclusively through well-defined interfaces that define the inputs and outputs.

When a decision isn’t completely within the domain of a single organizational unit, it flows to a higher-level organizational unit for resolution according to a well-defined governance process. Decisions flow up until they reach an executive who completely owns them. That executive makes the decision and sends instructions back down the command hierarchy.

It’s the sort of approach that looks great in the PowerPoint. Here in the world, it keeps those with the deepest knowledge and clearest view away from the most important decisions. Instead, the person with the right authority makes them.

It’s this meaning of “hierarchy” — hierarchical decision-making, not hierarchical organizational design — that’s supposed to be dead, replaced by collaborative, informal networks of whichever employees are most qualified to handle the situation.

Executives who prefer the hierarchical decision model must distrust either the expertise, judgment, loyalty, or willingness to collaborate of those closest to the action.

Which brings up the question: Why are these executives hiring and keeping people like this?

Comments (4)

  • “I’ll believe hierarchical organizational designs are obsolete when someone convinces me an outline is obsolete as a way to organize thoughts.”
    One word: Wikis. 🙂

  • Bob,

    Nice try, but I don’t think you have completely disentangled the hierarchy from the decision-making process yet.

    The problem is that you have delegated “responsibility”.

    So if we have informal, collaborative networks taking decisions without the involvement of the managers, how can those managers be responsible for the decisions taken?  And unless managers are informed about the decisions taken, then how can they judge whether the collaborative efforts are working or not?  But then if managers have to be informed about every decision taken, then aren’t we back at square one?

    “Responsibility” means that the manager must be able to control the ultimate outcome.  And there are many situations where multiple valid approaches means that intelligent people (no matter how excellent theirexpertise, judgment, loyalty, and willingness) may choose different solutions that, taken as an aggregate, detract from the organisation’s overall efficiency.

    To avoid micromanaging decisions, I would suggest that managers need to create principles-based frameworks that allow individual actions to be guided in a single, cohesive direction without anticipating each possible scenario in advance (ie not generating rigid procedural instructions).

    An important corollary: The process for creating these principles needs to be both transparent and accountable to ensure that staff accept their validity.

  • According to Elliott Jaques’ social science research that provided the basis for his “Requisite Organization” (RO) theory, there are two pertinent principles that explain the predominance of hierarchical structures:

    (1) Not all work has the same complexity, as represented by the “time span of discretion” required to do the work.

    In RO theory, there are at least eight discrete levels of work that can be identified based on increasing time span of discretion required (complexity). For example, basic clerical work is at Level One, in that a clerical worker plans his work (filing, typing, answering phone calls, routing mail, etc.) based on what he can complete each day (i.e., the time span of discretion is one day or less). Going up to Level Three, the Administration Manager must plan his work based on his responsibilities within the framework of the annual budget and operational plan (i.e., the time span of discretion is one year).

    Depending on the amount of the organization’s true work, this principle implies that each role (position) in an organization belongs at some given level of work. Further, there is some maximum level of work for any organization, depending on the highest required time span of discretion for that organization (that would be the level of the President, CEO, or whatever you choose to call it).

    (2) Not all people have the same current or future cognitive capability for dealing with the complexity of work represented by any particular time span of discretion (in RO this measure is called “complexity of information processing” or CIP).

    According to RO theory, every individual has both a current CIP capacity and some maximum CIP capacity. The current CIP determines fitness for current roles (with respect to time span of discretion), while the maximum CIP determines the fitness for prospective future roles. Further, since not all people have the same current CIP, not all people have the same potential for performing at any given level of work (again, based on the required time span of discretion as a proxy for the work complexity).

    The composite implication of these two principles is that every organization has work to do that is at different levels of complexity, with employees that are best suited for working at one level of complexity or another. In other words, hierarchy is required by the very nature of work itself and the innate complexity-processing ability of the workers.

Comments are closed.