Hierarchy

Hierarchy is the ordering of an organisation into levels of authority. Each role reports to one above it, and the resulting chain determines who instructs, who approves and who is accountable for what.

In an HR context it shows up as three practical things: the reporting line on the employee record, the approval routing for leave, expenses and hiring, and the escalation path when something needs a decision above the current level.

It is worth separating from two adjacent ideas. Grade fixes the pay band and benefits eligibility. Designation supplies the external title. An organisation can promote someone in grade without changing their position in the hierarchy, and frequently does.

Four arrangements cover most organisations: functional, divisional, matrix and flat. Many use more than one at once.

Most Indian organisations of any size are functional with divisional elements, and adopt matrix reporting for projects without formally acknowledging it. The unacknowledged matrix is the difficult case: an employee takes direction from a project lead who has no input into their rating and no visibility of their leave.

There is no correct span of control. Span of control is the number of people reporting directly to one manager. It and the number of levels are two views of the same arithmetic: for a given headcount, widening spans reduces levels and narrowing them adds levels.

The right width depends on how similar the work is, how experienced the team is and how much of the manager's time goes to their own delivery rather than to managing. A team of eight doing similar work with an experienced manager is straightforward. Eight direct reports doing eight different things, with a manager who also carries individual targets, is not.

Adding levels has costs that are easy to underestimate. Each additional level slows decisions, dilutes information passing up and down, and creates a title expectation that has to be met. Removing levels, usually called delayering, transfers that load onto the remaining managers and is often reversed within a few years.

The signal that levels have grown past usefulness is when routine approvals require three signatures and nobody can name what each is checking.

Organisation charts drift, and the drift is worth detecting because several systems depend on the chart being accurate.

• Leave and expense approvals route to the recorded manager. Where someone has moved teams and the record was not updated, approvals go to a person with no visibility of the work.
• Appraisals are assigned on the reporting line. A wrong line means a rating written by someone who did not observe the year.
• Escalation paths and business continuity assume the chart. A vacancy that was never backfilled leaves an approval step with nobody in it.
• Statutory and safety responsibilities are frequently attached to a position rather than a person, and an unfilled position leaves the responsibility unallocated.

The maintenance discipline is unremarkable: update the reporting line at the point of transfer rather than at the next appraisal cycle, and review vacant supervisory positions monthly rather than annually.

Nothing formal replaces them, so the hierarchy becomes informal. Flat organisation is a popular aspiration and a partial one: removing job levels does not remove the need to decide who allocates work, who resolves disagreements and who is accountable when something fails.

Influence concentrates around tenure, proximity to founders or visibility, and none of it is written down. For an employee this is harder to navigate than a stated structure, because there is nothing to appeal to and no defined path to advance along.

The version that works usually keeps a light formal structure and pushes decision authority downward within it, rather than abolishing the structure. That distinction matters: the goal worth pursuing is fewer approvals, not fewer defined relationships.

For HR specifically, a flat structure still requires a recorded reporting line for every employee, because leave, appraisal and escalation all need one whether or not the organisation describes itself as hierarchical.

Because the employee now has two managers and one rating. Matrix structures exist because work crosses functions and someone has to coordinate it, and this is the problem they create that is rarely designed for.

The functional manager owns their career, their pay band and usually their leave approval. The project or dotted-line manager sees the actual work day to day and often has the better view of performance. If only the first writes the appraisal, the rating reflects the manager with less information.

• Decide explicitly who writes the rating and who is required to contribute to it, and put the second person's input into the record rather than leaving it as a conversation.
• Settle priority conflicts by rule rather than by escalation. Where two managers both need the same person this week, someone should already know which one wins.
• Record the dotted line in the system. An unrecorded second reporting relationship cannot be included in any process that runs off the employee record.
• Keep leave and expense approval on the primary line. Splitting approvals across two managers produces gaps rather than rigour.

The version that fails is the unacknowledged matrix, where a project lead directs someone's work daily with no formal standing at all. The employee absorbs the conflict, and it surfaces as a retention problem rather than a structural one.

Organisations add levels reactively, usually to solve a title problem or to give a good manager somewhere to go, and the levels persist long after the reason has passed.

A few checks are worth running annually.

The last one is worth checking most often, because it is invisible until the person takes leave. A supervisory position left unfilled quietly redistributes its approvals to whoever is nearest, and the org chart continues to show a structure that no longer exists.

When removing a level, move the authority down with it. Delayering that removes the manager but keeps the approval at the level above simply lengthens the chain, which is the opposite of the intention and the reason many flattening exercises are reversed within two years.