Yes, on average coaching works: meta-analyses of workplace coaching find positive, moderate effects on individual outcomes such as performance, well-being, and goal attainment, and on organizational outcomes. Whether it works in a given case depends on clear goals, a strong coach-client relationship, and support for change back at work.
This guide summarizes the main research findings in plain language, explains what the numbers mean, and sets out the conditions under which coaching is most and least likely to work.
The main research findings
Three bodies of evidence are most useful for HR leaders and buyers.
- Individual outcomes. A meta-analysis by Theeboom, Beersma and van Vianen (2014) found significant positive effects on performance and skills, well-being, coping, work attitudes, and goal-directed self-regulation. Effect sizes ranged from g = 0.43 for coping to g = 0.74 for goal-directed self-regulation.
- Organizational outcomes. A meta-analysis of 17 studies by Jones, Woods and Guillaume (2016) found a positive effect of coaching on organizational outcomes (d = 0.36). It found no difference between face-to-face coaching and e-coaching, stronger effects for internal coaches than external ones, and no clear effect of engagement length.
- The relationship. A longitudinal study by Molyn, de Haan, van der Veen and Gray (2022) found the working alliance between coach and client predicted later coaching effectiveness and wellbeing.
What the numbers mean in plain terms
Effect sizes such as g and d describe how far the average coached person moved compared with people who were not coached, in standard units. As a rough rule of thumb used across social science, around 0.2 is small, around 0.5 is moderate, and around 0.8 is large. Coaching sits mostly in the moderate range, with its strongest effects on how people pursue goals.
| Question | What the evidence suggests | Source |
|---|---|---|
| Does coaching improve individual outcomes? | Yes, moderate positive effects across five outcome types | Theeboom et al. 2014 |
| Where is the effect strongest? | Goal-directed self-regulation (g = 0.74) | Theeboom et al. 2014 |
| Does coaching help the organization? | Yes, positive effect on organizational outcomes (d = 0.36) | Jones et al. 2016 |
| Is virtual coaching as good as in person? | No difference found between face-to-face and e-coaching | Jones et al. 2016 |
| Are external coaches better than internal? | Internal coaches showed stronger effects | Jones et al. 2016 |
| Does longer coaching work better? | Duration did not clearly change results | Jones et al. 2016 |
| Does the coach-client relationship matter? | Yes, working alliance predicts effectiveness | Molyn et al. 2022 |
| Does AI coaching work? | Early trial: improved goal attainment only | Terblanche et al. 2022 |
When coaching is most likely to work
- The goal is specific and owned by the person. Coaching is strongest on goal pursuit, so a clear goal plays to its strength.
- The person wants to change. Coaching assigned as a last warning rarely works.
- The relationship is strong. Fit is part of the mechanism, not a nice extra.
- The environment supports change. A manager who notices and reinforces new behavior makes it stick.
- Each session ends in a commitment that is reviewed next time.
When coaching is least likely to work
- The real problem is structural: unclear roles, conflicting targets, or decision rights nobody agreed.
- Goals are vague, such as “be more strategic,” and never translated into behavior.
- The person is coached in isolation, while the team around them stays the same.
- Nobody checks progress, so nobody can say whether it worked.
The first point deserves emphasis. DorWay sees organizations as a Network of Agreements™, and many problems that look like individual performance gaps are actually unclear agreements between people. Coaching one leader will not fix an agreement that involves five. In those cases, working on team agreements and accountability across the team is the better use of budget.
How to find out if coaching worked for you
Research tells you about averages. To know about your own program, set a baseline and check against it. Our guides to measuring coaching effectiveness and coaching ROI cover methods. A structured tool such as the PACT™ Leadership Assessment can provide the before-and-after picture. For a summary of benefits by level, see benefits of coaching.
FAQ
Is there scientific evidence that coaching works?
Yes. Meta-analyses of workplace coaching, including Theeboom and colleagues (2014) and Jones, Woods and Guillaume (2016), find positive effects on individual and organizational outcomes.
How effective is executive coaching?
Most studies of workplace coaching, including many with managers and executives, find moderate positive effects. Results vary widely with goal clarity, the coach-client relationship, and support at work.
Why does coaching sometimes fail?
Common reasons are vague goals, poor fit between coach and client, a person who does not want to change, no support from their manager, and problems that are structural rather than individual.
Does online coaching work as well as in-person coaching?
The Jones, Woods and Guillaume meta-analysis found no difference in effectiveness between face-to-face coaching and e-coaching.