Does workplace coaching improve performance

TL;DR

Does Workplace Coaching Actually Improve Performance? The Evidence

  • Meta-analyses agree: moderate, real effects - g around 0.4 to 0.6 depending on outcome.
  • The famous ROI figures (5.7x, 788%) rest on self-report and single cases; treat them as marketing.
  • Internal coaches outperformed external ones in the 2016 meta-analysis - context beats prestige.
  • Session count showed no effect on outcomes; what the coach knows about the person appears to matter more than hours.

Coaching may be the largest workplace intervention ever adopted on testimonial evidence. Organisations spend billions on it annually, and most buyers, if asked for proof, would produce a satisfied executive rather than a study.

The studies exist. They are better news for coaching than sceptics expect, and worse news than the industry's brochures suggest - and the details point somewhere genuinely useful.

Does workplace coaching really work?

Yes, with moderate and consistent effects. Theeboom and colleagues' 2014 meta-analysis found significant positive effects across every outcome category measured - performance and skills at g = 0.60, wellbeing at g = 0.46, coping at g = 0.43. Jones, Woods and Guillaume's 2016 meta-analysis confirmed positive organisational outcomes, and a 2023 meta-analysis in Frontiers in Psychology found an overall effect of g = 0.43. These are moderate effects - genuine, useful, and smaller than the industry's marketing implies.

An effect of g = 0.5 or so means a typical coached person ends up better off than roughly two-thirds of an uncoached comparison group. Worth paying for; not a transformation machine. The caveats belong alongside: the studies lean on self-report, several use within-subject designs that flatter effects, and the pool of rigorous trials is still small - Theeboom's landmark analysis drew on just 18 studies.

The ROI theatre

What is the ROI of executive coaching?

The famous figures do not survive inspection. The oft-quoted 5.7 times return comes from a 2001 study of 100 executives using self-reported estimates with no control group; the 788% ROI claim comes from a single company case study of 30 people. Anthony Grant, one of coaching psychology's founding researchers, argued ROI is a poor measure of coaching success - too dependent on unverifiable self-report - and proposed wellbeing and engagement measures instead. The defensible claim is moderate effect sizes, not spectacular percentages.

The 5.7x figure has been repeated in coaching proposals for over two decades, usually without its provenance: executives were asked to estimate the monetary value of their own coaching, and the estimates were averaged. Ask people who chose an expensive intervention whether it was worth it, and arithmetic does the rest. Grant's critique lands because it comes from inside the field - a researcher who believes in coaching arguing that its sales pitch undermines its science.

The findings buyers actually need

Two results from the 2016 meta-analysis deserve more attention than they get.

Are internal coaches as effective as external coaches?

The most surprising finding in the coaching literature says internal coaches may be more effective. Jones, Woods and Guillaume's 2016 meta-analysis found coaching by internal coaches showed substantially larger effects than coaching by external coaches. The plausible mechanism: internal coaches know the context - the people, the politics, the real constraints - and can coach inside it, and they are present for the follow-through rather than gone after six sessions. Context, in coaching, appears to beat prestige.

The second finding is stranger still: the 2023 meta-analysis found no relationship between the number of sessions and the size of the outcome, and no significant difference between virtual and face-to-face delivery. Take the two findings together and a pattern emerges. What predicts coaching outcomes is not the hours bought or the medium used, but what the coaching works with - how well the coach understands this particular person in this particular context. An internal coach starts with that understanding. An external coach spends the first three sessions acquiring it, at day rates.

What the coach knows

That pattern points at the real scarce resource in coaching: accurate knowledge of the person. A coach who knows how their coachee prefers to make decisions, take feedback, handle conflict and structure their work can go straight at the friction. A coach who does not spends sessions finding out - or worse, coaches the person they assume is in the chair.

This is the gap Sariio was designed to close, and it is why coaching sits at the centre of the suite rather than at its edge. Every coachee's preference map - 60 work preference pairs on continuous scales, from a ten-minute survey retaken at least twice a year - is available to the coaching conversation from the first minute, along with the team context around them. The AI orchestrates the preparation an internal coach would otherwise assemble by hand: where this person's preferences sit, how they differ from the people they work with, and what has moved since the last reading - because the delta between readings is often the coaching agenda in miniature. The evidence says coaching works through understanding and follow-through. A living map supplies the first and gives structure to the second - it never replaces the coach, and no reading ever becomes a verdict on the person. Related: the GROW model, reviewed and the 70 percent manager.

See what a coach sees in a preference map - the individual survey is free.

Sources