Executive Coaching ROI Measurement: Evidence Before Claims
Nour
Founder, Coachful
Executive coaching ROI measurement asks whether benefits attributable to an engagement justify its full cost. The arithmetic is usually easier than establishing what changed, why it changed and whether the change has a defensible monetary value. A polished percentage can conceal weak answers to all three questions.
This guide follows one fictional engagement from its evaluation brief to a completed sponsor report. Its purpose is to help you test a financial claim before using it. It covers measurement design and interpretation, rather than reviewing published coaching-return studies or supplying a quarterly audit pack. Every case, amount and result below is illustrative, not a Coachful customer outcome.
Separate four claims that require different evidence
“The executive valued coaching” and “coaching generated savings” are different statements. Useful evaluation keeps the distance between them visible.
| Claim | Evidence needed | What does not establish it |
|---|---|---|
| The engagement happened | Attendance and delivery records | A completed session does not demonstrate learning |
| A behavior changed | Comparable observations of the agreed behavior, with opportunities to use it | A positive satisfaction score |
| Coaching contributed to an operational change | A credible explanation tested against timing, implementation and alternative causes | A favorable before-and-after number alone |
| The attributable benefit exceeded cost | Defensible attribution, monetary valuation, period and complete cost calculation | Putting a currency symbol beside every benefit |
HM Treasury’s Magenta Book, sections 1.8.1–1.8.3, distinguishes evaluation of implementation, impact and value for money. It is government evaluation guidance, not evidence that coaching achieves a particular return. The distinction is useful here: successful delivery, an observed outcome and a worthwhile investment each require their own examination.
Start with the decision the report must inform. Renewing one executive’s engagement, changing a leadership program and making a public marketing claim have different consequences. A short reflective review might inform the first decision; it cannot automatically justify the third. Choose effort proportionate to the decision, while keeping the claim proportionate to the evidence.
Write the measurement agreement before collecting answers
- Name the intended work change. Replace “better leadership” with a behavior someone can recognize in a defined situation.
- Define the operational measure. State the unit, population, data source, calculation and period. Decide what is excluded.
- Record the starting position. Collect a baseline using the same definition you will use later. Preserve missingness and unusual events.
- Identify other influences. List planned reorganizations, new systems, staffing changes and seasonal pressures before results appear.
- Agree who may see what. Separate private coaching content, participant-approved examples and organization-owned operational records.
- Set a decision and review date. Decide what favorable, unfavorable or inconclusive evidence would mean. Do not make “renew regardless” the unspoken rule.
The executive coaching intake guide develops the client and sponsor conversation behind this agreement. A sponsor’s payment does not itself settle the sharing boundary. The ICF Code of Ethics, standards 2.1–2.5, addresses confidentiality, information-exchange agreements, records and responsibilities when using technology. Apply the actual agreement and relevant obligations to the engagement.
A practical measurement question is “Who will verify this?” The executive may describe an experiment; an operations analyst may own meeting records; finance may approve a valuation. Do not ask the coach to certify financial results using private reflections they cannot independently verify. Name the owner of each evidence stream and the boundary of their judgment.
Worked engagement: reducing repeated coordination
In this fictional case, Amina leads a six-person management group. Decisions repeatedly return to the whole group because ownership and escalation conditions are unclear. Her sponsor wants less coordination time without an increase in missed handoffs. They agree six coaching sessions over twelve weeks, followed by a sponsor review already included in the engagement.
The proposed mechanism is specific: Amina practices ending a decision discussion with a named owner, a decision boundary and a condition for escalation. Managers can then proceed without convening everyone again. Coaching may help her practice that behavior. It cannot by itself change reporting software, staffing or the organization’s approval rules.
1. Complete the brief
| Field | Agreed fictional entry |
|---|---|
| Decision | Close, redesign or renew this engagement after twelve weeks |
| Behavior | End an eligible decision discussion with a recorded owner, boundary and escalation condition |
| Time measure | Weekly total person-hours in recurring coordination meetings attended by the same six management roles |
| Baseline | Four weeks before coaching, averaging 36 person-hours a week |
| Follow-up | Ten observed weeks after the initial practice period, averaging 24 person-hours a week |
| Quality check | Missed handoffs logged using the same definition in both periods |
| Evidence owners | Operations analyst: meeting and handoff records. Executive: approved behavior examples. Finance: valuation and costs |
| Sharing | Sponsor receives the measurement brief and approved work examples, not session transcripts or private reflections |
Person-hours matter. A one-hour meeting attended by six people consumes six person-hours; shortening it by fifteen minutes releases 1.5 person-hours. Counting the calendar reduction as six saved hours would multiply the result incorrectly. The analyst retains actual attendance and duration records, including canceled meetings and replacements, so the calculation can be reproduced.
