Tenure-long performance data cannot distinguish leader quality from inherited headroom, which means every internal-versus-external appointment policy built on such data rests on a confound its own evidence exposes.
Most boards hold two convictions about appointments at once and defend both with performance evidence. Insiders should be preferred: they know the customers, the plant and the informal decision rights, and do not spend two years learning what everyone else already knows. And when a business is genuinely in trouble the answer is somebody from outside, unattached to the decisions that produced it and willing to say so.
The tension is not that these are contradictory as principles. It is that they are supported by the same body of evidence, and that evidence cannot adjudicate between them.
A widely circulated ranking of chief executives by tenure-long shareholder return, built precisely to avoid reputation and celebrity, reports two findings that pull against each other and are never reconciled. [FACT CHECK REQUIRED] Insiders outperform outsiders, substantially. And leaders who inherit poorly performing organisations post the largest gains, because a business below its potential offers room that a business at its potential does not. Add a fact neither disputes — outsiders are disproportionately appointed to organisations already in difficulty — and the second finding predicts outsiders near the top of the ranking. They sit lower.
Something has to give, and what gives is the measurement. Both readings — that the insider advantage is understated, and that the headroom finding is an artefact of counting only the survivors — fit the same numbers. A measure supporting opposite conclusions with equal comfort is not weak but uninformative.
This matters well below chief executive level. The same confound sits inside the record of every division head, plant manager and programme director an enterprise is about to promote, and nothing in the succession paperwork corrects it.
The Strategic Context
Enterprises make three consequential decisions off tenure performance: whom to promote, whether to look outside, and what to pay. All three read a leader-attributable quantity off an outcome that contains the starting condition.
The published rankings do adjust, carefully. They strip out the national market and the sector, on the sound reasoning that a leader should be neither credited for a rising tide nor punished for a falling one. What they do not strip out is the condition of the enterprise at handover: the order book, the margin trajectory, the state of the plant, the unresolved disputes, the distance between where the business sat and where it could reasonably be run. That is the largest single term for any one business over any one tenure, and the appointing body sets it by choosing when to appoint.
The same structure repeats one level down. A programme director inheriting a recovery has measured headroom that one inheriting a well-run programme does not, and a portfolio-level talent decision reading raw delivery outcomes will prefer the recovery. Nothing in the portfolio's reporting flags the difference, because it was built to track delivery, not to attribute it.
What Leaders Commonly Misread
The first misreading is that adjusting for market and sector is the same kind of act as adjusting for starting condition, only more of it. Those adjustments remove influences external to the enterprise; inherited condition is internal, specific, and usually the dominant driver of how much movement was available at all. Removing the smaller confounds and leaving the largest produces a number that looks rigorous and is not.
The second misreading is that the insider-versus-outsider question is about people. It is about assignment. Insiders and outsiders are not appointed into the same distribution of situations, so their outcome distributions are not comparable and no statistical care makes them so. An enterprise reading such a comparison as a fact about candidate types has read its own allocation policy and mistaken it for a finding about people.
The third misreading is that the direction of the bias is obvious once pointed out, and this is where most executives stop too early. If distress genuinely creates room, outsiders were advantaged and still finished behind, so the insider case is stronger than the headline. If distress mostly persists and only survivors are counted, the headroom effect is largely selection, and the case for appointing outsiders into trouble rests on very little. The evidence licenses both, and choosing requires information the ranking does not contain and your own records do not contain either.
Reframing the Issue
The productive question is not which category of leader performs better, but what the enterprise's own appointment rule generates as evidence about itself.
Consider the loop. A board appoints outsiders mainly when things are bad, when the internal bench looks compromised and the appetite for discontinuity is highest. Outsiders therefore accumulate a record built disproportionately in hard conditions. The board later reviews that record, finds outsiders underperform, and concludes outsiders should be appointed only when things are bad. Every participant behaves reasonably, and the conclusion is unfalsifiable because the policy manufactures the data that confirms it.
Take a hypothetical industrial automation systems integrator with two divisional managing directors appointed in the same quarter. One inherited a full order book, a stable engineering team and a completed systems migration; the other a disputed major contract, the departure of two principal engineers and a migration abandoned halfway. Three years on, the second has posted the larger improvement in margin and the first the higher absolute margin. Which is the better leader, and which should run the combined division? The record answers neither, and the promotion is made on it regardless.
Why the Adjustment Stops Where It Does
The adjustments that get made are the ones somebody else computes
Country and sector adjustments are made because indices exist: published, standardised, cheap to subtract. Inherited condition has no index. Nobody publishes a series for the state of this division on the day its new leader walked in, so nothing gets subtracted — not because the term is small, but because no external party supplies the number. Measurement stops at the edge of the conveniently available data, and enterprises mistake that edge for the edge of the effect.
Headroom is the largest term, not a nuisance variable
For one business over one tenure, the distance between where it sat and where it could reasonably be run dominates measured improvement. A leader taking a business near the top of its operating envelope can be exceptional and post a flat line, because the available movement was small before they arrived. A leader taking one well below its envelope can be unremarkable and post a strong recovery, because reversion does part of the work unaided. Percentage improvement and absolute level are different claims, and boards accept one as evidence for the other several times a year.
Pricing an inherited condition in delay terms is a related exercise this article does not attempt, because [Related article: Every Day of Delay Has a Price] owns it — for any asset generating revenue, schedule and cost are not comparable constraints, since every day of delay carries a computable price at full margin. The concern here is attributing the inherited condition, not costing it.
