My working definition is short. Capability-based coaching differs from generic coaching in one respect: it treats a senior leader's development as a change to the organisation around them, not only to the person. Generic coaching improves a leader. Capability-based coaching changes what the leader's team, decisions and incentives reward, so the improvement is still there after the next quarter.
I write this for CEOs, CHROs and boards who have already paid for executive coaching and want to know why the effect faded, or who are about to buy it and want to buy it once. I worked with a CEO (a sharp operator who'd scaled the business from $50M to $300M in eight years) who was visibly frustrated. His executive team wasn't executing with the precision he expected. The question I get asked is rarely whether coaching works. On average it does. The useful question is what decides whether it lasts, and the research and my own practice point to the same answer: the system the coached leader returns to.
Why capability-based coaching starts with the organisation, not the individual
Start with what the evidence says. A 2023 meta-analysis of 37 randomised controlled trials of workplace and executive coaching, covering 39 samples and 2,528 people, found a moderate positive effect (g = .59), with caveats: the authors note signs of publication bias, and effects were larger when outcomes were self-reported than when they were observed (de Haan and Nilsson, 2023). An earlier meta-analysis of 17 studies found positive effects on organisational outcomes overall (Jones, Woods and Guillaume, 2015). Both papers also report that the length of the coaching assignment did little to explain the result. Buying more hours is not the lever.
The lever sits outside the coaching room. Blume and colleagues pooled 89 studies of training transfer, which is the question of whether what someone learns shows up in their work. They found positive links between transfer and a supportive work environment, and they found that predictors such as motivation and work environment were stronger when the training was aimed at open skills, leadership development being their example, than when it was aimed at closed skills such as software (Blume, Ford, Baldwin and Huang, 2010). Leadership is the very skill set where the surrounding environment decides how much of the learning survives. That is the case for capability-based coaching in one sentence.
There is also a measurement trap that explains why so many boards think their coaching worked. Satisfaction scores and coach ratings are easy to collect, and they are usually good, because a well-run coaching relationship is a pleasant experience. They tell you the leader valued the sessions. They do not tell you whether the leader's team now decides faster, whether their direct reports are ready for the next role, or whether the board's questions about leadership depth have gone quiet. I ask for those three indicators before the first session so nobody can retrofit a success story afterwards.
That line sits on the work-with-me page of this site because it describes the order of operations I follow. Coaching is a delivery mechanism inside a designed architecture, not the offering itself. Steve Nassif, COO of Royal Health, says on my About page that "the rate limiting factor is often management capability, and that doesn't scale through osmosis."
The island problem
Consider a pattern you may recognise. It is an illustration and not a client story. A chief financial officer works with an excellent coach for six months. She leaves with sharper delegation habits and a clearer view of how she shows up in the room. Then she returns to a leadership team whose decision rights are unwritten, whose peers have never heard what she learned, whose performance reviews reward the old behaviour, and whose board still measures success the old way. Within months the old patterns reassert themselves. The coach was excellent. The system was unchanged.
Coaching without organisational alignment is a new engine in a car that still has the old transmission. I worked with a financial services organization that spent $4M over three years on executive coaching from a prestigious firm. Their coaching satisfaction scores were excellent. Their leaders reported positive experiences. But their business metrics didn't move. Their execution velocity actually declined. Why? Because the coaching wasn't integrated into how they made decisions or held themselves accountable. When we redesigned their approach, with the same coaches in many cases but integrated into a systemic capability-building strategy, everything shifted. Within 18 months, their execution velocity increased 40%. The difference wasn't better coaches. It was a better system. My Leadership Capability Stack is the diagnostic I use to find where: Strategic Intent, Leadership Architecture, Decision Systems, Capability Standards and Culture Output. Skills and behaviour are the visible top. The fracture is usually lower down, which is why a diagnostic conversation with the whole leadership team tends to teach a board more in a fortnight than a year of individual sessions would.
Five pillars that have to exist around the coach
The 5D sequence I use (Discover, Desire, Define, Deliver, Drive) is described in the 5D Leadership Architecture. When the work reaches Define and Deliver, I look for five pillars of infrastructure. Most coaching engagements supply one or two. Each one means something specific when the question is whether coaching will stick.
- A Leadership Operating System: Decision rights, escalation paths and meeting rhythms written down, so a newly coached leader can use a new behaviour without waiting for the whole team to guess what has changed.
- Embedded Development Pathways: Coaching tied to real work this quarter, with a defined standard for each level, rather than a parallel programme that sits beside the job.
- Pipeline and Succession Architecture: A visible route for the leaders below the coached executive, so development is a career pattern the organisation recognises and not a perk for the top ten people.
- Culture-to-Execution Infrastructure: Performance reviews, promotion criteria and board reporting that reward the behaviour being coached. Without this, coaching stays optional and friction wins.