The follow-up covers the same roles and recurring meeting categories. Ad hoc meetings are outside the primary measure but tracked separately for displacement. Otherwise, a leader could appear to save time simply by moving the same conversations into uncounted calls.
2. Check behavior and unintended effects
The analyst reviews ten eligible decision records in the baseline and ten in the final review window. Three baseline records and eight follow-up records contain all three agreed elements. The sample is small and selected from the specified windows; the report describes the records reviewed rather than declaring a permanent eighty-percent leadership competency.
Amina supplies two approved examples. In one, the named owner uses the agreed boundary without returning to the full group. In the other, a cross-team dependency triggers the escalation condition and the group meets again. The second example is appropriate use of the boundary, even though it creates meeting time. The goal is useful coordination, not suppressing escalation.
Missed handoffs average two a week in both periods. That offers no observed improvement on this quality measure. It also does not prove that shorter coordination is harmless in every situation. The analyst finds no rise in recorded ad hoc person-hours during the measured weeks; unrecorded informal conversations remain a limitation.
3. Test the first explanation
The raw time difference is 36 minus 24, or twelve person-hours a week. Across ten follow-up weeks, that is 120 fewer recorded person-hours than maintaining the baseline rate would imply. This is an observed comparison with a fixed baseline, not yet a causal coaching estimate.
During the same period, the company introduced a new reporting dashboard. Another management team, which did not receive this coaching engagement, moved from 32 to 26 weekly coordination person-hours. Its six-hour reduction makes a “coaching caused all twelve hours” explanation less persuasive.
Subtracting the comparison team’s reduction gives a six-hour differential change: twelve minus six. Over ten weeks that is sixty person-hours. However, the teams handle different projects, and there are insufficient earlier observations to assess whether their trends were comparable. The differential is a descriptive check, not an accepted causal effect.
The Magenta Book Annex A, sections A1.4 and A2.7, describes contribution analysis and difference-in-difference. The latter depends on assumptions about how groups would have moved without the intervention. A convenient neighboring team is not automatically a valid counterfactual. Contribution analysis tests an explanation and alternatives; it does not simply award coaching a guessed percentage.
The team therefore retains three findings: the target behavior appears more often in the reviewed records; coordination time fell; and the dashboard and other differences prevent isolation of coaching’s share. These statements can coexist without forcing a single causal verdict.
Value the result without inventing cash savings
Finance uses EUR75 per person-hour as a fictional internal resource valuation. It is an agreed planning assumption, not an advertised salary benchmark or cash saving. Multiplying 120 hours by EUR75 gives EUR9,000 of modeled capacity value. No salaries were reduced, no overtime payments were avoided and no incremental sales were verified.
That capacity could still matter. Managers might use time for customer follow-up, planning or unresolved work. To substantiate that additional use, specify and observe it. Do not immediately count the same hours as both EUR9,000 of released capacity and EUR9,000 of new revenue. A resource opportunity and a realized financial receipt require different evidence.
For this case, the report uses “modeled capacity value” consistently. Its cash-savings line reads “none verified in the evaluation period.” It does not say the engagement produced no value. It says the financial evidence does not establish a cash return.
Include the full resource cost
| Cost category | Fictional calculation | EUR |
|---|---|---|
| Coach fee, including agreed reviews | Actual contracted amount in the case | 6,000 |
| Amina’s session and preparation time | 12 hours × 100 | 1,200 |
| Administration and evaluation | 4 hours × 50 | 200 |
| Additional group practice time | 20 person-hours × 50 | 1,000 |
| Total resource cost | 6,000 + 1,200 + 200 + 1,000 | 8,400 |
The practice hours are additional to the coordination meetings being measured. This avoids deducting the same time twice. The example assumes no other incremental assessments, travel or software costs. A real evaluation would check those categories, taxes and its chosen accounting treatment. Participant time is a resource cost even where it creates no new invoice.
Compare benefits and costs over an explicit boundary. This simplified twelve-week evaluation includes ten weeks of measured time change and all engagement costs. It makes no claim about later persistence, discounting or a full organizational investment appraisal. Extrapolating the weekly change across several years would add an untested durability assumption.
Calculate scenarios, then decide what they permit
The net ROI formula is (attributable monetary benefit minus total cost) divided by total cost, multiplied by 100. The benefit-cost ratio is benefits divided by costs. A ratio of 1.5 means fifty-percent net ROI, not one-hundred-and-fifty-percent net ROI.
