Measuring the impact of coaching and consulting engagements

Sponsors want to know whether an engagement made a difference, and many other factors shape the outcome. The verified entries below show what the research can and cannot support.

• Coaching works on average, but less dramatically than headline figures suggest. Randomized trials show a moderate average effect, and the authors report publication bias, so the true effect is probably smaller (r-0001). Earlier workplace studies point the same way (r-0002).
• How you measure changes the answer. Self-reported results run higher than observed ones (r-0001), and effects are largest on coachees’ own goals and smallest on performance rated by others (r-0004).
• Business results are rarely measured. None of the recent workplace studies measured results-level outcomes (r-0003), and a review of 110 executive coaching studies found most relied on self-report and questioned ROI as a measure, recommending several data sources instead (r-0006).
• Relationship quality is a useful process signal, not proof of impact. A strong coach-client working alliance is consistently associated with better outcomes (r-0005).

Each entry separates what a study found from what its authors and our editors make of it, and carries an evidence grade explained on the methodology page.