Coaching Outcome Attribution Reviews: Evidence, Models and Worksheets
Nour
Founder, Coachful
A coaching outcome attribution review asks what changed, what evidence supports that account and how coaching may have contributed alongside other influences. Start by separating completed activities, observed outcomes, a reasoned contribution claim and an estimated causal effect.
The framework below includes four evidence pillars, four model categories, a ninety-minute review, six original stakeholder questions and five filled financial worksheets. It is a structured coaching evaluation aid, not an independent financial audit or a method for automatically proving that coaching caused a business result.
Choose the question before building the report
| Question | Suitable evidence | Limit |
|---|---|---|
| Was the agreed work delivered? | Relevant versions, dates, attendance or completion records. | Delivery is not the same as effectiveness. |
| What outcomes were observed? | Defined operational records, observations and client accounts. | A change during coaching is not automatically caused by coaching. |
| How might coaching have contributed? | A plausible sequence, evidence of its steps and examination of other explanations. | A reasoned contribution case is not a numerical allocation of credit by default. |
| What was the causal effect? | An appropriate evaluation design, comparison and analysis led by competent evaluators. | Before-and-after results or stakeholder agreement alone do not isolate the effect. |
The CDC evaluation framework distinguishes process, outcome, impact and economic evaluation. Its public-program context provides useful methodological distinctions rather than mandatory rules for every coaching engagement.
Choose a proportionate question and audience. Some reviews should remain about the work, the client’s experience and next decisions. A financial return calculation is unnecessary when a credible monetary estimate is unavailable.
Build four evidence pillars
| Pillar | Working record |
|---|---|
| 1. A defined starting point | Purpose, relevant outcome, measurement definition, source, period, context and missing data. |
| 2. Documented activity and practice | What the coach and client actually did, with appropriate source and timing. |
| 3. Milestones and outcome records | Produced, submitted, accepted and used statuses, plus the separate result being examined. |
| 4. Multiple perspectives and alternatives | Client, relevant observers and authorized business owners, with differing accounts and other influences retained. |
A baseline need not be a comprehensive psychological diagnostic. It may be a defined business record or an original reflection. Licensed assessments require their authorized administration and appropriate competence.
Habit logs show recorded practice, not every action taken outside the platform. Ninety-percent logging alongside lower staff turnover does not establish a causal link. A milestone can be complete while its intended outcome remains unknown.
HM Treasury’s Magenta Book describes contribution analysis as examining a plausible theory of change, evidence and other influences. It supports a reasoned contribution case while explicitly limiting claims of definitive causal proof. The source concerns government evaluation, not a validated coaching instrument.
Use a filled contribution record
In this original fictional case, Jules coordinates a cross-functional product decision process. Coaching includes practicing clearer questions about who owns unresolved decisions. During the same period, the organization appoints a new decision owner and changes its approval process.
| Field | Filled Jules record |
|---|---|
| Question | What happened to the decision process, and what role might the chosen communication practice have played? |
| Starting record | Twelve sampled decisions have a reported median recorded duration of twelve days. Definition and selection need review. |
| Coaching activity | Jules describes preparing an ownership question before relevant meetings and records examples of using it. |
| Later record | Ten sampled decisions have a reported median recorded duration of eight days. |
| Other influences | A new decision owner, a revised approval process, staffing and the mix of decisions may matter. |
| Contribution evidence | Jules and two observers describe clearer requests in specific meetings; those accounts concern communication. |
| Current conclusion | The records and accounts warrant examination, without allocating a percentage of outcome credit to coaching. |
| Owner and next decision | An authorized analyst reviews whether durations, selection and comparison are consistent before reporting a change. |
The difference between twelve and eight is four days. That numerical difference alone does not establish a comparable improvement or the time that would have been taken without coaching. The new owner’s authority and revised process are substantive alternative explanations.
