The business case for leadership coaching does not rest on an ROI multiplier, because there is no verified figure that applies broadly. It rests on a chain you can check: managers explain most of the variation in team engagement, replacing people costs half to twice their salary, and coaching shows a moderate positive effect in controlled research. I will show you each link, where it comes from and where it breaks.
Boards ask me to make the business case for leadership coaching more often than they ask almost anything else about my practice. Usually they have already sat through a pitch built on a number nobody can trace: a sevenfold return, a 700 per cent ROI, a "for every dollar spent" line lifted from a slide deck with no citation attached. I do not blame them for asking twice. A number with no source behind it is not evidence. It is marketing wearing a lab coat.
So this is my honest attempt at the thing most coaching content avoids: a business case built entirely from numbers I can point you to, from people whose job is measuring this stuff, not selling it. Three findings, three exact sources, and the reasoning that connects them. If you want the vague version, it's a five-minute search away. This is the version I'd actually put in front of a CFO.
Why the Usual Pitch for Coaching Doesn't Survive Scrutiny
Search "ROI of executive coaching" and you will find the same few figures repeated across hundreds of pages: a 700 per cent return, a five-to-one or seven-to-one multiplier. When I followed them back, the trail was thinner than the confidence. The 600 to 700 per cent range comes from older retrospective studies, and a 2016 systematic review by Grover and Furnham reports that the technique used to calculate those figures was rather subjective and tenuous. The seven-times figure appears on the International Coaching Federation's own ROI page, attributed to a PwC and Association Resource Center survey, with no description of how it was calculated. Call these what they are: claims waiting for a method, not findings.
I'm not saying coaching doesn't work. I've run it as my practice for years and watched it change how leaders decide, delegate, and hold a room. I'm saying the argument for it should rest on something sturdier than a stat nobody can walk you back to. There is real research on this, from organisations whose entire job is measuring workplace behaviour, and it tells a more specific, more useful story than any recycled multiplier ever could.
The Business Case for Leadership Coaching, in Three Numbers
Three findings carry this case, and I'd defend each in a boardroom. Each comes with a link to the original source so you can check it yourself rather than take my word for it.
70%: of the variance in engagement across business units: Gallup: managers account for at least 70% of the variance in employee engagement scores
Gallup's research on workplace engagement found that managers account for at least 70 per cent of the variance in employee engagement scores across business units. Read that carefully: not 70 per cent of engagement itself, but 70 per cent of the difference between your best-performing team and your worst. Two teams doing the same work, under the same company policies, with wildly different engagement scores. The dominant explanation for that gap isn't pay, or perks, or the mission statement on the wall. It's who runs the team.
50-200%: of annual salary: Gallup: the cost of replacing an individual employee can range from one-half to two times annual salary
The same research puts a number on what that gap actually costs. Gallup puts the cost of replacing an individual employee at between half and twice their annual salary, and calls that a conservative estimate. Scaled across a workforce, Gallup estimates voluntary turnover costs U.S. businesses close to a trillion dollars a year, and more than half the employees who leave say their manager or organisation could have done something to stop it. Over half also say that in the three months before they left, no manager or leader spoke to them about their job satisfaction or their future.
g = 0.43 to 0.74: range of effect sizes for workplace coaching: from coping (0.43) to goal-directed self-regulation (0.74), across five outcome categories in one meta-analysis
The third number comes from outside the coaching industry entirely: an academic meta-analysis published in The Journal of Positive Psychology, pooling the available quantitative research on workplace coaching. Using Hedges' g, a standard measure of effect size, the researchers found coaching produced a meaningful positive effect across every outcome category they tested, from g = 0.43 on coping through to g = 0.74 on goal-directed self-regulation. In plain terms, that's a moderate-to-large effect by the conventions researchers use to judge these things, and unlike a marketing multiplier it comes with a method you can inspect.
That finding is a decade old now, so I checked what has been published since. Two 2023 analyses update it. De Haan and Nilsson pooled 37 randomised controlled trials of workplace and executive coaching, with 2,528 participants across 39 samples, and estimated a standard effect of g = 0.59, which they call moderate. Cannon-Bowers and colleagues analysed 11 studies after removing a clear outlier and estimated g = 0.44, and concluded that coaching is an effective workplace intervention. Both are lower than the top of the 2014 range, which is worth noticing. Cannon-Bowers and colleagues point out that the 2014 analysis found pre-post designs gave significantly larger effect sizes than designs with a comparison group, so a range quoted from the earlier analysis may flatter the result.
How the Three Numbers Connect
On their own, these are three separate findings from three separate fields. Put together, they form the actual argument, and it isn't the one coaching brochures usually make.
The manager is the single biggest lever on whether a team stays engaged and stays put. Losing that team, or watching it disengage under a leader who never develops, is one of the most expensive and most preventable costs a business carries. Coaching has controlled research showing a moderate average effect on how a leader performs, not just how they feel about performing. Chain those three facts together and the case is plain: the most measurable, highest-impact place to spend a development budget is the person whose behaviour already explains most of the gap between your best team and your worst.
That's a different argument to "coaching pays for itself seven times over." It's narrower, it's checkable, and it tells you exactly where to point the investment: not at every leader equally, but at the ones sitting on top of your most expensive engagement and retention risk.
What I'd Actually Put in Front of a Board
- Start with the manager, not the department: Pull engagement or retention scores at the team level, not the company average. A company average hides exactly the variance Gallup's research is describing. Look for the widest gaps between teams doing comparable work.
- Name the leaders sitting on the biggest gaps: The business case isn't for coaching in general. It's for coaching the specific leaders whose teams are the outliers, good or bad. That's where the 70 per cent finding actually points.
