White paper · November 2026

What was worth it: measuring the return on executive coaching

Most ROI conversations happen at the end of an engagement, as an argument. This paper is about doing that work at the start, so the end is a report. It is written from the grids we have kept since 2018.

In December 2018 an analytics manager in financial services sent me her competency grid. Against almost every line she had rated herself 1 out of 4. Her peer, who had watched her work every day, gave her 3s and 4s.

That sheet changed what we measured in that engagement. It also changed how I think about the return on coaching. Before anyone asks “what did we get for the money?”, someone has to ask “where does this leader actually stand, and who says so?”

The cases below are real and masked. No client is named. Where a figure was told to us rather than measured by us, we say so. The figures for the practice as a whole, and their limits, are set out on Evidence and methodology.

Three ledgers, not one

Coaching returns value to three parties, and each keeps its own ledger. Mixing them is where most ROI claims go wrong.

LedgerWho reads itWhat goes in itHow it is evidenced
Return to the organisationCFO, business headMovement in agreed business measures: turnaround time, quality score, audit findings, team outputThe organisation's own MIS, before and after, with an honest share credited to coaching
Return on the sponsor's expectationReporting manager, CHRODid we get the leader we needed: the successor who is ready, the manager who now delegatesThe sponsor's judgement, written down at the start and asked again at the close
Return to the leaderThe coacheeReadiness for the next role, promotion, compensationRatings by others, role movement, and compensation as the leader reports it

A sponsor who pays for coaching is buying from the first two ledgers. A leader who pays for themselves is buying from the third. Most of our coachees paid for themselves; organisations sponsored the rest. So we have had to keep all three honestly, and not let a figure from one ledger pass as proof for another.

The method we have used since 2018

Every engagement starts with the role, not the person. In our first mail to a new coachee in 2018 the instruction was plain: start from the job description, the skills it needs, the ideal level and the industry benchmark. Only then look at where you stand.

We put that on one sheet, first called the Competency Self-assessment Grid and from 2019 N-GAGE, Nirvedha's Goal Achievement Grid Enabler.

  1. The leader names two to four goals. Each becomes a broad skill area with a weight; the weights add to 100%.
  2. Each area splits into elementary competencies, again weighted to 100%.
  3. For every competency the leader writes the critical task, the standard or benchmark, how it will be assessed, the evidence today, and the gap.
  4. Three ratings, each on a 1 to 4 scale: the leader's own, a peer's (later the senior's too), and the benchmark for the role.
  5. The sheet rolls these up into one figure per rater. We called it Present Job Market Value; in CoachOS today it is the Role Readiness Index.
Readiness = Σ over areas [ area weight × ( Σ competency weight × rating ) ÷ 4 ]

One row from a 2023 engagement shows what “evidence” means here:

CompetencyCritical taskStandardEvidence
Industry knowledgeDisbursement processing with maximum quality and least turnaroundDisbursal in 3 hours, loan booking in 1 hour, quality score 95%Monthly MIS shows the turnaround; quality scorecard averages 95%

The evidence column is where coaching meets the business. It ties a behaviour to a number the organisation already counts.

The first finding: the gap between self and others

The first number worth having is not a business number. It is the distance between how leaders rate themselves and how the people around them rate them. In our grids that gap runs both ways, and it is large.

Leaders misread themselves in both directions Readiness by competency area, % of the top level, from the N-GAGE grid The leader A peer Benchmark for the role 0% 25% 50% 75% 100% Case A · analytics manager, financial services, Dec 2018 Overall readiness Overall readiness: benchmark 53% Overall readiness: peer 65% Overall readiness: the leader 30.7% 31 65 Data stories that win buy-in Data stories that win buy-in: benchmark 67.5% Data stories that win buy-in: peer 90% Data stories that win buy-in: the leader 25% Being an empathetic leader Being an empathetic leader: benchmark 65% Being an empathetic leader: peer 72.5% Being an empathetic leader: the leader 32.5% Case B · head of operations, lending company, Apr 2023 Overall readiness Overall readiness: benchmark 89.1% Overall readiness: peer 80% Overall readiness: the leader 96.9% 97 80 Solution orientation Solution orientation: benchmark 95% Solution orientation: peer 82.5% Solution orientation: the leader 100% Adaptability and resilience Adaptability and resilience: benchmark 85% Adaptability and resilience: peer 75% Adaptability and resilience: the leader 100% Innovation Innovation: benchmark 88.75% Innovation: peer 75% Innovation: the leader 100% Time management Time management: benchmark 87.5% Time management: peer 87.5% Time management: the leader 87.5%
N-GAGE grids returned by two coachees, December 2018 and April 2023. Names withheld.

