Research and insights from the Institute

The gap between authority and influence

Authority arrives with the announcement. Influence arrives later, if at all, and is granted by people who are watching more closely than most new leaders realise.

Roughly six weeks into a new role, most leaders discover something they were not warned about. The title changed. The reporting lines changed. What did not change is the willingness of the people around them to follow, and no amount of re-reading the org chart makes that willingness appear.

This gap accounts for a large share of first-year leadership difficulty, and it is rarely diagnosed correctly. It gets attributed to a skills deficit, or to a difficult team, or to the predecessor. More often it is structural. Authority is conferred by an institution in a single moment. Influence is conferred by individuals, repeatedly, over months, and it is revocable.

The trouble is that authority is available immediately and influence is not. A leader under pressure to demonstrate early progress will reach for the tool that is in the room. They issue direction. They escalate. They invoke the decision rights the role came with. In the short term this produces compliance, which is easy to mistake for followership because from a distance the two look identical.

They are not identical. Compliance produces the work that was asked for. It does not produce the caught error, the unsolicited improvement, the concern raised early enough to matter, or the extra hour of thinking someone gave a problem because they cared how it turned out. That category of effort is discretionary by definition, and it is offered to leaders people believe in rather than leaders people report to.

What actually closes the gap

Tenure closes it eventually. Three things close it faster.

The first is asking before directing, particularly in the opening quarter. Not as a consultative technique, which teams detect quickly and resent, but because a new leader genuinely does not yet hold the context the team holds. Directing before asking communicates that the new authority outranks the team's accumulated knowledge. Few messages are more expensive to send in week three.

The second is being visibly wrong once, deliberately and early. Teams calibrate a new leader partly by watching how that leader handles being incorrect. Someone who never concedes error reads as someone managing an image. Someone who says plainly that they got it wrong and here is what changes as a result reads as someone who can be trusted with a harder problem later. The second reading is worth considerably more than the first.

The third is spending early credibility on the team's priorities rather than the leader's own. A new leader who uses their initial standing to unblock something the team has been stuck on builds a reserve that can be drawn down later, when an unpopular call has to be made and will be made anyway.

Authority can be granted in a memo. Influence is granted one decision at a time, and the people granting it keep a longer memory than most leaders assume.

Back to index

Why capable people avoid the hardest conversations

Avoidance is usually described as a courage problem. It rarely is.

The manager who has postponed a performance conversation for four months is frequently the same person who will walk into a hostile client meeting without preparation, or contradict a peer in front of the executive team, or deliver bad numbers to a board without softening them. They are not flinching. They are doing arithmetic, and the arithmetic favours delay.

Consider what the two sides of that calculation look like. The cost of having the conversation is immediate, certain, and personally unpleasant: an uncomfortable hour, a damaged relationship, a possible argument, the risk of handling it badly. The cost of not having it is delayed, uncertain, and distributed across other people: the team that keeps absorbing the problem, the client who eventually notices, the individual who is denied the information they would need to improve. One of those costs lands on the manager's desk this afternoon. The other lands somewhere else, later, on someone else.

Framed that way, avoidance is not weakness. It is a rational response to an incentive structure that almost every organisation accidentally builds.

Why the arithmetic rarely corrects itself

Three features of ordinary organisational life keep the calculation tilted.

• The delayed cost is never attributed. When a weak performer finally exits eighteen months late, the organisation records a departure, not a management failure. The eighteen months of absorbed cost appear in nobody's assessment.
• The immediate cost is highly visible. A conversation that goes badly is witnessed, discussed, and sometimes escalated. A conversation that never happened is invisible by construction.
• Most managers have no rehearsal ground. They are expected to be competent at the single hardest interpersonal task in the role, having practised it never, under conditions where the first attempt is also live.

The common remedy (training managers in conversational frameworks) addresses the third point and ignores the first two. This is why frameworks tend to survive about a quarter. A manager who has learned a structure for difficult feedback but still operates in a system that never charges them for avoidance will use the structure when the conversation is easy and avoid it when it is not.

What changes behaviour

Organisations that get more candour out of their managers tend to have done something to the incentive rather than only to the skill. The specific mechanisms vary. What they share is that avoidance stops being free.

That can look like a manager's own review including whether their team knows where they stand. It can look like skip-level conversations that surface, routinely, whether feedback is actually reaching people. It can look like a senior leader visibly doing the difficult thing early, which resets what the surrounding managers believe is normal.

None of these are sophisticated. They are simply the difference between telling managers that hard conversations matter and building an environment in which not having them costs something.

Back to index

Promoted, or developed?

Most organisations select leaders using evidence of individual excellence, then express surprise when the collective work goes badly.

Promotion decisions are made on the strongest available evidence. That evidence is almost always a record of individual output: the account closed, the system shipped, the region turned around, the number delivered four quarters running. It is legible, it is defensible, and it correlates with almost nothing about whether the person can run a team.