Because attribution is unresolved, finance does not put a central ROI estimate in the case report. It shows three deliberately hypothetical shares of the EUR9,000 capacity model. These are sensitivity assumptions, not probabilities, confidence intervals or measured coaching effects.
| Assumed coaching share of modeled capacity value | Modeled attributable benefit | Benefit minus EUR8,400 cost | Modeled net ROI |
|---|---|---|---|
| 0% | EUR0 | −EUR8,400 | −100% |
| 50% | EUR4,500 | −EUR3,900 | −46.4% |
| 100% | EUR9,000 | EUR600 | 7.1% |
The fifty-percent row happens to equal valuing the sixty-hour differential change at EUR75. That numerical match does not validate either the comparison design or a fifty-percent coaching share. The most favorable row attributes the entire observed reduction to coaching and still exceeds resource cost only slightly.
At this hourly valuation, recovery of EUR8,400 requires 112 attributable hours: 8,400 divided by 75. That is 93.3% of the 120-hour raw reduction. The available evidence cannot support that attribution, and the modeled capacity is not verified cash. The responsible conclusion is that a positive coaching ROI has not been demonstrated.
A negative scenario is not a statement that the executive failed. It reveals the consequences of assumptions within a limited financial model. Equally, liking the coaching does not authorize removing participant time or changing the period until the percentage becomes positive.
The completed sponsor report and decision
Purpose: Decide whether to renew Amina’s individual engagement after twelve weeks.
Observed: Eight of ten reviewed follow-up decision records contained owner, boundary and escalation condition, compared with three of ten baseline records. Recorded weekly coordination fell from 36 to 24 person-hours. Recorded missed handoffs remained at two a week.
Interpretation: Coaching plausibly supported Amina’s decision-closing practice, but the dashboard and differences between teams prevent a quantified attribution claim. The descriptive comparison does not establish a causal effect.
Financial position: Resource cost was EUR8,400. No cash saving was verified. EUR9,000 is the valuation of the entire raw time reduction before attribution. Sensitivity rows do not demonstrate an actual return.
Decision: Close the six-session engagement as contracted. Operations will monitor the same meeting and handoff measures for six further weeks within its existing responsibilities. No additional coaching sessions are commissioned. Reconsider targeted support if decision records lose clarity or handoffs worsen; do not renew solely to produce a favorable ROI number.
This decision preserves the behavior learning while resolving the commercial question. The sponsor can choose further development for other reasons later, under a new scope. The six-week operational check is not an implied extension of confidential coaching or free ongoing coach analysis.
The report also records the coach’s involvement in interpretation. A paid provider has an interest in favorable findings. Let the appropriate internal owner challenge definitions and calculations, and show disagreements rather than quietly editing them away.
Improve the next study without rewriting this result
Amina’s sponsor can learn from the weaknesses without retroactively promoting the comparison into a control group. For a future program, the evaluation owner could collect repeated observations before rollout, identify teams with relevant similarities and differences, and record when each team receives coaching and other changes. An evaluation specialist should assess whether the available design supports the intended causal question.
The details change the analysis. If coached teams volunteered after a particularly difficult month, some improvement might reflect movement away from an unusual low point. If a new dashboard reaches only those teams first, its effect travels with coaching. If managers share the coaching practice with comparison teams, “not coached” may no longer mean unexposed. Merely increasing the number of spreadsheet rows does not resolve these problems.
For Amina’s six-week operational follow-up, the analyst uses a smaller purpose: check whether the observed practice and coordination pattern persist. The same six roles and meeting categories remain in scope. The analyst checks ad hoc time and missed handoffs alongside recurring time, and records staffing changes. The review can identify a deterioration needing attention; it cannot repair the missing pre-intervention comparison history.
The agreed local review rule is to investigate if weekly coordination exceeds 30 person-hours or recorded missed handoffs exceed two in two consecutive weeks. Those are fictional management triggers chosen for this case, not validated coaching thresholds. The analyst investigates the cause before recommending a response. A new compliance obligation could reasonably require more meeting time without indicating that Amina lost the skill.
This is also where finance can examine whether released capacity was used. It might find that a canceled coordination slot was replaced with documented planning work, remained unallocated, or became additional individual administration. Report the actual use and its uncertainty. Do not assume every freed hour became productive output, or value all outputs at the same rate simply because the spreadsheet permits it.
Use a claim record before approving a headline
For each proposed report statement, create a small record with six fields: exact claim, source, calculation, evidence owner, competing explanation and permitted wording. This prevents a cautious analyst’s footnote from disappearing when a finding moves into a presentation or renewal proposal.
| Field | Completed Amina example |
|---|---|
| Proposed claim | Coaching saved EUR9,000 |
| Source and calculation | Analyst’s meeting records: (36 − 24) × 10 × EUR75 |
| Evidence owner | Operations verifies recorded hours; finance accepts the rate only as a capacity valuation |
| Competing explanation | Dashboard rollout, team differences and unknown unrecorded conversations |
| Decision | Reject the proposed savings headline |
| Approved replacement | Recorded coordination fell by 120 person-hours against the fixed baseline rate. Its modeled capacity value is EUR9,000 before attribution; cash savings and coaching’s causal share were not established. |
The replacement is longer because it carries the distinctions the short claim omitted. A presentation can shorten it to “coordination decreased; coaching ROI unproven,” then retain the calculation beside the detail. The executive, coach and sponsor should use the same meaning when discussing the result. Permission to share the finding is separate from evidence that the finding is true.