Case A rated herself 31% ready. Her peer put her at 65%, above the benchmark for her role. She was not short of skill; she was short of belief, and it showed most in the very thing she wanted to grow, telling stories with data. Coaching her on technique would have wasted the fee.

Case B rated himself 97%. His peer put him at 80%, and the role benchmark sat at 89%. The gap was widest in innovation and adaptability. On time management, the area he had chosen to work on, all three ratings agreed.

Without the second and third rating, both engagements would have been aimed at the wrong target.

That is the first return a sponsor gets: coaching pointed at the right gap.

From the gap to action

A gap on a sheet changes nothing until it becomes a dated action that someone can see done. Our sequence has stayed the same since 2018.

  1. Goals, made sharp. We mine, refine, define and then “shine” each goal into one line of eight to ten words, with its measure inside it.
  2. Strategies. Five to eight per goal, put in a sensible order. A goal without strategies is still a wish.
  3. Time-bound actions. Each strategy gets an action, an owner, a date, a status and the date it was actually done.
  4. Weekly tracking. A simple tracker, week by week, that the leader fills and the coach reads before each session.
  5. A second rating. The peer and the senior rate again, and we read the new gap.

Case B shows how quickly this moves when the gap is agreed. Within three weeks he had six actions on prioritisation and delegation, all marked implemented between 20 March and 10 April 2023:

CompetencyStrategyWhat he did
Work prioritisationOne slot for allA fixed daily slot for routine queries from his teams
Work prioritisationTime-bound discussionsTime limits on meetings, one issue settled at a time
Work prioritisationWork listingHis daily schedule shared with his direct reports and his own manager
DelegationBreak down tasksWork split by responsibility and handed to the right person
DelegationSingle point of contactOne owner made answerable for each task
DelegationDirecting peopleCross-team queries sent to the process owner, not to him

He also drew a delegation chart and brought his own boss into it. That is the system getting ready for the change, not only the leader.

Putting a number on it

A return is only a return once the fee is taken off. Many coaching ROI claims stop at the benefit and never show what was paid. We show both, and we discount the benefit before we count it.

The discount follows Jack Phillips' ROI Methodology. The people who saw the change are asked two questions: what share of this improvement came from the coaching, and how sure are you of that share? Both answers reduce the figure.

StepIllustrative figure
Improvement in the agreed business measure over the year, valued by finance₹40 lakh
Share the sponsor and peers attribute to coaching (average)40%
Their confidence in that share (average)60%
Benefit credited to coaching: 40 × 40% × 60%₹9.6 lakh
Coaching fee, six months₹5 lakh
Net return: 9.6 − 5₹4.6 lakh
ROI: net return ÷ fee92%
ROI % = [ ( benefit × attribution × confidence ) − fee ] ÷ fee × 100

These figures are an example, not a client result. The point is the shape: a modest, traceable number that a CFO can check line by line is worth more than a large one borrowed from a published study.

Two rules keep it honest. The business measure is agreed and baselined before the first session, from the organisation's own MIS. And it must be a measure the leader can actually move; if it depends mostly on policy or on another function, it measures the organisation, not the leader.

The third ledger, told honestly

Across the 103 leaders in our promotion cohort, the annual salary increase each reported at the point of promotion adds up to ₹32 Crore (USD $3.8M+). That is the return to the leader, and for those who paid their own fee it is the return that mattered most.

We hold that figure to the same standard as the rest of this paper, so here is what it is and what it is not.

Why share it at all? Because a number the leaders themselves gave us, across 103 people, is a real signal. It is just not proof, and we will not present it as proof. The full definition, window and exclusions are on Evidence and methodology.

The better practice is to record compensation at the start, with the leader's consent, and ask again a year after the last session.

Durability: measure the half-life, not just the peak

The reading that counts comes a year after the coach has left. Leadership behaviour is like any other capital: it wears down unless someone maintains it.

