An organisation can teach that good leadership depends on the situation and still promote on a number that says otherwise. The number wins.
A capability budget is approved. It funds a development programme for two hundred supervisors, a facilitated workshop series, and an assessment instrument that produces a profile for each participant. The programme is well regarded. The assessment is completed in forty minutes, produces two scores, and the scores go into the talent file.
Three years later the organisation has a promotion pattern it did not intend and cannot easily explain. The people advancing are not obviously the best leaders in the business; they are the ones who scored well. And what the score rewards turns out to be different — subtly but consistently — from what the programme taught.
Nothing improper has occurred. The instrument was applied consistently, the panel was fair, and the training was sound. The organisation simply held more than one theory of leadership at the same time, and only one of them could be recorded in a system. That one became policy.
This is a capability-investment problem rather than a human-resources one, and it belongs to executives for a specific reason: leadership instruments outlive the training that introduced them, and they keep making decisions long after everyone has forgotten what they were for.
The Strategic Context
Leadership development is one of the few enterprise investments where the measurement instrument is bought at the same time as the intervention and is rarely evaluated separately. A programme has a curriculum, a provider and a cost. The instrument arrives with it, appears to be part of the package, and then persists.
It persists because it produces data. Once a score exists, it can be filed, compared, tracked and cited — and an organisation with two hundred supervisors and no other comparable measure will use it, whatever caveats accompanied it originally.
That makes the choice of instrument a strategic decision disguised as a procurement item. Whatever theory of leadership the instrument encodes will, over time, become the organisation's real theory of leadership, because it is the only one that reaches promotion, succession and pay.
Three Theories in One Room
The teaching material supplied for this analysis makes the incoherence unusually visible, because all of it sits inside a single sequence of slides.
Leadership as a set of characteristics. The deck presents habits of effective people, credited to Covey (1989), introduced as characteristics important for all team members, and a list of ten key aspects of successful leadership addressed to the leader personally — be a team builder, have a sense of humour, have vision.
Leadership as a type. Leaders are then divided into three: autocratic, laissez-faire and democratic, each described by what its holder characteristically does.
Leadership as a situation. Two slides later, the deck presents a model credited to Hersey and Blanchard (1977) in which style must adapt to the maturity of the person being led, moving between directing, coaching, supporting and delegating according to that person's competence and confidence.
These are not complementary emphases. The first two locate the answer in the person; the third locates it in the situation and says explicitly that a leader should shift. If the third is right, the second is a description of someone who has failed to adapt. Nothing in the material reconciles them, and there is no reason to think the authors intended anyone to notice.
Most enterprises are in precisely this position. They teach a contingency model in a workshop, describe leaders as types in succession discussions, and list required characteristics in a capability framework — with no view about which is true.
The Tie Is Broken by Measurement
A theory that produces a number can be recorded, compared across candidates and carried into a decision. A theory that says it depends cannot. That asymmetry decides which theory an organisation actually runs on, regardless of what it teaches.
The same course supplies the instrument, and it is worth describing precisely. Thirty-five items, almost every one beginning "I would" — for instance, "I would let some members have authority which I could keep." Each is answered on a five-point scale from always to never. A scoring key converts the answers into two totals: a score for concern for tasks and a score for concern for people. The participant plots both on a diagram and reads off a result. It is credited to a 1995 work on leadership, with no publisher given [SOURCE DETAILS REQUIRED]; two-axis self-scoring instruments of this shape circulate widely and long predate that date, and the version supplied cannot be traced to an origin from this material [FACT CHECK REQUIRED].
The diagram is the part that matters. High task concern is labelled autocratic leadership — high productivity. High people concern is labelled laissez-faire leadership — high morale. Scoring highly on both is labelled shared leadership — high morale and productivity, and it is unmistakably the destination.
So the instrument answers the question the contingency model on the neighbouring slide deliberately leaves open. There is one right place to be, it is the same place for everyone, and it does not depend on the team, the task or the week.
What a Score of This Kind Is Evidence Of
Three properties determine what such an instrument can carry, and each narrows it further.
It scores self-report. Every item asks what the respondent would do. It measures a person's account of their own intentions — a real thing, and not the same thing as behaviour. Two people with identical scores may have entirely different records, and the instrument cannot distinguish them because it never looks.
It scores against a fixed optimum. A leader running an inexperienced crew on safety-critical work, who directs closely and checks constantly, will score toward autocratic — while doing exactly what the contingency model on the same course prescribes for that situation. The instrument penalises correct behaviour, and the person who learns to score well learns to describe themselves as balanced rather than to read situations.
It has no validation in the material that supplies it. No norms, no reliability data, no evidence that the scores predict anything. That is not a criticism of the instrument's authors — a teaching handout is not a technical manual. It becomes a criticism of the organisation the moment the score is used to allocate something.
None of this argues for abandoning measurement. It argues for matching the weight of the decision to the strength of the evidence. A self-report instrument is a perfectly good basis for a coaching conversation, in which the participant is the only person who acts on it. It is a poor basis for a promotion, in which someone else acts on it and a third party bears the consequence.
Where This Argument Stops
Several adjacent problems are deliberately not this one.
This is not about who gets promoted into delivery leadership or what that transition demands of the individual — the move from specialist to leader is a career change with its own requirements, argued in full elsewhere [Related article: From Specialist to Delivery Leader]. It is not about the structural under-powering of the role once someone holds it [Related article: Accountability Without Authority]. It is not about the audience for an organisational maturity rating, which is a different instrument aimed at an institution rather than a person [Related article: Who Is Your Maturity Rating For?].