In a matrix organization, technical teams, project owners, sponsors and the client can contribute different work. Do not turn a coach’s record of one conversation into proof that a launch would otherwise have failed.
Let a later finding change the conclusion
On reviewing the samples, the analyst finds that the later record excludes decisions still waiting for approval, while the earlier record includes them. The two reported medians were built from different definitions.
The team revises the briefing: “The available duration summaries are not yet comparable. We have examples of clearer ownership questions and a changed approval process, but no established estimate of cycle-time improvement or coaching’s causal effect.”
The next task is a consistent analysis with an explicit cutoff and treatment of pending decisions. Jules can still discuss whether the practice was useful. The reviewer does not hide the discrepancy or replace it with an arbitrary twenty-five-percent contribution assumption.
Stakeholder estimates can be recorded as opinions with their reasoning and limitations. Agreement among several people is not independent causal identification, and averaging their percentages does not remove shared incentives or missing evidence.
Consider four model categories carefully
| Category | Appropriate question | Common overclaim to avoid |
|---|---|---|
| 1. Cost avoidance | What expenditure was credibly avoided over a defined period, compared with what would otherwise occur? | Assuming a retained employee was about to leave or applying a universal recruiter percentage. |
| 2. Timing or revenue acceleration | Was an actual cash flow received earlier, and what accepted timing-value method fits? | Treating pipeline, earlier revenue and additional profit as interchangeable. |
| 3. Operational capacity | What time, effort or resource changed, for whom and over what period? | Calling salaried time released cash saved without an avoided expense. |
| 4. Milestone timing | How does an actual deliverable date compare with an appropriate reference, considering scope and dependencies? | Using a standard multiplier to turn completed milestones into enterprise value. |
A category helps define a question; it is not a standardized proof of advisory impact. Record the actual comparison, costs, uncertainty and qualified owner. Do not claim legal fees, settlements or redundancies were prevented merely because coaching took place.
Use five financial worksheets
The following are separate original instructional scenarios. They are not Jules’s business results and should not be added together. Use consistent units and periods, include relevant costs and distinguish initiative-level value from coaching’s contribution.
1. Talent retention
Fields: role or group, actual employment outcome, credible comparison, replacement-cost components, other influences, source, uncertainty and finance owner.
Filled example: a leader remains employed while pay, reporting lines and job scope change. No reliable evidence establishes that the leader would otherwise have left. Record continued employment and context; monetary benefit is not established. A salary multiple or a stakeholder’s guess does not supply the missing comparison.
2. Timing value
Fields: actual cash amount, timing difference, credible comparison date, accepted annual rate, day-count basis, costs and responsible owner.
Filled example: a hundred thousand dollars received thirty days earlier at an assumed ten-percent simple annual rate has an illustrative timing value of 100,000 × 0.10 × 30 ÷ 365, approximately 821.92. Use only if the cash timing and valuation assumptions are appropriate. The hundred-thousand-dollar principal is not additional benefit, and pipeline is not cash received.
3. Capital preservation
Fields: vendor contract, comparison expenditure, cancellation status, measured period, replacement needs, exit costs and other influences.
Filled example: an authorized initiative ends an unnecessary five-hundred-dollar monthly vendor service for a four-month comparison period. Gross avoided expenditure is two thousand dollars if that comparison is supported. Record whether the cancellation was already planned and whether another team drove it.
4. Capacity
Fields: time difference, people affected, observed weeks, record source, use of released time, relevant cost basis and cash-versus-capacity classification.
Filled example: two hours per week for three people over six weeks releases thirty-six person-hours. At an assumed fifty dollars per hour, the estimated capacity value is eighteen hundred dollars. It is not cash saved if payroll is unchanged and no avoided expense is established.
5. Cost and decision summary
Fields: initiative, period, accepted eligible benefits, full costs, excluded outcomes, attribution method, sensitivity and owner.
