How to Conduct a Coaching Skills Audit: Worked Guide
Nour
Founder, Coachful
When deciding how to conduct a coaching skills audit, begin with the question it needs to answer. Use specific examples, feedback and reflection to choose what to practice next. It can reveal a gap between what you intended and what another person experienced. It cannot make every judgment objective or establish your competence from one conversation.
This guide develops five review dimensions, three feedback sources, a complete fictional example, a reusable review record and a ninety-day development plan. The rubric is an original learning aid. It is not a validated assessment, an ICF scoring instrument or a substitute for a formal credential evaluation.
Agree what the review will be used for
Start with a precise question: “Do I help clients clarify the decision they want to make?” is more useful than “Am I an excellent coach?” Name the setting, the work to be sampled, who will review it, who receives the findings and the decision those findings may inform.
A private development review, mentor coaching, supervision, an employment assessment and a credential performance evaluation have different purposes. Do not present a mandatory workplace assessment as a confidential learning conversation if a manager will use the results for selection or discipline.
In a manager-as-coach setting, the employee may reasonably hear a question as an instruction. Identify the manager’s actual authority, required responsibilities and limits before evaluating a supposedly equal coaching relationship. Switching to advice or management may be appropriate, but make the role explicit.
Use the current professional framework relevant to your purpose. The 2025 ICF Core Competencies retain eight overarching competencies. The five areas below organize this article’s development review and do not replace that framework.
Keep this audit distinct from the quarterly client review, which examines the engagement and its next service decision.
Choose evidence before assigning a rating
A small sample offers a starting point, not a complete picture of practice. Choose material that fits the question and record why it was selected. Reviewing only your most polished conversation can hide routine difficulties; reviewing only a difficult exchange can create an unfair generalization.
A real recording requires an appropriate permission and information-handling process. Explain the actual recipients, use, storage, access and retention arrangements. A client’s agreement to be coached does not automatically authorize a recording, an AI transcript or sharing with a peer group.
Where a recording is unsuitable, consider a permitted observation, a carefully limited practice note or a simulation with informed participants. A reconstruction depends on memory; a simulation demonstrates behavior in that setting. State those limitations in the review.
Keep an evidence trail that is proportionate: the relevant excerpt or event, what was observed, the context, the reviewer’s interpretation and what remains unknown. Replacing names does not always prevent identification through a distinctive role or incident.
Use five development dimensions with observable anchors
For each dimension, identify what happened and its effect on the agreed purpose. An observation such as “I asked three questions before the client could respond” is easier to examine than “I lack presence.” The examples below are original prompts rather than official competency markers.
| Dimension | Look for in the chosen material | Useful development question |
|---|---|---|
| 1. Safety, agreement and boundaries | Purpose, roles, relevant confidentiality limits, choice about participation and how the conversation’s focus was agreed. | What did the client actually agree to, and where did I assume agreement? |
| 2. Listening and presence | Responses to the person’s words, checking understanding, appropriate pauses and attention to their context. | What did I notice, miss or interpret without checking? |
| 3. Inquiry | Clear questions suited to the client’s focus, room to answer and acknowledgement when an observation or advice is being offered. | Did my question invite exploration or smuggle in my preferred answer? |
| 4. Awareness and meaning | How the client examines an assumption, considers another view or decides that an observation does not fit. | What changed in the client’s own understanding, if anything? |
| 5. Action and review | Client-chosen next steps or decisions, realistic support, obstacles and an agreed way to revisit learning. | Who chose the action, and how will we learn from what happens? |
Safety here concerns the quality and limits of the working arrangements. A score cannot certify psychological safety, remove a power imbalance or guarantee that someone feels able to speak. Ask about the person’s actual experience rather than infer it from a calm appearance.
For inquiry, an open question is not automatically useful, and a short closed question may clarify an important fact. Do not call every concise question “Clean Language” or claim a specialist method without the relevant training and fidelity.
If ratings help, define them and keep their limits visible
You can use descriptive notes alone. If a small rating scale helps select a practice focus, agree its meanings before reviewing the sample. The following original scale is a discussion aid, without validated norms or a pass mark.
| Label | Meaning within this review |
|---|---|
| 1: Needs attention | A specific observed choice obstructed the agreed purpose or needs follow-up. Record what happened. |
| 2: Mixed | Useful and less useful choices appeared in the sample. Identify both. |
| 3: Supported in this sample | There is a clear relevant example of the agreed behavior serving the conversation. |
| 4: Adapted in this sample | The behavior was adjusted thoughtfully as the client’s context or preference changed, with a checked response. |
| Not observed | This sample did not provide enough relevant evidence. It is not a zero. |
| Not suitable to rate | The question needs another process, role or source of evidence. Explain why. |
Avoid adding the five ratings into an apparent scientific competence score. A four in one area cannot cancel a serious ethical concern elsewhere. Record the concern and use the appropriate process rather than wait for an overall average.
