What the controlled research shows
Most coaching statistics come from client satisfaction surveys, which tell you how buyers feel rather than whether coaching caused anything. The stronger evidence comes from meta-analyses pooling controlled studies, and it is more modest and more trustworthy.
Those reviews consistently find positive effects across five areas: performance and skills, wellbeing, coping, work attitudes, and goal-directed self-regulation. The effect sizes are moderate. Coaching reliably helps; it does not reliably transform.
Why it works — the mechanisms with support
- Goal-setting. One of the most robust findings in organisational psychology: specific, difficult, accepted goals outperform vague intentions. Coaching is largely a structured goal-setting and review mechanism.
- Deliberate practice with feedback. Skill improves through targeted repetition against accurate feedback. Most leaders get repetition without feedback, since honest feedback thins out as authority rises.
- Self-efficacy. Belief in your capacity to execute predicts persistence and performance, and it is built through mastery experiences rather than encouragement — which is why coaching works on real situations.
- Accountability. Committing to a specific action in front of another person materially raises follow-through.
- Reflection. Experience alone teaches little; experience plus structured reflection is where learning happens. Senior calendars leave almost no room for the second part.
What the evidence does not support
Being straight about this matters more than the positive findings, because the field is full of overclaiming.
- Specific ROI multiples. The 700% and 788% figures come from self-reported surveys, not controlled measurement.
- Neuroscience claims. Coaching marketing frequently invokes brain rewiring. The underlying neuroplasticity is real; the claim that a particular coaching method drives specific neural change is not established.
- One method beating another. Comparative evidence between coaching approaches is thin. What predicts outcomes most consistently is the working relationship, not the model.
- Personality-matched coaching. Intuitive, weakly supported.
The most reliable finding
Across the research, the strongest single predictor of coaching outcomes is the quality of the coach–client relationship — the same finding that dominates psychotherapy outcome research. Credentials set a floor. Fit determines the result.
Which has a practical consequence: use the free first session as an actual assessment, and speak to more than one coach before deciding.
Frequently asked questions
Is executive coaching evidence-based?
Yes, with moderate rather than dramatic effects. Meta-analyses of controlled studies find consistent positive results for performance, wellbeing, coping, work attitudes, and goal attainment. The dramatic ROI figures circulating in marketing come from client surveys and are considerably weaker evidence.
How does coaching actually change behaviour?
Through mechanisms with independent support: specific goal-setting, deliberate practice against accurate feedback, building self-efficacy through real mastery experiences, accountability, and structured reflection. None is exotic; the value is that coaching supplies all of them consistently.
Is there real neuroscience behind coaching?
Neuroplasticity is real, but claims that a specific coaching technique produces specific neural changes go beyond the evidence. Be sceptical of coaching marketing leaning heavily on brain imagery.
Which coaching method is most effective?
The comparative evidence between methods is thin. What predicts outcomes most reliably is the coach–client working relationship, which is why fit matters more than methodology when choosing.