This is not a criticism of the people promoted. It is an observation about what the selection process can see. An organisation cannot easily assess a candidate's capacity to develop others, because the candidate has not been asked to develop anyone. It cannot assess their judgement under ambiguity, because their previous role supplied structure. It cannot assess whether they can accept a worse outcome now for a better capability later, because nothing in an individual contributor role rewards that trade.

So the organisation selects for the thing it can measure and hopes the rest arrives. Sometimes it does. When it does not, the failure is usually described in terms of the individual (they were not ready, they could not let go, they micromanaged) rather than in terms of a process that never assessed the relevant capability in the first place.

The difference the promotion does not close

A manager who was promoted has been recognised for what they produced. A leader who was developed has been prepared for something structurally different: work whose output is other people's work.

That shift breaks several instincts that were previously assets. Speed becomes a liability when the fastest path is to do it yourself and the necessary path is to let someone slower learn it. High standards become a liability when they are applied to someone else's first attempt rather than to a finished product. Deep expertise becomes a liability when it makes a leader the bottleneck on every decision that touches their old specialism.

None of this is intuitive from inside the transition, which is why it is rarely self-corrected. The behaviours that stopped working are the same behaviours that earned the promotion, and abandoning them feels like abandoning the thing that made the person good.

What development adds

Development does not replace the promotion criteria. It supplies the part the criteria could not observe.

Practically, that means giving people leadership work before the title (a small team, a cross-functional effort, an ownership stake in something they cannot personally execute) and assessing how they handle it. It means being explicit that judgement, not expertise, is the currency of the new role. And it means providing a place where the transition can be discussed honestly, because the one thing a newly promoted leader rarely has is somewhere to admit that they are finding it difficult.

The organisations with strong leadership benches are not the ones that promote more carefully. They are the ones that stopped treating promotion as the development event.

Back to index

Delegation is not a time-management tactic

Delegation is usually taught as a productivity technique. Identify the work that does not require you, hand it off, recover the calendar, focus higher. Nothing in that description is false, and it produces a specific and predictable failure.

A manager delegating to clear their own plate will select for the wrong criteria. They hand off what is low-stakes, low-visibility and easy to explain, because that work is the least risky to lose control of. They retain what is consequential and judgement-heavy, because it matters too much to hand over. The result is a manager with a marginally lighter calendar and a team that has developed nothing, since the only work that develops anyone is the work that requires judgement.

The high-stakes work is not the work that is too important to delegate. It is the only work worth delegating.

Three distinctions that separate development from offloading

Delegate the decision, not only the task. Handing someone a fully specified task, build this report, to this format, by Thursday, develops execution. Handing someone the underlying problem, we have no reliable visibility into this, work out how to fix it, develops judgement. Teams stop growing when every piece of work arrives pre-decided, and they stop growing quietly, because the output still looks fine.

Accept a worse result the first few times. This is the part most managers will not do, and it is not irrational: a capable person given genuine ownership will occasionally produce something the manager could have done faster or better. That gap is the actual price of a team that does not need the manager in every decision a year from now. Managers who refuse to pay it are not being rigorous. They are financing this quarter's quality with next year's capability.

Delegate the visibility, not just the work. When a manager assigns real ownership but continues to field every upward question about it, the team learns that the ownership was decorative. Redirecting a senior stakeholder to the person who now owns the work is a small act with a disproportionate effect: it is the moment the delegation becomes real to everyone watching.

The measure that matters

The useful test of delegation is not how much time a manager recovered. It is what the team can now do without them that it could not do six months ago.

By that measure, a manager whose calendar is still full but whose team now handles three categories of decision independently has delegated well. A manager with two free afternoons and a team that escalates exactly as much as it did last year has simply moved some tasks around.

Back to index

The expert's first year in charge

The hardest part of the transition is not learning to manage. It is the quiet loss of the thing the person was good at.

An organisation promotes its strongest engineer, clinician, analyst or specialist into a management role and, in doing so, asks them to stop doing the work that made them valuable. This is usually framed as a step up. Experienced from inside, it frequently feels like a demotion in competence: yesterday the person was unambiguously excellent at something. Today they are a novice at something else, and the something else is much harder to get feedback on.

The first year tends to follow a recognisable arc.

Months one to three: doing both jobs

The new manager keeps a substantial share of technical work, partly because the team still needs it and partly because it is the only part of the week that feels productive. Management tasks get done in the margins. Nobody objects, because output holds steady and the transition looks smooth from above.

Months four to eight: the bottleneck forms

The technical work has not reduced but the management load has grown. Decisions queue behind the manager because they are still the deepest expert and everyone knows it. The team, reasonably, routes everything to the person most likely to give a correct answer. The manager experiences this as being extremely busy. The team experiences it as waiting.

This is the point at which the transition either turns or calcifies. The available correction is unattractive: deliberately step back from the technical work, allow decisions to be made less well by other people, and accept a visible dip in the quality of things the manager could have handled personally.

