The evidence
Chris is what happens between the sessions.
Leadership programs work. What they struggle to do is survive the fortnight after everyone goes back to work.
Take one common shape: a 360 at the start, six facilitated sessions, and ninety minutes of one-to-one coaching. Whatever the shape, the arithmetic is the same. There are the eleven days a fortnight when the facilitator is not there, and that is where most of the value of most programs quietly disappears. The evidence below is not that an AI coach beats a human one. It is that the gap between sessions is where programs leak, and that something in the gap is what stops them leaking.
Why good programs still fade, and what stops it
Three turns in the argument, and nine findings behind them. Each one leads with why it matters; the study sits underneath if you want to check it.
What most organisations have already paid for
Almost every organisation has run some version of this: an assessment, an honest conversation about the results, and a hope that the behaviour follows. The research on what happens next is unusually clear, and it is not comfortable reading.
A 360 on its own changes very little.
You paid for the instrument, everybody filled it in, the debriefs were genuinely useful, and a year later most of the behaviour is where it was. That is not a failure of the tool or of the people. Feedback produces awareness, and awareness is not behaviour. The report was never the intervention, though it is usually priced as though it were.
Smither, London & Reilly (2005), Personnel Psychology · a review of longitudinal studies
The people who improved were the ones who picked something and worked on it.
Not everyone improved equally, and the difference was not talent or seniority. It was whether the person turned the feedback into one specific thing they were going to do, and then went and did it. Later work sharpened the mechanism: what separates an intention from an action is naming when and where you will do it, not how much you want to. This is the whole reason the product asks a leader for one practice attached to a real occasion, rather than a development plan with nine strands. Somebody worked this out in 1979 and it has held up everywhere it has been tested since.
Nemeroff & Cosentino (1979); Gollwitzer & Sheeran (2006) · multisource feedback with and without goal setting, and a meta-analysis of goal-striving experiments across domains
The practice a leader cannot pick is the one they cannot see.
Naming when and where works for the leader who can already see what to change. It does nothing for the more interesting case: the one who sincerely commits and does not move, because something invisible is holding the old behaviour in place and doing its job. Robert Kegan named that difference, between what runs you and what you can hold up and look at, and the developmental half of this field, instruments and coaching alike, acknowledges him as its foundation. In leadership the obstacle is rarely a missing skill and usually a protected one, which makes that shift the unlock, and it is why the questions come before the practice. That is a claim about what Chris is built on, and it is the only one made here: whether it does the frame justice will be settled by leaders using it and saying so, not by this page.
Kegan (1994); Kegan & Lahey, Immunity to Change; Bachkirova; Berger & Atkins (2009) · constructive-developmental theory and the coaching built on it, stated as the frame rather than as a finding
The program is not the problem
It would be convenient for us if facilitated leadership development did not work. It does, and the evidence is stronger than for most things organisations spend money on. That matters, because it moves the question from "does this work" to "what happens to it afterwards".
Well-designed leadership programs genuinely change behaviour at work.
This is the largest body of evidence in the field, and it is decisive. Programs with a needs analysis, real feedback, practice rather than lecture, and sessions spread over time produce behaviour change that shows up on the job. So the session in the room is doing its job. If the value leaks away later, that is a different problem with a different fix, and blaming the workshop is the wrong diagnosis.
Lacerenza, Reyes, Marlow, Joseph & Salas (2017), Journal of Applied Psychology · a meta-analysis of leadership training
One-to-one coaching works, and buying more sessions does not buy more change.
The number of sessions does not predict the result. That is a genuinely useful finding for anyone sizing a budget: the instinct is to add hours, and the evidence says the hours are not where the lever is. What a leader does in the weeks between the sessions is doing more of the work than the sessions themselves.
Jones, Woods & Guillaume (2016), JOOP; Theeboom et al. (2014) agrees · two independent meta-analyses
The gap between the sessions is the problem, and it can be closed
Everyone who has run a program knows the feeling: the room is alive on the day, and a fortnight later people are back to how they worked before. That leak is measurable, it has known causes, and closing it is the single best-evidenced thing you can do with the money you have already committed.
Learning does not survive contact with the job unless something maintains it.
Transfer has two halves: getting the learning onto the job at all, and keeping it there over time. Programs are designed for the first half and assume the second. The second is where the money goes, quietly, with nobody in the room when it happens. Nobody sends a report saying the behaviour stopped.
Baldwin & Ford (1988); Blume et al. (2010) · the founding review, and a later meta-analysis
The leak is not inevitable. It has causes you can do something about.
