Delayering

Delayering removes levels of management. Where a decision previously passed through six people between the front line and the executive, it now passes through four.

The case for it is that each layer adds delay, filters information, and creates work for the layers around it in the form of reporting, coordination and approval. Removing one shortens the path and, the argument goes, speeds up the organisation.

The other reason, usually less prominently stated, is cost. Management layers are expensive, and removing one produces a visible saving immediately.

These are different objectives and they favour different designs. A delayering aimed at speed targets the layers that add delay. One aimed at cost targets the most expensive layers. They are not necessarily the same, and being unclear which is being pursued is why some delayering exercises produce savings and no improvement in pace.

The layer being removed was doing something. Some of it was coordination overhead created by the existence of the layer itself, which genuinely vanishes. The rest does not.

• Approving, checking and prioritising, which now falls to a manager with twice the span.
• Developing people, which is the first thing dropped when a manager's span doubles.
• Absorbing and translating information between levels, which now happens less or not at all.
• Handling the exceptions and disputes that never reach a written process.

Where this work is not deliberately reassigned, automated or genuinely stopped, it lands on the surviving managers and on the people below them, uncounted and unacknowledged. The organisation records a headcount saving and absorbs an unmeasured workload increase.

The discipline that makes delayering work is doing the analysis first: list what the layer does, decide for each item whether it stops, moves up, moves down or is automated, and then remove the layer. Organisations that remove first and discover afterwards usually rebuild something like the layer within two years.

Delayering necessarily widens spans of control, and span is a constraint on attention rather than on capacity.

The right span depends heavily on the work. Highly similar, well-proceduralised work supports very wide spans. Varied, judgement-intensive work does not, and a wide span there means the manager is unavailable when judgement is needed.

The cost of a wide span rarely appears in output figures in the first year. It appears in development that stops happening, in problems raised late because there was no easy moment to raise them, and eventually in attrition among people who felt unmanaged.

Removing layers removes rungs, and this is the consequence organisations most often fail to plan for.

A structure with six levels offers five promotions between the front line and the top. A structure with four offers three. The distance between rungs is now larger, promotions are rarer, and the people most affected are the ambitious ones the organisation most wants to keep.

Telling employees that the organisation is flatter and more empowered does not address this. The question they are asking is what happens next for them, and the honest answer after delayering is that the next step is bigger and further away.

• Build lateral moves and scope growth as genuine progression rather than as consolation.
• Decouple pay progression from level, so someone can grow substantially within a level.
• Create specialist or technical tracks that reach senior pay without requiring management.
• Be honest in recruitment about what the structure offers, since discovering it later drives early attrition.

Without these the flat structure exports its ambitious people to organisations with more rungs, which is a slow and expensive way to lose the benefit of the exercise.

Delayering is frequently reversed within a few years, and not because the original analysis was wrong.

Layers exist for reasons: spans became unmanageable, a manager needed a deputy, a function grew, someone needed a promotion and a level was created to give them one. Those pressures survive the restructure and reassert themselves.

The regrowth is usually incremental and unremarked. A team lead here, a deputy there, a coordinator role that becomes a management role. Nobody decides to add a layer; one accumulates.

Organisations that hold a flat structure do so by managing the causes rather than by repeating the exercise: keeping promotion decoupled from creating a supervisory role, resisting the deputy pattern, and reviewing the structure periodically rather than only during a cost programme.

If the underlying pressures are not addressed, delayering is a periodic correction rather than a change of shape, and each cycle carries the full cost of restructuring.