Months nine to twelve: which identity wins

Managers who make the turn generally did one specific thing: they found somewhere to be a beginner without it being career-threatening. A peer group, a coach, a programme, a senior leader willing to discuss the transition honestly. What matters is less the format than the existence of a space where the admission "I do not know how to do this part" carries no cost.

Managers who do not make the turn usually remain competent and are frequently well-regarded. They simply continue to run the team as its senior technical resource. The cost shows up indirectly: the team's most capable people leave for places where they can grow into the decisions that are currently being made above them.

What organisations get wrong

The common intervention is a management training course somewhere in month two, which arrives before the difficulty has been felt and is therefore heard as abstract. Support tends to be far more useful in months four to eight, when the bottleneck is real and the person actually wants an answer.

The second common error is leaving the expectation unstated. A great many technical managers have never been told explicitly that they are expected to stop being the best practitioner on the team. In the absence of that instruction, they reasonably assume the opposite.

Back to index

What performance reviews actually measure

A review process is designed to measure contribution. What it reliably measures is legibility: how visible, well-documented and easily described someone's contribution was to the person writing the review.

These two things overlap. They are not the same, and the gap between them explains most of what frustrates people about review season.

Work that is legible has clear boundaries, a named owner, a measurable output and a moment of completion. Work that is illegible is diffuse: the problem that never became a crisis because someone quietly handled it, the colleague who became effective because someone spent unrecorded hours helping them, the bad decision that was prevented in a corridor. The first category writes itself into a review. The second has to be remembered by someone who was not there.

Organisations therefore end up rewarding a particular profile (people whose contributions are naturally well-shaped and self-documenting) and under-rewarding a different one, which includes a disproportionate share of the people holding teams together.

The predictable second-order effect

People notice what the process rewards, usually faster than leadership expects, and adjust. The adjustment is not cynical in most cases. It simply becomes rational to prefer work that will be visible at review time over work that will not, and to prefer starting a named initiative over strengthening an existing one.

Over several cycles this produces an organisation with an unusual number of new initiatives, a shortage of people maintaining anything, and a review process that keeps returning good scores.

What partially corrects it

The fix is not to abandon reviews, which do useful work: they force a documented conversation about performance that would otherwise be avoided, and they create a record. The fix is to widen what counts as evidence.

• Ask peers a specific question rather than a general one. "Rate this person's collaboration" collects noise. "What did this person do that made your work easier this year, and what would have been harder without them?" collects the illegible category.
• Make maintenance visible on purpose. If nothing in the process distinguishes a system that ran without incident because of sustained attention from one that ran without incident because nothing happened to it, the attention will not continue.
• Separate the development conversation from the compensation conversation. When both happen in the same hour, the person is listening for the number, and the developmental content is not received in any meaningful sense.

A review process cannot measure everything that matters. It can at least be designed so the people who know what happened are asked the questions that would surface it.

Back to index

Executive communication does not fail for lack of clarity

When a message does not land across a large organisation, the diagnosis is almost always the same: it was not clear enough. The deck gets rewritten. The language gets tightened. A communications partner is brought in to sharpen the narrative. The message is delivered again, more clearly, and lands about as well as it did the first time.

Clarity was rarely the constraint. Most executive messages are perfectly comprehensible on first hearing. The constraint is that a single clear statement, delivered once from the top, is competing against everything else an employee is using to work out what is actually going on: their manager's tone, what happened to the last three initiatives, which projects just got funded, who was promoted, and what their own week looks like.

Against that field of evidence, a well-written statement is one data point. It is not usually the strongest one.

Three assumptions that cause the failure

That saying something once means it has been communicated. Senior leaders reach saturation with a message long before the organisation reaches first contact. By the time an executive is tired of repeating a priority, most of the company has heard it twice. The point at which leadership finds the message boring is roughly the point at which it is beginning to work.

That the cascade preserves the content. A message passed down four layers is not passed down. It is re-authored at each one, by people summarising it in their own words to an audience with different concerns, under time pressure, without having been briefed on which parts are load-bearing. What arrives at the front line may be accurate in outline and unusable in practice.

That the words carry more weight than the behaviour. If the stated priority is customer retention and the visible resourcing favours new acquisition, the organisation will conclude, correctly, that acquisition is the priority. Employees are experienced readers of the gap between what an organisation announces and what it funds, and they weight the second far more heavily.

What communicating at scale actually requires

It requires repetition well past the point of leadership boredom, in varied forms, over quarters rather than weeks.

It requires equipping the middle layer rather than instructing it: giving managers not just the message but the answers to the three questions their team will actually ask, including the awkward ones about what this means for headcount, workload and existing commitments.

And it requires that at least one visible decision in the following quarter be consistent with the message, ideally a costly one. A resourcing choice that lines up with the stated priority does more communicating than a year of townhalls, because it is the only form of evidence that cannot be produced cheaply.

Back to index

The translation problem

Strategy does not travel down an organisation intact. It is translated at every level, by people who were never told which parts must survive.