This is the turn. Whether learning survives depends on things an organisation controls: whether there is any structure holding a leader to it, whether anyone follows up, and whether the environment they go back to makes the new behaviour possible. None of that is luck, and none of it requires more training.
the transfer-of-training literature, Baldwin & Ford onward · the founding review of the transfer literature
What goes in the gap is a design decision, and the choice changes the result.
This is the step where the evidence is thinnest and it is worth saying so. The transfer literature establishes that maintenance matters; the leadership meta-analysis recommends booster activities explicitly; and the one trial that put boosters after leadership training compared two kinds of them, coaching against emails, rather than boosters against nothing. So the direction is well supported and the size of the effect, for leadership specifically, is not settled. What follows from that is a reason to measure our own, not a reason to borrow somebody else's number.
Lacerenza et al. (2017); Tafvelin, von Thiele Schwarz & Stenling (2023), Scandinavian Journal of Work and Organizational Psychology · a leadership-training meta-analysis that recommends boosters, and a trial comparing kinds of booster after leadership training
Following up is the active ingredient, and it works better when the team is in it.
Across a very large sample, the leaders whose people saw them improve were the ones who kept going back to what they were working on, and who told the people around them what they were trying to change. Being seen to work on it is part of the mechanism, not a nice extra. That is the Loop, and it is why the people around a leader get asked what they experience.
multi-rater follow-up research · large-scale multi-rater follow-up research
It is finished when it runs without anyone thinking about it.
Decide to act, turn the intention into the behaviour, repeat it, and eventually stop needing to decide. That last stage is the point of the whole exercise: a practice a leader still has to remember is not yet a change in how they lead. Commit, practise, log the rep, get the nudge: the four stages of the research map almost exactly onto the four things the product asks for.
Lally & Gardner; Trenz et al. (2024), JOOP · habit formation meta-analyses, including workplace-specific work
Spread out beats concentrated, whatever the room.
The same content, spaced over weeks, produces more learning than the same content in a block. This is one of the oldest and most replicated findings in how people learn anything, and it is what makes modular delivery the preferred design: sessions spread over months, with practice and reinforcement running in between rather than a gap. A two-day offsite is easier to schedule, and that is its only advantage.
Donovan & Radosevich (1999)
What we actually pay attention to in a leader
Coaching that could be about anything ends up being about nothing, so Chris works from a defined picture of what leadership is made of. Fifteen parts, in plain English, and the research each one stands on.
Why we can build this, and say so
Every instrument in the drawer describes a similar shape, and not because they copied one another. They descend from the same public primary literature, and Bob Anderson says as much in his own paper on the Leadership Circle Profile: Karen Horney for the reactive half, Robert Fritz for the creative and reactive orientations that name its two halves, and Kegan, Wilber, Gilligan, Hall, Kohlberg, Beck, Cook-Greuter and Torbert for the developmental stages, with Kegan the heaviest single influence. He built a model from the primary sources, and the Universal Model of Leadership is his. We built ours from the same literature, and every dimension below shows what it stands on.
The strongest evidence that this shape belongs to nobody is that it was arrived at more than once. Horney described three strategies in 1945: moving toward people, moving against them, and moving away from them. J. Clayton Lafferty took those three into the Life Styles Inventory as its defensive styles. Bob Anderson took the same three into Complying, Controlling and Protecting. Two instruments that compete with each other, built by different people a generation apart, reached the same three moves because both went back to the same book. Ours is another reading of it rather than a version of theirs, which is why we cite her and not a brand name. The method converges too: Anderson credits Burns and Ellis for the cognitive work underneath his profile, Lafferty drew on Ellis as well, and the coaching evidence above rests on that same cognitive-behavioural lineage.
It happens from the other direction too, with no theory at the start. Kouzes and Posner began in 1983 by asking people what they did at their personal best and analysing the cases for what recurred, rather than testing a model they already had; the five practices came out of that. Zenger and Folkman worked from a very large body of 360 assessments to find which behaviours actually separate the most effective leaders from everyone else. Neither began with Horney, and both arrive at the same split this model uses: what a leader does with themselves, and what they do with the people around them. A shape that turns up whether you start from a clinical theory or from a mountain of assessments is a shape worth building on.
How you hold yourself, and what you do with people
Every serious model of leadership splits the same way: there is what a leader does with themselves, and what they do with the people around them. Most leaders are visibly stronger on one side, and the side they are weaker on is usually the one they think about least.
The oldest empirically supported split in the field, running from the Ohio State studies’ initiating structure and consideration through every task-versus-relationship distinction since. It is also the spine of the emotional intelligence models, which divide into awareness and management, of self and of others.