- Scaling Mechanisms and Measurement: Peer coaching, a leadership community and a few measures tied to business outcomes, so capability compounds after the external coach leaves.
These pillars depend on one another. A leadership standard with no accountability changes nothing. Accountability with no cultural support creates resistance. Support with no link to the business turns performative. The provider you choose should be able to describe how their work reaches all five, even if they deliver only some of it themselves. For a related view of where coaching is the wrong tool entirely, read when executive coaching is the wrong intervention and executive coaching or leadership systems: which to fund.
Six questions to ask before you sign
Most organisations choose a provider on credentials, referrals or availability. I would choose on whether the provider understands your specific capability gaps and has a method for closing them. Ask these six questions. They work whether you are hiring an individual coach, a firm or an internal programme, and they pair with my guide to how to choose an executive coach.
- Do they diagnose organisational gaps before designing the coaching?: A capable provider starts with a diagnostic: interviews with the board, the C-suite and high-potential leaders, and a look at where decisions actually stall. A provider pitching the same programme to every client is not doing this work.
- Do they work with the leadership system or only with individuals?: Look for time spent with the full leadership team on how decisions get made and what the team's standards are. If the coach never meets the peers of the person being coached, the island problem is built in.
- Is the coaching tied to your strategy and to measures you already track?: Ask which of your three-year priorities the coaching serves, and which business indicators would move if it works: decision cycle time, escalation volume, retention of high-potential leaders. Satisfaction surveys alone cannot answer this.
- Do they address systems as well as behaviour?: A leader coached well can still fail if performance management, incentives and communication norms punish the new behaviour. The right provider examines those systems with you.
- What persists after the coaching ends?: If the coach disappeared tomorrow, would peer coaching, standards and routines keep going? Good providers build that persistence and make themselves less necessary over time.
- How will they measure sustained change, not early progress?: Early progress is easy. Ask how they will track whether the behaviour holds six and twelve months later, and whether it spread to peers and direct reports.
What to measure, and when to expect it
The research gives no support for the idea that a longer engagement automatically means a better one. De Haan and Nilsson found the effect did not depend much on the length of the assignment, and Jones and colleagues found no moderation by number of sessions or longevity. So do not anchor on hours. Anchor on integration and on indicators that belong to the business.
- Decision cycle time and escalation volume at the level below the coached executive
- Promotion readiness of the leaders who report to them, judged against a written standard
- Retention of high-potential leaders in the part of the organisation the coaching touches
- Whether peers can describe, unprompted, what has changed in how the coached leader works
- Whether the behaviour holds when pressure is high, checked at six and twelve months
On timing, my own expectation is that you should see first behaviour change inside a quarter and structural change by the second or third review cycle. That is a working expectation from practice, not a research finding, and a provider who promises a faster miracle should be asked how they know. If you want my fuller view of timelines, how long executive coaching takes to show results goes through it, and whether executive coaching is worth the investment covers the return question.
A practical sequence for the C-suite
Put together, this gives a sequence a board can hold a CEO to. None of it is complicated, but most organisations skip steps because they want a quick result: they hire a coach, run six months, measure satisfaction, declare success and move on.
- Diagnose: Map your actual leadership gaps against your strategic priorities, with interviews rather than a generic survey.
- Define: Write what good leadership looks like at each level in behavioural terms specific to your organisation.
- Align: Agree with the board and CEO what the coaching is for and how success will be measured, in business terms.
- Embed: Wire the new behaviours into performance reviews, promotion criteria, succession planning and board reporting.
- Connect: Build peer accountability so executives coach each other and not only receive coaching from outside.
- Sustain: Measure change over twelve to twenty-four months and keep a leadership community going after formal coaching ends.
The professional services case study on my work-with-me page shows the pattern in a different setting. A CHRO inherited a fragmented leadership culture after a merger, with no common language and no shared accountability. The solution listed there is a culture-to-execution architecture with embedded development pathways across 120 leaders. The published results are a 45 per cent improvement in leadership alignment score, a 35 per cent drop in voluntary attrition among leaders and merger integration completed six months ahead of schedule. These are my own figures from my site, so treat them as a description of the approach and not independent evidence.
The distinction that decides the spend
Generic coaching asks how to make this leader better. Capability-based coaching asks what the organisation must change for a better leader to stay better. The first question has a coach as its answer. The second has a coach as one part of its answer, alongside decision rights, standards, incentives and a rhythm that keeps all of it alive. A C-suite that buys the second kind gets a better leader and an organisation that keeps the improvement after the coach has gone.
If you are weighing options for your own leadership team, the leadership assessments give you the diagnostic without a coaching commitment, CapabilityAI, the AI coaching companion for the work around each session, is set out on its own page, and the Architecture Accelerator installs the surrounding structure. Or apply to work with me and we will start with the diagnosis.