A reorganisation is the most available instrument a senior leader has. It requires no budget approval, produces immediate visible activity, and demonstrates that something is being done.
It is also one of the most expensive things an organisation can do to itself, and the cost is almost never counted.
For a practical reference on time tracking software, Monitask provides the website.
What a reorg can actually fix
A short list, and worth being precise about it.
Misaligned incentives between groups. Two teams whose objectives pull against each other will fight regardless of how reasonable the people are. Restructuring so that the conflict sits inside one accountable unit genuinely resolves it.
Span of control that has broken. A manager with twenty-two reports cannot manage. This is arithmetic. See span of control.
A missing function. Nobody owns something that needs owning.
Genuine strategy change. The company is doing something materially different and the structure describes the old thing.
A capability that must be centralised or must be distributed, and currently is the wrong one.
What it cannot fix, and is frequently used for
Unclear decision rights. The most common misdiagnosis. Moving the boxes does not establish who decides; it establishes new boxes with the same ambiguity, and the same arguments resume with different participants. See decision rights.
A performance problem in one person. Restructuring around an individual is a way of avoiding a conversation, and everyone can see it.
Poor communication. A reorg reliably makes this worse for six months.
Strategy that has not been decided. Structure is downstream of strategy. Reorganising in the absence of a decision produces a structure that describes nothing.
Cultural problems. Structure shapes behaviour slowly and weakly compared to what leaders reward and tolerate.
The need to demonstrate action. Honest to name, because it is a real driver. A new leader under pressure to show change has an instrument that produces visible change immediately, and it will be used unless something stops it.
The costs nobody counts
Relationship rebuild. Working relationships are built over months and reset in an afternoon. Everyone spends the next quarter working out who to ask, who decides, and who to trust. This is by far the largest cost and it appears in no budget.
Institutional memory scattered. People who understood why something was the way it were are now somewhere else. The new team makes decisions without the reasoning behind the old ones, and reverses things deliberately done. See documentation as infrastructure.
Work in flight. Projects mid-delivery get re-scoped, re-owned or dropped. The sunk cost is real and invisible.
Attrition. Reorgs cause departures, disproportionately among people with options, and the departures arrive two to four months later where they are attributed to something else.
The pre-announcement period. Rumours circulate for weeks before an announcement and productivity falls throughout. People update their resumes. This cost begins before the decision is communicated and is routinely underestimated.
Manager time. Weeks of it, at every level, on an activity that produces nothing for a customer.
The next one. Frequent reorgs teach people that structure is temporary and not worth investing in. After the third, people stop rebuilding relationships and start waiting.
If you are going to do one
Establish what problem it solves, in one sentence, before anything else. If the sentence is about clarity, alignment or agility, it is not a sentence — write a specific one or do not proceed.
Check it against the list above. If the problem is decision rights, communication, strategy or one person, a reorg will not fix it and will cost you.
Do it once and completely. Serial partial reorgs are worse than one large one, because the disruption cost recurs and the stability never arrives.
Announce fast after deciding. The gap between decision and announcement is when the rumour damage happens, and it is entirely under your control.
Say what does not change. People assume everything is in play. Naming what is stable removes most of the anxiety at no cost.
Name decision rights explicitly in the new structure, since ambiguity is what people are actually worried about and what most reorgs fail to resolve.
Protect work in flight. Decide explicitly what continues and who owns it, rather than letting it fall between the old and new structures.
Do not do it more than roughly every two years. Below that interval the organisation never reaches the stability where a structure pays off.
Measuring whether it worked
Almost nobody does this, which is why reorgs continue to be used for problems they do not solve.
Before: record time-to-decision on a few typical decisions, attrition, and the specific problem you claimed to be solving.
Six months after: measure the same things.
If the stated problem is unchanged, the reorg did not work — and that is worth knowing before the next one is proposed as the solution to the same thing.
The uncomfortable question
Before proposing one, ask honestly:
Would I be doing this if I had to explain the cost in departures and lost momentum, and if someone would check in six months whether it worked?
Reorgs survive as an instrument largely because neither of those things happens.
For broader public guidance and background, consult the OECD.