What you build, and what you do under threat
The second split is the one that explains most of the frustration in leadership. The same person who is thoughtful on a good day becomes short, or controlling, or unreachable on a bad one. That is not two people and it is not a character flaw. It is what competence does when it feels threatened, and it is predictable enough to work with.
Karen Horney (1945), Our Inner Conflicts: under threat, people move toward others, against others, or away from others. That is the source the reactive half of every modern instrument descends from, which is why we cite her rather than anybody’s brand name for the same three moves.
How you hold yourself
Whether you are steady enough that people can read you, honest enough that they know where they stand, and clear enough about what you stand for that it survives a bad quarter. This is the domain leaders neglect because it does not look like work, and the one their teams notice first.
Grounded: how you hold yourself
Steady under pressure
Stays level when it gets hard, and does not transmit stress downward.
Inefficacy, the feeling that you are not accomplishing much, has the strongest negative link to job performance of all burnout dimensions.
Corbeanu et al. (2023), meta-analysis of burnout and job performance · meta-analysis · 3 anchors in total
Self-aware
Knows their own pattern well enough to name it while it is happening.
Improvement following multisource feedback is generally small; feedback alone does not change behaviour.
Smither, London & Reilly (2005), review of longitudinal studies · meta-analysis · 2 anchors in total
Straight
Says the same thing in the room and out of it.
"Has high ethical standards and integrity" ranks in the top ten strengths at every level of management.
Hogan 360 global benchmark · practitioner benchmark, not peer-reviewed
Purposeful
Can say what they stand for, beyond this quarter’s targets.
Level 5 leaders channel ambition into the cause rather than themselves, and that distinguishes sustained company performance.
Collins (2001), Good to Great
What you do with people
Whether people finish their sentences around you, raise a problem while it is still small, get their share of your attention when nothing is going wrong, and hear specifically what they did well. Most of what a team calls culture is decided here.
Connected: what you do with people
Listens properly
People finish their sentences, and sometimes change what the leader thinks.
Psychological safety’s strongest outcome is learning behaviour: people experiment, seek feedback and reflect on mistakes when it is present.
Frazier et al. (2017), meta-analytic review · meta-analysis
Makes it safe
People raise problems early, without first calculating what it will cost them.
Psychological safety retains unique predictive power for performance and helping behaviour even after accounting for leadership, work design and relationships.
Frazier et al. (2017), meta-analytic review · meta-analysis · 3 anchors in total
Develops people
Spends time on people’s growth, not only on their output.
Team leadership behaviours account for a meaningful share of team learning behaviours.
Koeslag-Kreunen et al. (2018), When Leadership Powers Team Learning: A Meta-Analysis · meta-analysis · 2 anchors in total
Recognises
Notices the specific thing, close to when it happened.
Employees recognised at least weekly are around five times more likely to feel connected to their workplace culture. Frequency matters more than size.
Gallup workplace recognition research (2022) · practitioner benchmark, not peer-reviewed
How you get results through other people
Whether people know what matters this fortnight, whether decisions actually close, whether poor performance gets addressed early and cleanly, and whether you ask for more than is comfortable. Leaders are usually promoted for this domain, which is exactly why it hides the cost of the other two.
Driving: how you get results through others
Makes it clear
People know what matters most right now, and why.
Specific, challenging goals produce higher performance than "do your best" goals, and the effect strengthens with difficulty while attainable.
Locke & Latham (2019), fifty-year retrospective on goal-setting theory · meta-analysis
Decides
Closes open loops rather than leaving them running.
"Is visionary and strategic" is the competency that most separates top-quartile executive leaders from every other level.
Hogan 360 global benchmark · practitioner benchmark, not peer-reviewed
Holds the line
Addresses poor performance early and cleanly, rather than working around it.
"Challenging poor performance" is one of only two development opportunities common to every level of management in the benchmark.
Hogan 360 global benchmark · practitioner benchmark, not peer-reviewed
Sets the ambition
Asks for more than is comfortable, and means it.
Transformational leadership effects on performance are consistent across individual, team and firm levels, and culturally robust.
Agag et al. (2024), cross-cultural meta-analysis · meta-analysis · 2 anchors in total
What happens under pressure
Everybody has a move they make when something feels threatening: smoothing it over, gripping tighter, or going quiet. It is not a fault, it is what competence does under threat, and it is the most useful thing a leader can learn about themselves because it is the part they cannot see while it is happening.
After Horney (1945). Every leader has a home one, and knowing which is more useful than trying not to have it.
Appeasing
Buying safety with agreement. Moving toward.