Case A shows this from the inside. In 2018 her grid caught a leader who undervalued herself by more than 30 points. She grew, moved on, and came back to us in 2025 for a new transition into sales and business development in AI. Her 2025 goals again included dropping self-doubt and communicating with confidence. The skill had grown; the old belief had come back under a new kind of pressure.

That is not a failure of the first engagement. It is what change looks like over seven years. But it tells a sponsor something a close-out report cannot: some gains hold, some need maintaining, and you only learn which by asking again later.

So we add one date to the contract: a stakeholder pulse 12 months after the last session, using the same grid and the same raters. If the change is still visible to the team, that is the return confirmed. If it has faded, it tells the sponsor the system around the leader was never ready, and that is worth knowing too.

An honest note on our own record: our grids hold strong starting and mid-point ratings. Full end-of-engagement and 12-month rescores are the part we have done least well, and the part we most need to strengthen.

Ethics and confidentiality

Measurement gives the sponsor visibility. It must not give the sponsor the coaching room.

On putting a share of the fee at risk against results: it can sharpen focus, and some buyers ask for it. It can also pull the coach towards the sponsor's number and away from what the leader actually needs, as Case A shows: her real gap was belief, not a business metric. Where any fee is at risk, the measures and the raters must be agreed by all three parties at the start, and the coach must keep the right to say a measure belongs to the system, not the leader.

A protocol for sponsors and coaches

Most of the measurement happens before the first session. Everything after it is reading the same grid again, with the same people.

The work of measurement is done before the first session Before session one Role, JD and benchmark Goals and weights Baseline from 3 raters Business measure, MIS Sponsor's expectation CoachOS: N-GAGE, 3-party e-signed contract During Dated actions per goal Weekly tracker Rescore at month 3 Resync if goals drift CoachOS: session notes, portal, action status At the close Same grid, same raters Attribution, confidence Fee and net return Leader approves report CoachOS: sponsor report, Role Readiness Index A year on Stakeholder pulse Role and compensation Read the half-life CoachOS: planned
Nirvedha measurement protocol, four phases.

We built CoachOS so that a coach does not have to rely on spreadsheets passed around by email, as we did from 2018. The grid, the three-party agreement, the actions and the sponsor report sit in one record per engagement. The sponsor report goes out only after the leader approves it. The method stands without the software; the software just makes it hard to skip a step.

The question for the sponsor

The analytics manager in Case A did not need more skill. She needed someone to show her the gap between her own rating and her peer's. The operations head in Case B did not need more confidence. He needed the same sheet to show him the opposite gap.

In both, the most valuable number was cheap to collect, and it came before a single coaching hour was spent.

So before you ask your coach what return you will get, ask yourself one question. Do you know how the leader you are about to invest in rates themselves, and how far that is from what their team sees?

If you don't, you are not yet ready to measure the return. You are only ready to argue about it later.

Asked in these words

Can you really measure the ROI of executive coaching?

Yes, if the measurement is set up before the first session. Agree a business measure the leader can actually move, take its baseline from the organisation's own MIS, rate the leader's behaviour through self, peer and benchmark, then at the close ask the sponsor and peers what share of the change came from coaching and how sure they are. Discount the benefit by both answers and take off the fee. What is left is a modest number a CFO can check line by line.

What should a company measure before coaching starts?

Three things: the leader's starting readiness for the role as rated by the leader, a peer and the role benchmark; one business measure from the organisation's own records that the leader can influence; and the sponsor's expectation in plain words, such as a successor ready for a named role. Without a baseline, no claim at the end is credible.

Why compare a leader's self-rating with a peer's rating?

Because the gap between them is often the real coaching target, and it runs both ways. In one of our engagements a manager rated herself 31% ready while her peer rated her 65%; in another a head of operations rated himself 97% against his peer's 80%. The first needed belief, the second needed a truer view of himself. Coaching aimed without that gap is aimed at the wrong thing.

How long after coaching should results be checked?

At the close, and again 12 months after the last session, with the same grid and the same raters. Some gains hold and some fade when the system around the leader is not ready for the change. Only the later reading tells a sponsor which.

If you are about to sponsor coaching and want the baseline set before the first session, that is the conversation to have first.

Start a conversation

Evidence and methodology: how every figure on this site is counted

How to hire a coach

All cases are masked and no client is named. The ROI arithmetic is illustrative. The ₹32 Crore compensation figure is as reported by the leaders, covers 1 January 2017 to 31 May 2026, and is not verified against payroll.