Executive Coaching Research: What Meta-Analyses Show and Don't

What peer-reviewed meta-analyses and reviews report about the effectiveness of executive and workplace coaching, and the limits their own authors acknowledge.

Exec Coach · Last reviewed October 23, 2026 · 6 min read

If you are deciding whether to fund executive coaching, you will hear plenty of confident claims about its value. Some come from careful research. Many come from testimonials, satisfaction surveys or return-on-investment figures with little explanation of how they were calculated.

This guide looks at what peer-reviewed research syntheses actually report about coaching in organizations, with a focus on leaders and executives. It names each study, explains what it measured, and spells out the limits the authors themselves acknowledge. The honest summary is encouraging but not conclusive: the evidence points to real benefits, yet it is thinner and messier than marketing materials suggest.

Why "does coaching work?" is a hard question

Coaching is difficult to study. Every engagement is tailored, so two leaders "receiving coaching" may get very different interventions. Goals differ from person to person, which makes common outcome measures awkward. Senior leaders are a small population, so samples are small. And randomly assigning executives to receive or not receive coaching is rarely practical inside a real company.

Researchers deal with this by pooling studies. A systematic review gathers and appraises the available studies on a question. A meta-analysis goes further, combining their results statistically into an average effect size, a standardized measure of how much difference the intervention made. Larger numbers mean larger effects, but the figures are only as good as the studies behind them.

Workplace coaching: Jones, Woods and Guillaume

A meta-analysis by Rebecca Jones, Stephen Woods and Yves Guillaume in the Journal of Occupational and Organizational Psychology (2016) combined 17 studies of workplace coaching delivered by internal or external coaches. It deliberately excluded coaching by a person's own manager and peer coaching.

The authors found positive effects overall (δ = 0.36) and across outcome types: skill-based outcomes (δ = 0.28), affective outcomes (δ = 0.51), and individual-level results (δ = 1.24).

Two moderator findings are useful for buyers. Effects were stronger for internal coaches than for external coaches, and studies that used multisource feedback, such as 360-degree reviews, showed smaller positive effects. The authors found no difference by format (face-to-face versus blended face-to-face and e-coaching) or by duration, whether measured by number of sessions or length of the intervention. Note that this study covers workplace coaching broadly, not only executives.

Executive coaching in controlled studies: Nicolau and colleagues

A 2023 meta-analysis in Frontiers in Psychology by Andreea Nicolau, Octav Sorin Candel, Ticu Constantin and Ad Kleingeld focused specifically on executive coaching. It reviewed 20 studies that used control groups and measured people before and after coaching, a stronger design than simple before-and-after surveys.

They grouped outcomes into three types. Effects were largest for behavioral outcomes (Hedges' g = 0.73), followed by personal characteristics (g = 0.51) and attitudes (g = 0.34). They also found significant positive effects for specific outcomes such as self-efficacy, psychological capital and resilience. The number of sessions did not moderate the results; the length of the coaching program mattered only for attitude outcomes.

The authors flag real limits. Interventions, measures and outcomes varied considerably across studies, and some findings rest on as few as three studies, which makes them less stable.

A systematic review and its warnings: Grover and Furnham

Simmy Grover and Adrian Furnham's 2016 systematic review in PLOS ONE examined 52 quantitative studies, published since 2003, of business, executive and leadership coaching delivered by professional coaches. They concluded that coaching is an effective tool that benefits organizations, with generally positive findings for outcomes such as self-efficacy and goal attainment, while results for objective job performance were mixed.

The review is just as valuable for its criticisms. The authors report that just under half of the studies had a control group, that only five followed people up months after coaching ended, and that there is no consistency in outcome measures across the field. They warn that self-report measures, which dominate the literature, are not appropriate for judging objective organizational outcomes, and they note that some single data sets had been split across multiple papers, which can inflate the apparent volume of evidence.

The relationship matters: Graßmann and colleagues

If coaching helps, what drives the benefit? One well-studied factor is the working alliance, meaning the quality of the collaborative relationship between coach and client. Carolin Graßmann, Franziska Schölmerich and Carsten Schermuly's meta-analysis in Human Relations (published online in 2019) pooled 27 samples covering 3,563 coaching processes.

They found a moderate and consistent relationship between a strong working alliance and client outcomes (r = .41), and a negative relationship with unintended negative effects of coaching (r = -.29). The link held regardless of the type of client, the coach's level of expertise or the number of sessions. Because these are correlations, they show that good relationships and good outcomes go together, not that one strictly causes the other. Still, the finding supports taking the leader's choice of coach seriously.

What the evidence does not tell you

Read together, these studies support a cautious "yes, coaching can help." They do not support several common claims.

They do not establish a reliable financial return. None of the syntheses above reports a validated ROI figure for executive coaching, so treat precise multiples with skepticism unless you can see how they were calculated.
They do not tell you how a particular coach will perform. Averages hide wide variation among coaches, clients and settings.
They mostly measure short-term change. Few studies follow leaders for long after coaching ends.
They rely heavily on what coachees report about themselves. Ratings from managers, peers or business data are less common.
They involve different kinds of coaching. Results for workplace coaching in general may not translate directly to senior executives, and vice versa.

It is also worth remembering that effect sizes depend on how outcomes are measured. The Jones finding that studies using multisource feedback showed smaller effects is a reminder not to treat any single number as a fixed property of coaching.

How to use the evidence as a buyer

The research suggests a few practical moves for HR and people teams:

Set specific, observable goals so you can tell whether anything changed, rather than relying on satisfaction alone.
Give leaders a real choice of coach, since relationship quality is consistently linked with outcomes.
Collect more than self-report where you can, such as manager observations or stakeholder input, while respecting confidentiality.
Be candid in internal reporting about what you can and cannot attribute to coaching.

For a step-by-step approach, see our guide Evaluating an Executive Coaching Program: Metrics, ROI and Limits.

Key takeaways

Peer-reviewed meta-analyses report positive average effects of workplace and executive coaching, including in studies with control groups.
Nicolau and colleagues found the largest effects of executive coaching on behavioral outcomes and smaller effects on attitudes.
The field still leans heavily on self-report, small samples and short follow-up, as Grover and Furnham document.
A strong coach-client working alliance is consistently associated with better outcomes.
No synthesis here supports a precise, general ROI figure for executive coaching, so be wary of anyone who quotes one without showing the math.

Sources

Spotted something out of date? Let us know.