Ask an HR leader whether they have a succession plan and most will say yes. Ask when it was last opened for a reason other than a board meeting, and the room gets quieter. The gap between those two answers is where succession plans fail: not dramatically, but by decaying in a shared drive while the org moves on without them.
The stakes of that decay are not abstract. Research published in Harvard Business Review estimated that badly managed CEO and C-suite transitions destroy close to a trillion dollars of market value a year in the S&P 1500 alone, and that better succession practice would lift company valuations 20 to 25 percent. The failure modes behind those numbers are consistent enough to catalogue. Here are the five that account for most of the wreckage, and the habits of the plans that survive.
Failure mode 1: The plan is a spreadsheet nobody owns
The typical succession plan is a spreadsheet built under deadline pressure, and spreadsheets have a specific decay curve. Eight people have edit access, so nobody is responsible for accuracy. There is no history, so when a readiness call changes from "1 year" to "Ready Now," no one can say who changed it, when, or on what evidence. Columns get added until the file becomes unreadable, then someone makes a clean copy, and now there are two truths.
Within a year the plan describes an org that no longer exists. Leaders sense it before they can prove it, and the sensing is enough: when a real vacancy hits, they make the decision from memory and relationships, exactly as they would have without a plan. The document didn't fail loudly. It just stopped being consulted.
The fix is ownership plus provenance. One named owner per plan. Every change attributed and dated. Whether that lives in disciplined process or in software with an audit trail matters less than that it exists at all; what kills the spreadsheet is that nobody can tell which numbers are still true.
Failure mode 2: Assessments without criteria
Open a failing plan and read the successor column. The names are there, but ask why those names and the answers are adjectives: strong, impressive, ready with support. Nobody wrote down what the role requires, so nobody could score candidates against it, so the plan records opinions formed in hallways.
Opinion-based plans fail two tests that matter:
- The challenge test. When a skipped-over leader asks "why him and not me?", an adjective-based plan has no answer that survives the meeting. A criteria-based plan does: here are the role's weighted dimensions, here are the scored gaps, here is the development that would close yours.
- The turnover test. When the CHRO who held the opinions leaves, opinion-based plans leave with them. Criteria stay.
The corrective is boring and effective: a success profile per critical role, scored with evidence. It converts succession from a debate about people into a comparison against a standard, which is a debate an organization can actually finish.
Failure mode 3: The plan is a secret
Some plans fail from over-sharing, but far more fail from the opposite: the succession plan is treated like a classified document, and the people named in it are the last to know.
The failure sequence runs on a delay. Your Ready-in-1-Year successor for the VP of operations doesn't know she's on the plan. What she knows is that her last two raises were fine, her scope hasn't changed in two years, and a recruiter just described a bigger job with her name on it. She resigns. The plan, consulted at last, now lists a successor who works somewhere else.
Successor attrition is the specific tax secret plans pay, and it compounds: each departure of a named-but-never-told successor also removes cover from every other role that person backstopped.
The survivable version doesn't require publishing the org's depth chart. It requires that people carrying succession weight experience investment they can feel: stretch assignments, executive sponsorship, honest conversations about trajectory. People stay for a future they can see. They don't stay for one that exists only in a locked file.
Failure mode 4: Development without deadlines
Even plans with good criteria and honest scores can die in the development column, where gaps get diagnosed and then treated with intentions. "Needs more financial exposure." "Should build board presence." No assignment, no date, no owner. Twelve months later the same gaps appear in the same review, reworded.
A readiness window with no development engine underneath it is a forecast of nothing. "Ready in 1 year" is only true if something happens during the year: the budget cycle they run, the escalation rotation they hold, the board session they present. Remove the assignments and "Ready in 1 year" silently becomes "Ready in 1 year, indefinitely," which is the same as not ready.
The tell is easy to audit in your own plan right now: count the development actions that have a completion date and a named owner. If the answer is close to zero, the plan's readiness projections are decorative.
Failure mode 5: Nobody stress-tests the plan
The subtlest failure: a plan where every role has a named successor, reviewed and scored, that still collapses on first contact, because the roles were only ever planned one at a time.
Real departures chain. The COO leaves; the plan promotes the VP of supply chain; her seat is now open, and her own successor turns out to be the same person another plan was counting on for the plant-director role. One resignation, three exposed seats, and the discovery happens live, during the worst week to discover anything.
Boards know this exposure intimately: in the NACD's 2024 public-company survey, 30% of directors admitted their board had no successor identified for the CEO, and another 10% didn't know. Single-role planning can't see the chained version of this because the exposure lives between the plans, not inside any one of them. Stress-testing finds it early: model the departure, follow the cascade two or three levels down, and watch which plans go thin. Do it for the five roles whose loss would hurt most, and do it again after every reorg, because reorgs rewire the cascade paths and quietly invalidate last quarter's test.
The orgs that do this treat it like a fire drill: slightly tedious, occasionally alarming, and the reason the real event is a procedure instead of a panic.
What the survivors do differently
The plans that hold up share a short list of habits, none of them exotic:
- One owner per plan, with changes attributed and dated and a clear split of who owns what between HR and the business, so the document stays trustworthy enough to consult.
- Weighted, observable criteria per role, so assessments survive challenge and personnel changes.
- Readiness windows someone would act on, where Ready Now means "we'd hand over the job," not "we'd feel bad running a search."
- Development with dates, treated as the engine of readiness rather than a wish list.
- Stress tests on a schedule, because bench depth that hasn't been tested against a cascade is a rumor.
- A scoreboard, so decay shows up as a metric trending the wrong way before it shows up as a failed transition.
- A calendar that enforces the cadence: quarterly touches, twice-yearly deep reviews, immediate reruns on trigger events.
None of this requires believing in succession planning as a philosophy. It requires treating the plan as an operational system with maintenance requirements, the way you already treat payroll or security access. Systems with owners, evidence, and review schedules keep working. Documents with none of those things become archaeology, and every unplanned vacancy digs them up just long enough to confirm it.