Filled example using only the vendor scenario: stated costs are twelve hundred dollars of relevant advisory work, three hundred dollars of staff time, two hundred dollars of implementation and a three-hundred-dollar exit charge. Total costs are two thousand dollars, with the exit charge included once.
If the four-month comparison is accepted, two thousand dollars of benefit divided by two thousand dollars of costs gives a gross benefit-cost ratio of 1. Net ROI is (2,000 − 2,000) ÷ 2,000 × 100, or zero percent. If only two months are supported, benefits are one thousand dollars and net ROI is minus fifty percent.
If the comparison is unsupported, report that ROI is not established. Even an accepted initiative-level calculation does not allocate all value to coaching. ROI Institute’s Performance Consulting distinguishes gross benefit-cost ratio, net ROI and isolation of effects; its broader method is not replaced by these arithmetic examples.
Ask six neutral stakeholder questions
Explain the purpose, audience, response use and confidentiality limits before seeking feedback. Invite relevant observations, an option to say there was insufficient opportunity to observe, and optional examples. These original prompts cover five areas plus context.
| Area | Original question |
|---|---|
| 1. Decision transparency | What have you observed about how the reasons for relevant decisions are explained? |
| 2. Collaboration | What has helped or hindered coordination across the teams involved? |
| 3. Discussion under pressure | What have you observed about how disagreement or uncertainty is handled in demanding situations? |
| 4. Delegation | How clear or unclear have responsibilities and decision authority been in situations you observed? |
| 5. Development and support | What have you observed about how the leader supports others’ learning or chosen responsibilities? |
| 6. Context and alternatives | What other changes, limited observation or missing information should affect how your account is interpreted? |
If an original one-to-five scale is used, define the actual question, anchors, period and nonnumerical options. A change in an ordinal self-rating is not a percentage increase in leadership ability. These questions are not copied assessment items or a validated diagnostic.
Reviews at elapsed days one, ninety and three hundred sixty are planning examples. Three hundred sixty days is not an exact calendar year. Report respondent counts, opportunity to observe, missing answers and changes in who responded before comparing summaries.
Run a six-step ninety-minute review
| Step | Time | Output |
|---|---|---|
| 1. Confirm purpose and baseline | 10 minutes | Question, audience, definitions and actual starting record. |
| 2. Review milestones | 15 minutes | Versions, dates, produced/submitted/accepted/used status and dependencies. |
| 3. Examine practice and context | 15 minutes | Relevant actions, client account, missingness and other changes. |
| 4. Review stakeholder evidence | 15 minutes | Observations, disagreements, opportunities to observe and alternative explanations. |
| 5. Consider financial questions | 20 minutes | Appropriate models, full costs, uncertainty and whether an estimate is justified. |
| 6. Decide and communicate | 15 minutes | Supported conclusion, open questions, actions and any separate next-service decision. |
Ten plus fifteen plus fifteen plus fifteen plus twenty plus fifteen is ninety minutes. This assumes preparation is available. Narrow the scope or defer a conclusion when evidence, permission or specialist review is missing.
A deliverable review asks whether the work met its agreed criterion. An attribution review examines the explanation of outcomes. They can inform each other without a fixed quarterly-versus-annual rule or a guaranteed preparation time.
Answer three difficult questions directly
| Question | Evidence-based response |
|---|---|
| Could favorable market conditions explain this? | “They may contribute. Here are the relevant changes, comparison limits and evidence for the proposed mechanism. We have not isolated coaching’s share numerically.” |
| Would the executive have done this anyway? | “That is a counterfactual question. The available records do not establish an alternative twelve- or eighteen-month learning timeline.” |
| Can a qualitative leadership change be quantified? | “Some records or ratings can be summarized, with definitions and limitations. A proxy measure does not automatically establish the underlying capability or a causal financial return.” |
A credible answer can be that the available evidence is insufficient. Do not use confidence, a conservative-sounding percentage or a polished graph to imply a method that was not undertaken.