A strategic priority is set in language appropriate to the level where it was set. "Shift the portfolio toward recurring revenue" is a perfectly serviceable sentence in an executive session. It is close to meaningless to a regional sales manager deciding how to allocate their team's time next Tuesday, and that is not a failure of the manager.

Between those two points sits a translation chain. Each layer receives the priority in the language of the layer above and converts it into the language of the layer below. A director translates portfolio language into programme language. A senior manager translates programme language into team objectives. A team lead translates objectives into this week's work.

Each of those conversions is a judgement call, made without guidance, by someone who has not been told which elements are essential and which are illustrative. Small losses at each step compound. The version that reaches the front line is often a reasonable interpretation of the version above it and bears little resemblance to the original intent.

Why it is rarely caught

The chain has no error-detection. Information about how the strategy was interpreted travels back up the same layers that translated it, and each layer reports the version it believes to be true. Leadership receives confirmation that the message landed, which is accurate: something landed. Nobody is in a position to compare what arrived against what was sent.

Meanwhile the ordinary mechanisms for noticing the drift are weak. Metrics lag. Objective-setting processes typically check whether objectives exist, not whether they add up to the strategy. And a mid-level manager who has misread a priority has no reason to think they have, because their interpretation was reasonable given what they received.

What reduces the loss

Three practices do most of the work.

• State the non-negotiables explicitly. Every strategy has a small number of elements that must survive translation and a larger number that can be adapted locally. If leadership does not distinguish between them, translators will guess, and they will guess differently from one another.
• Test the translation at the bottom, not the top. Ask a frontline manager two levels down what the strategy means for their team's week. Their answer is the only reliable measurement of whether the handoff worked. Asking their director produces the director's version.
• Give the layers a shared artefact. A single short document that every layer works from (not a deck that gets rebuilt at each level) removes several of the conversion steps entirely.

Organisations tend to invest heavily in getting the strategy right and almost nothing in getting it transmitted. The second problem is more tractable than the first and is more often the one that decides the outcome.

Back to index

Silence is a message

Leaders tend to think of communication as the things they say. Teams experience it as the full set of things a leader could have addressed, of which the things actually addressed are a subset. The remainder is not neutral. It is read.

When a reorganisation is rumoured and leadership says nothing, the silence does not register as an absence of information. It registers as information: that the rumour is probably true, that it is bad enough to require careful handling, and that people will be told late. Two of those three inferences are frequently correct, which is why the interpretation sticks.

The instinct behind the silence is usually sound. Leaders wait because the details are unsettled, because premature communication would cause unnecessary alarm, or because legal or commercial constraints genuinely prevent disclosure. These are real reasons. They do not stop the silence from being interpreted, and the interpretation runs regardless of whether the reasons are good.

The default reading is pessimistic

In the absence of information, people do not suspend judgement. They fill the gap, and they fill it with the least favourable plausible explanation, not from cynicism, but because that is the prudent bet when the downside of being wrong is personal.

This produces a specific asymmetry. A leader who says nothing while a decision is being worked through believes they have communicated nothing. The organisation has in fact received a fairly detailed message, assembled from the silence, the timing, who has been seen talking to whom, and what happened the last time this pattern occurred.

What can be said when nothing can be said

The useful move is rarely to break confidence. It is to communicate about the silence rather than leaving it unattributed.

There is a substantial difference between no comment and a plain statement that a review is under way, that no decisions have been made, that people will hear by a stated point, and that leadership is aware the wait is uncomfortable. The second version discloses nothing confidential. It removes the ambiguity that the pessimistic reading feeds on, and it supplies a date, which converts an open-ended anxiety into a bounded one.

The same principle applies to smaller silences. A leader who does not respond to a raised concern has communicated something about that concern. A leader who says the concern was heard, is not being actioned, and here is why, has closed the loop even while declining the request. The second costs a sentence and preserves the willingness to raise the next one.

Nothing said is not nothing communicated. The only choice a leader has is whether to author the message or leave it to be assembled by other people.

Back to index

Strategy dies in the handoff

Most strategies that fail were not poor strategies. They died between the offsite where they were agreed and the first ordinary Tuesday on which somebody was supposed to work differently because of them.

The sequence is familiar. An executive team spends two days deciding priorities and produces a document. The document is presented downward, losing resolution at each level, until a frontline manager receives something like "we are focusing on mid-market now" with no attached guidance about what to do differently with their existing commitments, their existing capacity, or their existing targets.

At that point the strategy has technically been communicated and has functionally not been handed over at all.

What survives the handoff

Strategies that specify what stops. A team at capacity cannot absorb a new priority by goodwill. If the strategy names only what should start, the addition becomes optional in practice; it will be deprioritised the first time it competes with an existing commitment that someone is already accountable for. Naming what stops is the difference between a priority and an aspiration.

Strategies that reach a manager's own numbers. The workable test of a handoff is whether a manager two or three levels down can describe, specifically, how their team's weekly metrics should look different in six months as a consequence. If they cannot, the strategy has not arrived, however many times it has been presented.