When monetization is the wrong next step
Consider a contrasting fictional executive, Luis, whose objective is to raise an uncomfortable risk at board meetings. There are only two relevant meetings during his engagement. In the first he raises the issue but provides an unclear request; in the second he supplies the evidence and a decision request. The board’s final commercial outcome is not known.
Assigning a large “avoided loss” to the second intervention would require assumptions about what would otherwise have happened, the likelihood and size of the loss, and Luis’s contribution. Those inputs are not available. Multiplying a possible loss by a guessed prevention percentage would make uncertainty look like a result.
His completed evaluation instead records the two opportunities, the change in the request, his own reflection and an authorized description of the board response. It states “financial ROI not estimated.” The sponsor closes the agreed engagement and asks Luis to continue using the preparation method. A meaningful evaluation can end with a specific nonfinancial finding.
Build an evidence trail in Coachful
Coachful is our #1 recommendation for organizing the agreed coaching work and reflection behind this evaluation. Use its goals, tasks and forms to retain the process record, then prepare the authorized sponsor report deliberately. Organization-owned operational data and financial analysis still need their appropriate owners.
The demo goal screen shows linked tasks, habits and reflection tabs. For Amina, create an agreed task such as “Use the owner, boundary and escalation close in the next eligible discussion.” Completing that task records reported completion, not verified organizational savings. Review the specific work example with the client before advancing the claim.
- Save the agreed objective. Use a goal description that identifies the behavior and review date. Keep private personal objectives separate from approved sponsor language.
- Build a short worksheet. Ask for the discussion date, actual closing words, response, what happened next and what remains unverified. Include “no eligible opportunity” so absence of opportunity is not recorded as failure.
- Choose actions after discussion. Create the selected next task or goal deliberately. A form response is not permission to invent an action or share an answer.
- Review recap settings. Session recaps send automatically by default. Where review is needed, use Sessions, the settings gear, Session AI, then Session recaps and Approve before sending. Check any session-type override too.
- Prepare the separate report. Review approved examples against operational records and the sharing agreement. A client recap is not an automatic sponsor report.
The form-builder image illustrates question authoring, not a validated ROI assessment. Check the intended client’s access and the actual response visibility before using a worksheet. Do not upload unnecessary employee-level salary or performance data merely because a prompt can accept text. Explore Coachful’s features for the supporting workflow.
Troubleshooting an evaluation already in progress
There is no reliable baseline. Search for a comparable existing source with its original definition. If none exists, describe current performance and establish a prospective baseline. Do not reconstruct a precise “before” score from someone’s enthusiasm about the “after.” Label retrospective recollections as recollections.
Only favorable respondents replied. Report invited, responded and relevant-observation counts separately. Four responses from ten invited colleagues cannot establish that the other six agree. Ask whether respondents actually saw the target behavior; use the 360 feedback review guide when interpreting stakeholder perspectives.
The measure improved because difficult work moved elsewhere. Inspect exclusions, canceled work and transferred burden. Lower meeting time is incomplete if another team absorbs extra coordination. Add the displaced resource and quality effects before interpreting the change.
A sponsor wants one persuasive number. Supply the strongest supported statement, its source and its limit. If the model changes sign under plausible assumptions, show that range. Move the decision forward through observed behavior, costs and remaining uncertainty instead of presenting an unsupported headline.
Frequently asked questions
What is a good executive coaching ROI percentage?
No percentage is a universal requirement or credible guarantee. A return depends on the measured outcome, attribution, valuation, costs and period. Define what decision the actual engagement must support, then report the evidence it produces.
Can a participant estimate coaching’s contribution?
Yes, as a clearly attributed judgment. Ask for the reasoning and competing explanations. Do not relabel the estimate as a causal measurement or treat an extra confidence discount as statistical validation.
Should saved staff time count as a financial benefit?
It can be modeled as resource value with an explicit rate and use assumption. Cash savings require evidence of actual avoided or reduced expenditure. Keep the categories separate and avoid counting the same benefit twice.
Does Coachful calculate executive coaching ROI automatically?
The workflow described here uses goals, tasks, forms and reviewed records. Those records do not establish causal attribution, verify organization-owned outcomes or replace the financial calculation. Prepare the report with the appropriate evidence owners.
What if the evaluation cannot establish a positive return?
Report that finding, including nonfinancial outcomes and limits. Decide whether to close, revise or commission further work against the original purpose. An inconclusive financial result can still support a concrete, responsible decision.