Prepare a six-part briefing pack
| Part | Filled Jules example |
|---|---|
| 1. Purpose | Understand the decision-process experience and what can be concluded. |
| 2. Work and milestones | Relevant communication practice and organizational approval process are described separately. |
| 3. Evidence | Client and observer examples plus the two source summaries, with definitions visible. |
| 4. Findings and limitations | Duration summaries are not comparable; the new decision owner and process are relevant influences. |
| 5. Financial position | No monetary attribution estimate is established for this case. |
| 6. Decision and next steps | Authorized analyst to review consistent records; client chooses useful next practice or ending. |
Keep private reflections, identifiable raw feedback and unauthorized business records out of the pack. Confirm the permitted audience and version. A sponsor’s role does not automatically create permission to read everything.
For a fifteen-minute executive briefing, use three minutes for purpose and supported conclusion, five for evidence and context, four for limitations and open questions, and three for decisions and owners. A next engagement is a separate commercial choice rather than the required final slide.
Complete six preparation checks
- Purpose and audience: the decision and permitted recipients are clear.
- Starting record: definitions, sources, periods and context are documented.
- Activity and milestones: relevant versions, dates, status and missing records are available.
- Perspectives and alternatives: observations, differences and other influences remain visible.
- Financial review: any estimate has an appropriate comparison, consistent period, full costs and qualified owner.
- Briefing and action: supported conclusions, limitations, approved sharing and next owners are explicit.
Checking these items does not certify a financial audit or establish that a causal claim is correct. It makes the basis of the working review explainable.
Use four implementation milestones when appropriate
| Illustrative annual planning point | Working decision |
|---|---|
| Month 1 | Agree questions, measurement definitions, audience and an appropriate starting record. |
| Month 3 | Review delivered work, early experience and whether the proposed explanation remains plausible. |
| Month 9 | Check data quality, comparison needs, relevant stakeholders and specialist preparation. |
| Month 12 | Review supported findings and next decisions, including a suitable finish. |
Use actual dates for a calendar-month service. For cohorts, define who is included, the comparison, missing observations and differences in exposure. A group average cannot show that every individual improved or protect identifiability in a small group.
Coachful can organize suitable worksheets, submitted responses, relevant goals and follow-through. Client private journals and coach-team private review notes are separate from shared recaps and permitted sponsor material. Portal records do not constitute an immutable independent audit trail or automatic causal model.
The deliverable audit guide develops completion and acceptance records. The quarterly ROI guide develops financial examples. Explore Coachful’s features for appropriate records and delivery options.
Frequently asked questions
Can habit completion prove coaching caused an outcome?
No. It documents recorded activity. A causal claim requires an appropriate design and evidence beyond simultaneous changes.
Do stakeholder contribution percentages isolate the effect?
Not by themselves. They are opinions that need reasoning and limitations; averaging them does not create independent causal identification.
What if business results decline during coaching?
Examine the defined outcome, context, client experience and other influences. Decline alone neither proves failure nor establishes a hidden coaching benefit.
Should every review calculate ROI?
No. Report monetary value as unestablished when the comparison or valuation is not credible.
Is three hundred sixty days an annual calendar review?
It is a defined elapsed-day point. Use the actual calendar year or agreed dates when that is the intended period.
Does the briefing require a contract extension?
No. Continuing, changing the scope and finishing are separate decisions with their own appropriate authorization and arrangements.
A convincing attribution review makes its reasoning visible. State what is known, what remains uncertain and what evidence could change the conclusion. That gives clients and sponsors a sounder basis for decisions than a guaranteed return claim.
Primary context: CDC Program Evaluation Framework (2024); HM Treasury Magenta Book Annex A, contribution analysis (current 2026 edition); ROI Institute, Performance Consulting; and Coachful records and sharing guidance. Worksheets, cases, prompts, agendas and financial scenarios are original examples, not validated assessments, independent audits or observed customer outcomes.