The appointing body is inside its own data
The board chooses when to appoint and whom, and that choice sets the starting condition, so any correlation between appointment type and outcome is partly a record of the board's own timing and nerve. A performance record containing the board's appointment behaviour cannot evaluate the board's appointment behaviour, which is exactly the use it is put to in most succession reviews.
A hypothetical national sporting body shows the shape of it. After governance failure and falling participation, the board recruits an outsider as chief executive; three years later participation has recovered and the appointment is recorded as a success. But the comparison set is other national bodies never in crisis, and the board evaluating the appointment is substantially the one whose oversight produced the crisis. The verdict is not wrong so much as unattributable, and the body will use it to set appointment policy for a decade.
Decision Framework
The inherited-condition control. It is a handover instrument, not an analytical technique, and works only if applied before the incoming leader has done anything.
Band the inheritance at handover. Score the unit on six dimensions using evidence dated before the handover: demand against capacity, margin trajectory across preceding periods, capital and asset condition, key-person exposure, unresolved legacy commitments such as disputes, remediation and deferred maintenance, and distance from the unit's demonstrated operating envelope.
Assign one of four bands, and sign it. Outgoing leader, incoming leader and appointing executive all sign within thirty days of handover, and the band cannot be revised afterwards. If the three cannot agree, that disagreement is the finding and goes in the file.
| Band | Condition at handover | What good performance looks like | What the record will not tell you |
|---|---|---|---|
| Distressed | Below the demonstrated envelope on three or more dimensions | Recovery to the envelope with the causes of the fall closed out, not deferred | Whether the recovery was leadership or reversion |
| Stable | At the envelope, no material legacy commitments | A durable extension of the envelope, not one strong period | Whether the extension survives the leader |
| Performing | Above prior envelope, momentum inherited | Continuation without erosion, plus a structural change that outlasts the tenure | Whether the momentum was the predecessor's |
| At ceiling | At or near the envelope on four or more dimensions | Holding the line while building the next envelope, invisible in-period | Almost nothing, from in-period numbers alone |
Set the evaluation baseline at appointment, not at review. Write down, before the leader starts, what a competent steward should achieve from that band over that horizon. A baseline written afterwards is a description of what happened.
Apply the admissibility rule. No performance record enters a promotion, succession or appointment-policy discussion without a band attached. A record without a band is evidence about the enterprise's assignment rule, not about the person, and should be described that way in the paper.
Audit the assignment rule annually. Cross-tabulate handover band against whether the appointee was internal or external. If external appointments concentrate in the Distressed band, the enterprise's outsider record is uninterpretable and every paper citing it must say so on its first page.
From Strategy to Execution
Immediate. Retrospectively band every unit that changed leadership in the last two years, using only evidence dated before each handover. Where that evidence does not exist, record the band as unavailable rather than estimating it, and treat the record as inadmissible. Expect the unavailable cases to include your most contested promotion.
Medium term. Make banding a condition of handover, with the succession paper template carrying the band, the pre-agreed baseline and the admissibility statement as required fields. What the verdict should say is settled by this instrument; who delivers it to the individual, and what the choice of messenger does to whether it is believed, is treated separately in ERANORTH's article on who delivers the verdict.
Long term. Change what taking an At-ceiling unit does to a career. If the only way to be seen is to be handed a recovery, capable people will decline stewardship roles and the units closest to their potential will be run by whoever could not refuse them. Whether the enterprise can generate enough internal candidates to have a genuine choice is a separate constraint, and [Related article: Your Pipeline Is Capped by Your Portfolio, Not Your Budget] owns it — the number of leaders an enterprise can grow is capped by the composition of its work portfolio rather than by its development budget.
Signals to Monitor
Watch for internal candidates quietly negotiating for the troubled unit rather than the sound one; that preference is a measurement finding, not an appetite for challenge. Watch for "turned it around" entering promotion papers with no stated starting point, and track whether external appointments are concentrating further into distressed units. Watch attrition where units sit at the top of their envelope, since good stewards leave first and quietly. And watch a unit in the two years after its celebrated leader departs, because a legacy that evaporates on schedule was inherited momentum being spent. Whether the capability such a leader built survives them turns on a different question — [Related article: Which Half of Your Capability Programme Has an Internal Buyer?] owns it, since you can predict which parts of a capability programme survive by asking which have an internal buyer with a commercial reason to keep delivering them.
Questions for the Leadership Team
- For each of our last six senior appointments, what evidence dated before the handover describes the condition of the unit the appointee took on?
- Cross-tabulated against that condition, where have our external appointments actually been made, and does our internal record of outsider performance survive the comparison?
- Which of our current leaders inherited a unit already at the top of its envelope, and how does that show up in how they are assessed and paid?
- What did we write down before each appointment about what competent stewardship would achieve, and if nothing, what are we comparing outcomes against?
- In the two years after our most admired leader left their previous unit, what happened to it, and who tracked that?
- Which promotion in the past three years would have gone the other way if both candidates' inherited conditions had been banded and signed at handover?
Closing Perspective
The choice here is narrower than it looks. An enterprise either records the condition of a unit at the moment its leader takes it, in a form that cannot be revised later, or accepts that its appointment policy is a preference wearing the clothes of evidence. There is no third option in which the raw track record is rehabilitated by seniority or by the confidence of whoever presents it.
That places the responsibility on those who commission the papers, not those who write them. A board told its evidence cannot separate leadership from headroom, and which goes on citing the same track record in the same terms, has decided to be persuaded rather than informed — and will keep appointing the people its own measurement error has been quietly selecting for years.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.