Strategies with an owner below the executive team. Strategy owned collectively by leadership is, operationally, owned by no one. The plans that hold tend to have a named individual well below the top table whose job explicitly includes making the transmission work: not communicating it once, but checking a quarter later whether anything actually changed.

The diagnostic

A straightforward test, available to any executive team, is to ask three managers at different points in the organisation what the strategy requires them to do differently. Not what the strategy is (most people can recite that) but what is now different in their week.

Three vague answers indicate the handoff has not happened yet, regardless of how well the strategy was received when it was announced. The boardroom is where strategy is decided. It is rarely where strategy fails.

Back to index

Why plans do not survive their first quarter

Most annual plans are built on an assumption that is knowably false at the moment of writing: that the organisation has spare capacity.

Planning cycles produce a list of initiatives for the year. The list is ambitious, which is treated as a virtue. Resourcing is assigned. Owners are named. Somewhere in the first quarter, the plan quietly stops describing what the organisation is doing, and by mid-year the gap is wide enough that a re-planning exercise is scheduled, which produces a second plan with the same structural flaw.

The flaw is that the plan is built against nominal capacity rather than available capacity. A team of eight is planned as eight people's worth of new work. In practice, a substantial share of that team's year is already committed: to maintaining what exists, to the regulatory work that is not optional, to hiring and onboarding, to the incidents that will certainly occur without being individually predictable, and to the meetings that the plan itself generates.

None of this is hidden. It is simply absent from the planning document, because planning documents describe intentions and the committed load is not an intention.

What the first quarter reveals

The first quarter is when the two numbers meet. The plan requires more capacity than exists, and something has to give. What gives is rarely decided deliberately.

In the absence of an explicit choice, teams triage by urgency rather than importance: the thing with a date attached beats the thing that matters more but can slip. Strategic initiatives, which usually have the softest deadlines, are the reliable casualty. Not because anyone decided to deprioritise them, but because nobody decided not to.

Three corrections

• Plan against committed capacity, not headcount. Establish, before allocating new work, what share of each team's year is already spoken for. The number is usually uncomfortable and always useful. Plans built on the remainder are smaller and considerably more likely to happen.
• Make the trade-off explicit at the point of addition. When something is added mid-year, and it will be, the question of what comes out should be answered at the same meeting, by the person adding it, not left to a team to resolve silently three weeks later.
• Review the plan against the calendar, not against itself. A quarterly review that asks whether initiatives are on track will generate optimistic answers. A review that asks where the last quarter's hours actually went produces a more accurate picture and a much shorter conversation.

A plan that assumes unavailable capacity is not an ambitious plan. It is a plan that has delegated its prioritisation decisions to whoever is busiest in March.

Back to index

Measuring intent, mistaking it for outcomes

Organisations measure what is available to be measured. Activity is available immediately. Outcomes are available late, if at all, and are usually contaminated by factors nobody controls. Given that asymmetry, measurement systems drift toward activity, and the drift is rarely noticed because the dashboards stay green.

The pattern is easy to describe and hard to see from inside. A transformation programme reports the number of teams migrated. A sales enablement effort reports sessions delivered and attendance. A culture initiative reports participation rates. Each of these is a real number, accurately collected, and none of them answers the question the work was commissioned to answer.

What makes this durable is that activity metrics are not wrong. Teams did migrate. Sessions were delivered. The measure is truthful. It is simply measuring intent (that the organisation did the things it planned to do) and being read as evidence of effect.

Why the substitution persists

Three forces hold it in place, and all three are rational.

Activity data arrives inside the reporting cycle. Outcome data frequently arrives after the sponsor has moved roles, which means nobody in the room has an incentive to wait for it.

Activity is attributable. If revenue moves, a dozen factors could explain it and the programme's contribution is arguable. If forty workshops were run, forty workshops were run. Under scrutiny, the defensible number wins.

And activity is controllable. A programme team can guarantee delivery of sessions. It cannot guarantee that behaviour changes. Reporting on the thing you control is a reasonable instinct when you will be judged on the report.

A more honest structure

The correction is not to abandon activity metrics, which are useful for managing delivery. It is to stop presenting them as evidence of effect, and to name a single outcome measure in advance: before the programme starts, while the sponsor is still in the room and the number can be argued about honestly.

That measure should be one the organisation already collects, so it cannot be gamed by the programme, and it should be specific enough to be falsifiable. "Improved leadership capability" is not a measure. "Proportion of open roles filled internally" is, and it is imperfect, and imperfect and falsifiable beats comprehensive and untestable.

The discipline is in fixing the measure early. A measure chosen after results are known is not a measure. It is a selection.

Back to index

Why change programmes fail in the first ninety days

Change initiatives rarely fail because the direction was wrong. They fail because the opening quarter was run as a communications exercise when it needed to be an operating one.

The standard opening: leadership announces the change, delivers a townhall, stands up a steering group, and instructs managers to cascade. Ninety days later the programme has a name, a channel and a logo, and almost nothing about how work gets done has moved.