Both of the two universal development opportunities across every level of management sit here: "challenging poor performance" and "taking on too much, delegate more".
Hogan 360 global benchmark · practitioner benchmark, not peer-reviewed · 2 anchors in total
Controlling
Buying safety with grip. Moving against.
"Spreading yourself too thin" and "delegate more" is one of the two universal development gaps in the benchmark.
Hogan 360 global benchmark · practitioner benchmark, not peer-reviewed · 3 anchors in total
Withdrawing
Buying safety with distance. Moving away.
Inefficacy has the strongest negative relationship with job performance of the burnout dimensions.
Corbeanu et al. (2023), meta-analysis · meta-analysis · 2 anchors in total
There is a method, and this is the only place it is named
A leader is never told which technique they are in the middle of, because naming it turns a conversation into a procedure. Somebody deciding whether to buy this is entitled to know exactly what it is.
Cognitive-behavioural coaching
Beck; Ellis (1988); Burns (1980)What a leader believes about a situation drives what they do in it, and the result then appears to prove the belief right.
A leader who is convinced their team cannot handle bad news withholds it, the team is then surprised by something, and the leader concludes they were right not to say anything. Nothing in that loop is stupid, and effort will not break it, because effort is what keeps it running. Chris works on the belief and the evidence for it, which is why a practice is framed as a test of what you expect to happen rather than a chore to complete.
GROW
Whitmore, Coaching for PerformanceThe shape a useful conversation takes: what you want, what is actually true, what you could do, and what you will do.
The last part is the one that gets skipped, and skipping it is how a leader has a genuinely enjoyable conversation that changes nothing. A conversation that never reaches "what will you do" is one they liked. Chris is built to get there before the end.
Acceptance and commitment
HayesFor the beliefs that will not shift no matter how much evidence you put in front of them.
Some things a leader believes about themselves are not going to be argued away, and trying is how coaching turns into a debate. The alternative is to stop fighting the thought, accept that doing the thing will feel uncomfortable, and take the smallest action that moves toward what they actually care about. That last step is where the One Big Practice comes from.
Why this and not something else
Cognitive-behavioural work is the best-evidenced approach to changing behaviour that exists, and its coaching forms translate directly: instead of arguing with what somebody believes, you help them test it against what actually happens. Coaching built on psychological method shows its strongest effects on exactly the two things a leadership program needs to survive the gap between sessions: whether people reach the goals they set, and whether they believe they can.
Wang et al. (2021), meta-analysis of psychologically informed coaching · Locke & Latham (2019) on goal setting · workplace evidence for acceptance and commitment approaches
What this is not
This is coaching, not therapy. It is about work, about now, and about what somebody is going to do next. The same techniques applied to grief, trauma or mental illness need a qualified clinician, and Chris is built to notice when a conversation has that kind of weight and hand it on rather than carry on regardless. Any organisation with a duty of care to its people should ask this question, so we answer it before it is asked.
Three things we could tell you, and do not
All three are common in this market and all three would help us. Two are simply wrong and one is overstated, and any of them takes about five minutes to check.
Not claimed
Positive psychology interventions produce d = .61.
That figure has been corrected downward twice as the meta-analytic base improved and publication bias was accounted for. The current position is considerably smaller, and still real.
White, Uttl & Holder (2019), re-analysing the earlier meta-analyses
What we say instead. Positive psychology interventions have a small but genuine effect. We use the current estimate rather than the one that reads best.
Not claimed
Emotional intelligence is twice as important as IQ, and IQ is only 20% of success.
Both numbers trace to popular writing rather than to a study, and the underlying claim does not survive the meta-analytic work on how emotional intelligence relates to job performance once cognitive ability and personality are controlled for.
Joseph & Newman (2010)
What we say instead. Emotional intelligence predicts performance in roles where managing yourself and reading other people is the job. Leadership is one of those roles. That is a narrower claim, and it holds.
Not claimed
Leadership effectiveness explains 37.6% of variance in business performance.
It explains variance in perceived business performance, rated by the same people who rated the leadership. Same raters on both sides of a correlation inflates the relationship.
Anderson, n = 486
What we say instead. We say perceived business performance, every time, and we point at the shared-rater problem before anyone else has to.
Behind Chris sits a working library of 898 research findings, read and extracted rather than linked to. 378 of them, drawn from 188 separate published works, are rated strong or meta-analytic and come from outside our own writing. Of those, 128 are meta-analyses, which pool many studies rather than resting on one. These figures are counted from the library itself every time this page loads, so they cannot quietly drift from what is actually in it.