The blunt test of a COO succession plan is simple: if you were unreachable for the next six weeks, would operations hold? If the honest answer is no, you do not have a succession plan. You have a single point of failure with your name on it.

Most COOs treat succession as a retirement problem — something to sort out three years before they hand in the badge. That framing is why so many companies get caught out. People leave for the wrong reasons at the wrong times: a competitor's offer, a health scare, a family move, a board that wants a change. A plan built only for a graceful, scheduled exit is useless the moment the exit is neither graceful nor scheduled.

This guide treats succession as an operating discipline you run continuously, not a document you write once. The goal is not to name your replacement. It is to build a system where the operation keeps its shape and momentum whoever sits in the chair — including a stranger who has never met your best supplier.

Two plans, not one: emergency cover and planned handover

The first mistake is trying to solve two different problems with one document. Emergency cover answers "who runs operations tomorrow morning if I disappear tonight?" Planned handover answers "who becomes the next COO over the next 18 to 36 months, and how do we get them ready?" They need different owners, different depth, and different triggers.

Weak succession planning collapses these together and, in practice, does neither. Strong planning keeps them separate. Emergency cover is short, blunt, and current: a named interim, their decision authority, the passwords and approvals they need, and a one-page brief on the three decisions that cannot wait a week. Planned handover is longer, developmental, and reviewed quarterly. If your "plan" is a slide deck the board saw once and nobody has opened since, you have a planned-handover artefact and zero emergency cover — the exact gap that hurts most.

A useful test: ask your finance lead who signs off a large unplanned payment if you are on a plane and out of contact for 48 hours. If they pause, your emergency layer is theoretical. Pair this work with a proper business continuity plan so the operational side — systems, vendors, escalation paths — is documented alongside the people side.

Spot successors on trajectory, not on last year's results

Identifying candidates is where most plans quietly go wrong. The obvious move is to rank your direct reports by current performance and anoint the top one. But the COO role is not "the best functional lead, promoted." A brilliant head of supply chain who cannot chair a cross-functional argument, or a sharp finance operator who has never owned a P&L trade-off, is a strong performer and a weak COO candidate.

Assess for the specific demands of the seat, and score trajectory over status. Strong candidate identification looks at three things at once, and weighs how fast someone is closing the gap, not just where they stand today.

Run this as a calibrated exercise, not a solo judgment. Pull in the CEO and, ideally, one board member so the read is not just your personal chemistry with a candidate. Deepen the pipeline through a real talent development program rather than betting everything on one favourite — a single-successor plan has the same fragility problem you are trying to solve, just moved one level down.

Develop through real accountability, not shadowing

The default development plan is "have them shadow me and go on a leadership course." Shadowing teaches candidates how you do the job. It does not teach them to carry the weight of a decision that is theirs to get wrong. Development that transfers judgment puts the candidate on the hook for outcomes with room to recover from mistakes.

Strong development is a deliberate escalation of real accountability. Give a candidate a genuinely hard, cross-functional problem to own end to end — a facility consolidation, a systems migration, a margin-recovery push — with a real budget, a real deadline, and the authority to make the calls. Then let them run it while you resist the urge to rescue. Rotate them through the functions they have never carried, because a candidate who has only ever run operations will freeze the first time procurement, HR, and IT are all on fire at once.

Two moves make development real rather than ceremonial. First, hand over a live relationship you personally hold — your toughest supplier, your most demanding internal partner — and let the candidate own it for a quarter. Second, put them in front of the board on a substantive item, not as an observer but as the person answering the hard question. Both expose the gaps that a training room never will. Strengthen the board-facing muscle deliberately, because it is the most common failure point for internal candidates who have spent their careers one level below the room; board communication skills are learned in the room, not before it.

Transfer the knowledge that lives only in your head

Every operation runs on undocumented knowledge: why you never push that vendor on price in Q4, which manager needs the decision framed a particular way, the real reason a process step exists that looks redundant on paper. This is the tacit layer, and it is exactly what a successor lacks and cannot Google. Weak knowledge transfer produces a binder of standard operating procedures nobody reads. Strong transfer captures the reasoning behind the rules.

Do not try to document everything — you will run out of energy before you run out of process, and most of it does not matter. Document the decisions that are expensive to get wrong and non-obvious to an outsider. A practical method: keep a running "decision log" for a quarter where you note the genuinely hard calls, who was involved, what you weighed, and what you decided. That log teaches judgment in a way a procedure manual cannot, because it shows the trade-offs, not just the answer.