Awareness is not the gap. Ask employees to describe the change and most will do so accurately within a fortnight. The gap is that nothing in the actual mechanics of the job (the approval chains, the meeting rhythm, the numbers a manager is held to) was redesigned to match the new direction. People do what they are measured on. They have always done this, and a townhall does not alter it.

Three markers that separate programmes that hold

Something measurable changed in the first quarter. If ninety days pass without a single scorecard, dashboard or performance conversation being altered, the initiative is a campaign. This is the clearest early indicator available, and it is visible long before results would be.

Managers were told what to stop. Change programmes characteristically add (a process, a tool, a report) without subtracting. Managers, already at capacity, absorb the addition by quietly deferring it. They are not resisting. They are triaging, and the programme gave them nothing to triage against.

A visible win landed inside the window. Programmes that promise returns in eighteen months lose the organisation's attention in four. The first win does not need to be large. It needs to be real, attributable, and early enough that people conclude the thing is actually happening.

Why the communications version persists

Because it is available and it feels like progress. A townhall can be scheduled next week. Redesigning a set of operating metrics requires negotiating with the people who own them, several of whom will object, and produces nothing visible for a month.

The first option can be completed by the programme team alone. The second requires the organisation to change something it currently rewards. Only one of them is change management.

Back to index

What culture surveys miss

A survey measures what people are willing to write down. Culture is what they do when nobody is collecting responses.

Culture surveys are good instruments for a narrow purpose: tracking sentiment consistently over time. They are frequently asked to do something else (to describe the culture) and they are not built for it.

A team can return strong scores on psychological safety and still operate under a firm unwritten rule that junior people do not speak before senior people have. An organisation can score well on collaboration while its functions routinely withhold information from one another to protect their own numbers. The survey is not lying. It is capturing sentiment, which is real, while missing the behavioural norm that sentiment is coexisting with.

There is also a structural problem. The survey is administered inside the same environment whose pressures it is attempting to measure. If people are guarded about raising problems, that guardedness does not switch off because the form says anonymous. The conditions suppressing the behaviour are the conditions under which the measurement is taken.

More diagnostic than the instrument

Look at the last several promotions in a function. Which behaviour was actually rewarded: the person who hit the number regardless of method, or the person who hit it and left the team stronger? Promotion decisions describe a culture with more precision than any survey item, because they are decisions rather than statements.

Watch what happens after the senior person leaves the room. If the conversation opens up materially in the five minutes after a decision-maker exits, that is a finding, and it is one the survey will not return.

Examine which mistakes were absorbed and which were punished. Cultures reveal themselves fastest when something goes wrong. Was the person who flagged a problem early treated better or worse than the person who concealed it until it was unavoidable? Organisations teach this lesson once and people remember it for years.

Using the survey properly

The instrument becomes more useful when its scope is narrowed and its results are triangulated. Run it consistently, treat movement over time as more informative than any single score, and pair it with the behavioural evidence above before drawing conclusions.

The failure mode is not running surveys. It is treating the score as the culture, and then being surprised when a well-scoring organisation behaves in ways the score did not predict.

Back to index

Culture is what gets rewarded

Every organisation has two cultures. One is written down. The other is inferred by employees from a running record of what the organisation actually does, and it is the one that governs behaviour.

The written culture is produced deliberately, usually by a committee, and describes the organisation at its best. The inferred culture is assembled passively from evidence: who got promoted, who was tolerated, which mistakes ended careers and which were absorbed, what happened the last time someone raised an inconvenient objection.

Employees read the second set of evidence continuously and weight it far more heavily, for the straightforward reason that it is predictive. The values on the wall describe aspiration. The promotion record describes consequences.

The three signals that construct it

What is rewarded. Not what is praised: praise is cheap and everybody knows it. What is rewarded with promotion, resources, access and interesting work. If the organisation states that collaboration matters and consistently advances people who deliver individually while being difficult to work with, it has communicated its actual position clearly.

What is tolerated. The standard of an organisation is set by the worst behaviour that goes unaddressed, particularly from senior people. One tolerated exception does more to define the real limit than a policy document, because it demonstrates that the limit is negotiable and identifies who can negotiate it.

What is punished. Especially whether the punishment falls on problems or on the reporting of problems. Organisations that respond badly to bad news do not receive less bad news. They receive it later, from further away, and usually after it has become expensive.

Changing it

Culture change efforts that work on the written culture (new values, new language, a campaign) tend to produce short-lived movement and a durable cynicism, because the evidence stream continues unchanged and employees can see it.

Efforts that work tend to alter one of the three signals in a way people can observe. A promotion that visibly rewards behaviour the organisation says it wants, particularly when a more obviously qualified candidate on the old criteria was passed over, communicates more in a week than a values launch does in a year.

This is expensive, which is the point. The signal only carries because it cost something.

Back to index

Innovation theatre

A hackathon produces prototypes, energy, photographs and a shortlist of promising ideas. Some months later, a reasonable question (what shipped?) frequently has no good answer.