Nor is it a general argument about what governance instruments assume about the work they were built for [Related article: What Kind of Work Were These Instruments Built For?]. The object here is narrower: an instrument aimed at a person, the theory it encodes, and the decisions that ride on it.
One boundary needs stating directly. When a capability investment is justified by an instrument that grades individuals, there is a level-of-analysis problem — a personal score being used to argue an organisational case — and that argument, along with the question of what the enterprise retains from such spending, belongs to its own piece [Related article: What Does the Enterprise Own After a Capability Investment?].
Two Illustrations
Both are hypothetical.
A contact-centre operator develops team leaders across four sites. The instrument in use produces the two-axis profile, and its results are recorded in the talent system because that is where results go. Within two years, promotion correlates strongly with the balanced profile. The consequence surfaces during a system migration: the leaders who hold the sites together in the difficult fortnight are the directive ones, and several of them have been passed over twice. Nobody decided that directness would be penalised. An instrument decided it, quietly, at forty minutes per person.
A passenger ferry operator runs a leadership programme for masters and senior crew. The curriculum, correctly, is contingency-based: in a berthing manoeuvre or an emergency you direct, and in a maintenance planning meeting you consult. The assessment attached to the programme is the same two-axis instrument. Crew learn quickly that the profile to report is the balanced one, and the assessment stops measuring anything at all except a candidate's grasp of what the assessment wants. The organisation has bought a mirror and filed the reflections.
Decision Framework
Build an instrument register. Most organisations cannot list the tools currently used to judge their leaders, which is the first finding rather than an obstacle.
| Field | What it records |
|---|---|
| Instrument | What it is, who supplies it, when it entered use |
| Theory encoded | Trait, type, contingency, or something else |
| Evidence of what | Self-report, observed behaviour, outcome, or peer perception |
| Decision it feeds | Coaching, development, selection, promotion, pay |
| Validation | What, if anything, connects the score to observed performance |
Three tests follow.
The coherence test. Where two instruments in use imply different answers for the same person, the organisation should decide which is right rather than continue running both. Incoherence is not neutrality; it means the decision is being made by whichever instrument has a field in the system.
The weight test. Move each instrument up or down the decision ladder until the evidence supports it. Self-report belongs at the coaching end. Promotion decisions need something a third party observed.
The correct-behaviour test. Describe a situation your enterprise genuinely faces where a leader must be highly directive. Score that leader on your instrument. If they come out badly, you have found what your instrument actually rewards, and it is not judgement.
Whether the person reading the score can tell that a claim inside it is unsupportable is a separate capability question, and a real one [Related article: Enough Technical Depth to Test the Answer].
From Strategy to Execution
Immediately. List every instrument in use across the enterprise and the decision each one feeds. Where a self-report instrument feeds a selection or promotion decision, stop that use pending review. This costs nothing and is reversible.
Within the year. Decide, at executive level, which theory of leadership the organisation intends to operate on, and make the instruments follow. If the answer is contingency — as it is in most operational businesses, and as the material behind this article teaches on one slide and contradicts on another — then the assessment must record situations and observed responses, not dispositions.
Structurally. Build one source of observed evidence: structured observation during real events, feedback tied to specific situations, or a record of decisions taken under pressure. It is slower and more expensive than an inventory, and it is the only thing that answers the question a promotion panel is actually asking.
The instrument problem also runs in the other direction, and it is worth reading beside this one: a scoring model applied to competing offers makes its decisive judgements before any offer exists, in the same quiet way [Related article: Your Scoring Model Decided Before the Bids Arrived]. And any instrument's results are only as representative as the population that generated them [Related article: Volunteers Are Not a Sample].
Signals to Monitor
- Promotion patterns converging on a profile. If successful candidates increasingly share an assessment shape, the instrument has become the criterion.
- Scores improving across cohorts with no change in outcomes. People have learned the instrument, which is the expected response to any measure attached to a consequence.
- Development conversations conducted in the instrument's vocabulary. When a leader describes their own weakness using the tool's categories, the tool has replaced the observation.
- Instruments outliving their programmes. A tool still in use three years after the provider left, with nobody able to say what it validates against.
- Divergence between assessment results and operational judgement. Where line leaders privately disagree with the profiles, the organisation is already running two theories and has decided which one is written down.
Questions for the Leadership Team
- What instruments do we currently use to judge leaders, and what theory does each encode?
- Which of them feed a promotion, a succession decision or a pay outcome — and what validates them?
- Do our instruments and our training agree about whether leadership depends on the situation?
- Take a role where directive leadership is plainly correct. How would that leader score with us?
- What evidence do we hold about how our leaders have actually behaved, as distinct from how they describe themselves?
- If we removed every score from the talent file tomorrow, what would we use instead — and is that better?
Closing Perspective
An organisation's real theory of leadership is not in its capability framework. It is in whatever produces the number that reaches a decision.
Most enterprises have never chosen that theory. They bought a programme, accepted the instrument that came with it, and let the instrument's assumptions propagate into promotion, succession and pay for a decade. The result is a leadership population selected for a shape that nobody at the executive table would defend if it were stated aloud.
The correction is not to stop measuring. It is to know what each instrument is evidence of, to keep it at the level of decision that evidence supports, and to accept that the most useful thing an organisation can record about a leader — what they actually did when it was difficult — cannot be collected in forty minutes.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.