Map decision rights explicitly so a successor knows what they can decide alone versus what needs the CEO or board. A RACI map across your top 20 recurring decisions removes the single biggest source of early-tenure paralysis — a new COO who does not know whether they are allowed to approve something and burns two weeks finding out. Pair the map with warm introductions: a successor inherits your authority but not your relationships, and a supplier or regulator who trusted you personally has to be re-earned face to face.

Time the handover in overlapping phases

A clean handover is not a single day when you clear your desk. It is a phased transfer of responsibility where authority moves before you do, so problems surface while you are still there to help. The specific months matter less than the sequence, and the sequence should always front-load real authority, not delay it until the end.

• Readiness (roughly 18 to 24 months out): candidate identified, development plan running, gaps named honestly.
• Delegation (12 to 18 months): the candidate owns whole areas outright, and you stop being the escalation point for those areas.
• Shadow-in-reverse (6 to 12 months): the candidate makes the calls and you observe, rather than the other way around; you intervene only when something is about to break.
• Public transfer (3 to 6 months): the organisation, the board, and key partners are told, and the candidate is visibly the operating lead.
• Exit (final weeks): you are available for questions but no longer in the decision path.

The most common timing error is compressing everything into the last three months because the plan lived on paper until the exit became real. A structured transition plan with dates attached forces the early phases to actually happen. If the transition follows a departure you did not plan, run the same sequence at speed — the order still holds even when the calendar does not.

Keep the CEO relationship and the board in the loop

Succession is not an operations project you run alone. The COO seat exists in a partnership, and a successor who clicks with you but grates on the CEO will not last. Weak plans surprise the CEO with a finished decision. Strong plans build the CEO's confidence in the candidate over months, through exposure and shared wins, so the eventual handover is a confirmation rather than a leap of faith. Treat the COO and CEO partnership as something the successor has to grow into, and engineer the situations where that trust gets built.

Give the board real visibility without handing them the pen. Directors should know your emergency cover exists, see the pipeline develop, and meet the leading candidate — but the operational development is yours to run. A board that hears "we have a named interim and two candidates 18 months out" sleeps better than one told "we'll sort it when the time comes," and that confidence is worth protecting for its own sake.

Key takeaways

• Run two plans: blunt emergency cover for a sudden exit, and a developmental handover plan for the planned one. One document cannot do both jobs.
• Assess candidates against the demands of the COO seat and weigh trajectory over last year's results — the best functional lead is not automatically the best COO.
• Develop through real accountability — owned budgets, live relationships, board-facing items — not through shadowing and courses.
• Document the reasoning behind hard decisions, not every process. A decision log transfers judgment; a binder of procedures does not.
• Move authority before you move on, in overlapping phases, so problems appear while you are still there to help.
• Build the CEO's and board's confidence in the successor over months, so the handover confirms a known quantity rather than gambling on a stranger.

Frequently asked questions

How far ahead should COO succession planning start? The developmental handover plan should be running 18 to 36 months before an anticipated transition, because building genuine operating range and board credibility takes years, not months. Emergency cover, by contrast, should exist right now regardless of any planned exit — it is the piece most companies skip and the one that hurts most when a departure is sudden. Should we promote from inside or hire externally? Keep both live rather than committing early. Internal candidates bring institutional knowledge and existing relationships but can be blind to their own gaps; external candidates bring fresh perspective but need time to learn the operation and re-earn trust. Running an internal development track and knowing the external market at the same time gives you a real choice when the moment comes, instead of a forced hand. What is the biggest mistake COOs make with succession? Betting everything on a single favourite and starting too late. A one-candidate plan carries the same single-point-of-failure risk you are trying to remove, and a plan that only wakes up three months before the exit skips the phases where authority is safely transferred. The fix is a genuine bench and a running clock, not a name written down. How do we transfer knowledge that only exists in the COO's head? Do not try to document everything — capture the decisions that are expensive to get wrong and non-obvious to an outsider. A running decision log that records the hard calls, who was involved, and the trade-offs weighed teaches a successor judgment in a way a procedures manual never can. Add explicit decision-rights mapping and warm introductions to the people whose trust was personal to you. What role should the board play? The board provides oversight and confidence, not day-to-day management of the process. Directors should confirm that emergency cover exists, watch the pipeline develop, and meet the leading candidate, while the operational development stays with you and the CEO. A board that has seen a real bench form is far calmer in a crisis than one hearing about succession for the first time under pressure. How do we know a successor is actually ready? Readiness shows up in behaviour under real conditions, not in assessment scores. The clearest signals are a candidate who has owned a hard cross-functional problem end to end, held their own with the board on a substantive item, and managed a difficult external relationship without you in the room. If they have done those things and the operation held, they are ready; if they have only shadowed and studied, they are not yet tested.