This is not because the ideas were weak or the participants insincere. It is because the event was designed to generate ideas and the organisation's actual constraint is almost never idea supply.

Most established organisations have more good ideas than they can execute. What they lack is a path from a prototype to a funded, staffed, owned piece of work with a place in somebody's roadmap. A hackathon that does not address that path is running a generation exercise against a throughput problem.

What happens to the winning idea

The winning team returns to their existing roles on Monday, where their existing objectives are unchanged and their managers have committed capacity for the quarter. The prototype now requires a sponsor, a budget line, security review, an operations owner and a slot in a roadmap that was set before the event.

Each of those requirements is individually reasonable. Collectively they constitute a wall that a side-project with no organisational standing will not clear. The idea does not get rejected (a rejection would at least be a decision); it simply stops having anyone whose job it is to advance it.

What distinguishes the events that produce something

• A named sponsor with budget attends, and commits in advance to funding an outcome. Not to considering the ideas. To funding one, with a stated figure, before anyone knows what the ideas are.
• Time is allocated after the event, not only during it. A prototype needs a defined number of weeks of protected capacity to become viable. If that allocation is not agreed beforehand, it will not be found afterwards.
• The constraint the organisation actually has is the theme. An open brief produces ideas the organisation has no route to execute. A brief anchored to a real, current, resourced problem produces fewer ideas and more shipped work.

None of this makes the event less energising. It makes the energy load-bearing. The alternative is a well-run annual exercise that reliably generates enthusiasm and reliably converts none of it, which teaches the organisation's most inventive people that proposing things is entertaining and pointless.

Back to index

The permission problem

In most approval chains, every participant can stop a proposal and no single participant can start one. That asymmetry is the whole story.

Consider what a new idea has to pass through in a large organisation: a line manager, a function head, finance, risk or legal, sometimes procurement, sometimes a governance forum. Six checkpoints is unremarkable.

Each checkpoint holds a veto. None holds the authority to approve outright and commit the rest. The proposal needs six yeses and one no ends it.

The payoff structure at each checkpoint then does the rest. An approver who blocks a proposal that would have succeeded bears no consequence, because the counterfactual is unobservable. An approver who waves through a proposal that fails is identifiable and will be identified. Under those conditions, declining is the dominant strategy for any individual approver behaving sensibly.

The organisation then concludes it has a culture problem (people are risk-averse, people lack entrepreneurial spirit) when what it has is a structure that would produce this behaviour from anyone.

What the chain does to proposals that survive

The proposals that clear six checkpoints are not usually the best ones. They are the least objectionable ones. Each gate removes whatever a given approver found uncomfortable, and the distinctive element of an idea is frequently the uncomfortable part.

What emerges is technically approved and has had the interesting portion negotiated out of it somewhere around gate four. This is why organisations often experience their innovation output as both slow and disappointingly incremental. The process is selecting for incrementalism at every step.

Structural corrections

Exhortation does not fix an incentive asymmetry. Three structural changes do more.

Give someone the authority to say yes. A single accountable decision-maker for a defined class of proposal, below a defined threshold, removes the unanimity requirement entirely. The threshold can be small. What matters is that a yes exists that does not require six of them.

Set a default that runs the other way. Proposals below a certain size proceed unless someone objects within a fixed window, with a named reason. This shifts the cost of inaction from the proposer to the approver, which is where it belongs.

Make the blocking visible. If nobody records what was declined and why, the organisation has no ability to notice that its approval process is its main constraint. A simple log makes a previously invisible pattern arguable.

Back to index

Good innovation processes look boring

Innovation has an aesthetic problem. The activities that look like innovation (offsites, labs, open briefs, creative sessions) are weakly correlated with organisations that reliably produce new things. The activities that do correlate look like ordinary operational discipline, which makes them difficult to sell internally and easy to cut.

Organisations that produce a steady stream of new products, services or methods are usually doing something closer to portfolio management than to inspiration. They run many small, cheap, time-boxed attempts. They kill most of them quickly and without ceremony. They concentrate resources behind the small number that show evidence. None of this photographs well.

The three unglamorous disciplines

Cheap tests before expensive commitments. The central skill is not generating ideas but designing the smallest test that would produce real information about one. Organisations without this skill go from idea straight to a funded programme, which makes every bet expensive and therefore makes failure politically unacceptable, which in turn makes bets rare.

Killing things routinely. Most attempts should stop. In organisations where stopping something is read as failure, nothing stops; initiatives are quietly maintained on reduced resource, absorbing capacity indefinitely. A functioning portfolio requires that ending an initiative be a normal, low-drama outcome that the person who ran it is not punished for.

A funding rhythm that matches the evidence. Annual budgeting forces a full-year commitment at the point of least information. Portfolios that work tend to fund in stages, with each stage releasing more resource against evidence rather than against the original plan.