- Cost the alternative before you cost the coaching: Work out what it would cost to replace the people currently disengaging under that leader, using the 50 to 200 per cent of salary range as your floor. That number, not the coaching invoice, is what you're actually weighing the investment against.
- Set the outcome before the engagement starts: Decide in advance what you're measuring at six and twelve months: retention in that team, an engagement re-survey, a specific behaviour shift reported by direct reports. Without a baseline, you can't claim an effect either way.
- Review at the interval the research suggests, not the interval that's convenient: Geerts's 2024 framework for leadership development evaluation recommends collecting data at the start, during, at the end and again six to nine months later, because outcomes take time. Reviewing after a single quarter and calling it a verdict risks a false negative.
Where This Case Falls Apart
I'd rather tell you where this argument breaks than let you find out after you've signed a contract. It falls apart in three specific places.
First, attribution gets murky fast. If a team's engagement improves at the same time three other things changed, a market shift, a restructure, a new product launch, you can't cleanly credit the coaching. That's not a reason to avoid measuring; it's a reason to measure something specific enough that other explanations don't crowd it out.
Second, the research measures coaching's effect on the leader. It doesn't promise the organisation around that leader will let the change stick. A leader who returns from six months of coaching with a genuinely different way of delegating, into a structure that still rewards the old command-and-control habit, will drift back within a quarter. The effect size is real; the environment can still swallow it.
Third, this case is built on averages across many leaders and many studies. It tells you coaching works on average. It doesn't tell you this leader, in this role, with this specific gap, is a good coaching candidate. That's a separate diagnostic question, and skipping it is how organisations end up disappointed by a genuinely effective intervention applied to the wrong person.
I've watched all three of these failure modes up close, usually in that order. A board approves the spend on the strength of the aggregate research, a leader is nominated because they're available rather than because they're the right fit, and eighteen months later somebody asks why engagement hasn't moved. The research didn't fail. The step before it, working out which specific leader and which specific gap the coaching is meant to close, never actually happened.
The table below puts the limits of each finding next to the finding itself, so you can see how much weight each can carry in a board paper.
| Finding | What it shows | What it does not show |
|---|---|---|
| Managers account for at least 70 per cent of the variance in engagement across business units (Gallup, 2015) | Who runs the team explains most of the gap between your most and least engaged teams | That coaching a particular manager closes the gap in that manager's team |
| Replacing an employee costs one half to two times annual salary (Gallup, 2019) | A conservative range for the cost of losing someone, which Gallup applies to U.S. businesses | Your own replacement cost, which you should calculate from your own data |
| Effect sizes of g = 0.43 to 0.74 across outcome categories (Theeboom and colleagues, 2014) | Coaching had positive effects on every outcome category tested | That the upper range holds up, since pre-post designs gave larger effects than comparison-group designs |
| g = 0.44 across 11 studies (Cannon-Bowers and colleagues, 2023) | A moderate positive average effect on individual outcomes | Which coaches or approaches work best; very few studies report the coach's certification |
| g = 0.59 across 37 randomised trials (de Haan and Nilsson, 2023) | The best-controlled estimate of a moderate average effect | A figure free of bias: the authors flag significant publication bias, and self-reported outcomes ran higher than observed ones |
| ROI of 600 to 700 per cent (retrospective studies, as reviewed by Grover and Furnham, 2016) | That some organisations believe they saw a large return | A reliable return, since the calculation technique was rather subjective and tenuous |
- Baseline the team-level number, not the company number: Get the current engagement or retention figure for the specific team the leader runs, before coaching starts. Without this, any after-the-fact claim of improvement is a guess.
- Write down the specific behaviour you expect to change: Not "better leadership". Something closer to: stops making every decision alone in their head, or starts delegating real authority rather than just tasks. Vague targets produce vague verdicts.
- Confirm someone above the leader is actually watching for it: If the leader's manager or the board never checks whether the behaviour showed up, the engagement can succeed personally and still register as a non-event organisationally.
- Set the review date at six to twelve months, not six weeks: Outcomes take time to show, which is why evaluation frameworks recommend a follow-up six to nine months after the work. A six-week check-in captures mood, not outcome.
- Decide what happens if the answer is no: A real business case includes the honest version of failure: what you do if the team-level number hasn't moved by the review date. If nobody has an answer to that question, the case wasn't really costed in the first place.
The Distinction Worth Taking From This
If you remember one line from this piece, make it this: leadership coaching earns its budget line not because of a multiplier you can't verify, but because the manager is the largest measurable lever on the single most expensive people-problem most businesses carry, and coaching has controlled research behind its ability to change how that manager actually operates.
That's a smaller claim than "coaching returns seven times your investment." It's also one you can defend in front of a CFO who's read the same recycled statistic you have and has learned, correctly, not to trust it. I'd rather make the smaller, checkable case and have it survive scrutiny than make the bigger one and watch it fall apart the first time someone asks for the source.
If you are weighing this for your own organisation, the honest next step is not a coaching engagement straight away. It is the diagnostic work of finding out where your own version of that 70 per cent gap sits. That is the starting point for the Leadership Capability Architecture work I do with clients, which is also the thinking behind CapabilityAI, and it is a very different first conversation from "let's hire a coach and see what happens." My piece on the Leadership Capability Stack lays out the underlying architecture, and my breakdown of coaching as a leadership style covers the behavioural side: what a leader does differently once coaching changes how they operate day to day.
For the financial framing and an honest look at when it holds up, read is executive coaching worth the investment, where I run the arithmetic with an illustrative example. For the CHRO's view of building the evidence chain, see how a CHRO can prove leadership development works, and for why measurement so often fails, why most companies cannot measure leadership development ROI. If you are choosing a coach, how to choose an executive coach covers the questions to ask. This article is the evidence underneath those conversations, not a replacement for them. If you want to talk through where your own numbers sit, that is a conversation worth having directly.