Why the theatrical version wins anyway

Because it is legible to people outside the process. A lab with a name has a photograph, a launch and an announcement. A staged funding model with quarterly kill decisions has a spreadsheet.

When innovation is being demonstrated (to a board, to a market, to prospective recruits) legibility is the operative requirement, and the theatrical version supplies it. This is not necessarily wasteful. It becomes wasteful when the organisation mistakes the demonstration for the capability, funds the first, and concludes several years later that innovation does not work here.

Back to index

Why sales training does not stick

The conventional format (a multi-day workshop, a new methodology, a portal nobody revisits) has a well-documented failure pattern. Measurable skill gain in the room. Little detectable change in pipeline behaviour ninety days later.

The content is usually not the problem. Skill demonstrated in a workshop and skill demonstrated on a live call under quota pressure are different capabilities, and most programmes only build the first.

A rep in a training room has time to think, no consequence attached to the outcome, and an instructor available. The same rep on a real call has a buyer setting the pace, a number they are behind on, and one shot at the exchange. Under that load, people revert to whatever is automatic, and three weeks after a workshop, that is usually whatever they were doing before it.

What moves the numbers

Training built on live pipeline rather than hypotheticals. A rep who works a generic objection scenario learns something abstract. A rep who works the actual stalled deal currently sitting in their pipeline, with a coach present, learns something usable the following morning. Proximity to real deals is the single largest variable in whether training survives contact with a buyer.

Manager coaching afterwards, not just the event. Skills taught in a workshop decay within weeks absent reinforcement. Organisations where training holds are the ones where frontline sales managers were separately trained to coach the specific new behaviour in their existing one-to-ones. Without that layer the training is an event; with it, it becomes a habit. This layer is skipped more often than any other, because it requires developing managers rather than reps and the budget was allocated for reps.

A named behavioural metric, fixed in advance. "Improve our sales capability" cannot be tracked and therefore will not be managed. A specific behaviour (the proportion of discovery calls in which a budget conversation is opened, say) can be. Training tied to a trackable behaviour gets reinforced by whoever owns that number. Training tied to a general aspiration gets displaced by next quarter's priorities.

The uncomfortable implication

If reinforcement is the determining factor, then most of the value of a sales training investment sits in the weeks after the workshop rather than in the workshop itself, which means the standard procurement pattern, where the event is bought and the reinforcement is assumed, is buying the least valuable component.

Reps can absorb a new technique in two days. They generally do. What they return to is an environment in which nothing requires them to use it.

Back to index

The coaching gap on sales teams

Sales managers are selected for having been excellent sellers. They are then asked to do something nobody assessed them for and few were ever taught.

The frontline sales manager occupies the position with the largest single influence on a sales organisation's performance, and it is filled almost universally by promoting the strongest individual seller. The logic is understandable. The consequence is a management layer selected on one capability and evaluated on another.

Selling and coaching selling are related but distinct. A top performer has usually built a set of instincts they cannot fully articulate: they read a room, sense a stall, know when to push. That tacit skill is precisely what is hardest to transfer. Asked how they handled a call, a strong seller will often produce a description that is accurate about what they did and silent about the judgement underneath it, which is the part the rep needed.

What fills the vacuum

In the absence of coaching capability, the one-to-one defaults to the thing both parties know how to do: reviewing the pipeline. Deal by deal, stage by stage, forecast adjusted.

Pipeline review is necessary and it is not coaching. It inspects the output of a rep's skill without addressing the skill. A rep who leaves the meeting knowing which four deals need attention this week has received useful management. They have not been helped to handle the next discovery call differently, and the following week will produce a structurally similar list.

Two further defaults are common. The manager takes over: joining calls and closing deals personally, which secures the quarter and teaches the rep that difficult moments get escalated rather than learned. Or the manager motivates, substituting encouragement for diagnosis, which is pleasant and does not change what the rep does on Tuesday.

What closing the gap requires

It requires treating coaching as a taught skill rather than an expected consequence of seniority. Specifically: the ability to observe a call and identify the one behaviour worth changing rather than the nine that could be; the ability to ask a question that makes a rep reason aloud rather than supplying the answer; and the discipline to protect coaching time from being consumed by forecast pressure, which it will be in any month that is behind.

The last of those is the most commonly underestimated. Coaching is the first activity to be displaced when the number is at risk, and the number is at risk in most months. Organisations that sustain coaching generally had to make it structurally difficult to skip, rather than relying on managers to hold the line individually.

Where to start

The gap is usually easiest to see at the extremes: a manager who runs pipeline reviews exclusively, and a manager who can name the specific moment a call turned and why. Most sit somewhere between, coaching sporadically and inconsistently across their team. The practical starting point is not a training calendar but an honest look at how the last ten one-to-ones were actually spent, deal mechanics versus rep behaviour, because that ratio, more than any stated philosophy, is what a rep experiences as management.

Closing the gap is a sequencing problem before it is a skill-building problem. Managers coach better once an organisation has decided that coaching capability is something it will notice, measure, and protect on the calendar, not before.

Back to index