The ICF PCC evaluation guidance describes assessment of observed coaching using its own requirements and warns against treating its markers as a simple checklist. This article’s scale does not predict an evaluation result.
Bring together three feedback sources
| Source | What it can contribute | What it cannot establish alone |
|---|---|---|
| Self-review | Your intention, assumptions, attention and account of the choices you made. | The client’s unspoken experience or an unbiased verdict on performance. |
| Permitted observation | A reviewer’s account of specific behavior in a real or simulated conversation. | A complete picture of your practice or the client’s life outside the sample. |
| Client and appropriate stakeholder feedback | What the participant found useful, unclear or difficult; sponsor observations within agreed reporting boundaries. | A sponsor’s right to private session material or proof that coaching caused a business outcome. |
Ask a client a small number of relevant questions: “What helped you think?”, “When did my approach become less useful?” and “What would you like me to adjust?” Participation and the route for feedback need to be appropriate to the relationship.
If a sponsor is involved, ask only for information within the agreed role and reporting arrangement. The person paying for coaching does not automatically become an observer of confidential conversations. A sponsor’s satisfaction is also different from the client’s experience.
Anonymous feedback is difficult to guarantee in a tiny group. Explain what the system and process actually protect, and who can see responses. Do not advertise a confidential 360-degree assessment merely because you created a form.
A complete fictional example: an answer hidden in a question
In this fictional practice case, coach Dara rehearses with Nell, who plays an independent design-studio owner considering whether to stop a small but demanding service. Reviewers Ivo and Zoe examine the same short written practice excerpt and make notes separately. Everyone agrees the private learning purpose and four-person audience. This is an original simulation, not a real client transcript.
Client: I need to decide whether to keep the rush-design service. It interrupts other work, but two clients depend on it.
Coach: Would it not be better to delegate it, increase the price and tell those clients you need stronger boundaries?
Client: Perhaps, although I do not know whether I want to sell it at all.
Coach: I put several suggestions into that question. What decision would you like to work on today?
Client: Whether the service still belongs in my business, before deciding what to change.
Coach: What would you want to understand by the end of this conversation to make that decision more clearly?
Client: What it contributes, what it costs in time and whether there is a different way to serve those clients.
The first reviewer rates inquiry as “needs attention” because the opening question supplies three solutions. The second rates it “mixed” because the coach notices the problem and returns to the client’s decision. Both have identified real evidence; they initially weighted the repair differently.
They reread the same permitted material, agree that the review covers the whole excerpt and settle on “mixed,” while retaining both observations. They do not average one and two into a more precise-looking score of 1.5. If disagreement remains, they can record it instead of forcing consensus.
The useful next step is specific: practice asking one question about the client’s decision before offering a preferred solution. The excerpt does not establish whether the service should close, whether the whole session was effective or whether the coach has permanently changed.
Complete an evidence-to-practice record
| Field | Filled example |
|---|---|
| Review purpose | Examine whether my questions support the client’s own decision. |
| Sample and limitation | One permitted short practice excerpt; it cannot represent all my work. |
| Observation | I offered delegation, pricing and boundaries in a single question; I then acknowledged this and clarified the client’s focus. |
| Client response | The client distinguished deciding whether to keep the service from deciding how to change it. |
| Reviewer agreement and difference | Both saw the leading question; they initially weighted the repair differently. The whole excerpt was agreed as mixed. |
| Alternative interpretation | The coach may have mistaken the client’s request for an invitation to consult. The opening agreement needs checking. |
| Practice choice | Ask one question about the decision, then listen and check before choosing the next contribution. |
| Support | A mentor or suitable peer reviews a new permitted example, using the same narrow purpose. |
| Review evidence | A later excerpt, the client’s feedback where appropriate and the coach’s account of the choice. |
| Information decision | Keep only authorized relevant material with an agreed audience and retention period. |
Use the same fields for another dimension. For example, a review of accountability might find that the coach assigned three tasks without asking what the client wanted to try. The development focus would be client choice and feasibility, rather than adding more reminders.
Follow one inquiry skill through a complete practice review
Dara begins the development plan on 6 January 2027. The opening excerpt above is the baseline: both reviewers ultimately call the whole excerpt mixed, retaining the initial leading question and its repair. Dara selects one purpose for further practice: invite the partner's decision before introducing an answer. There is no overall five-dimension pass score.
Ask for support that the reviewers can actually provide
Nell agrees to two twenty-minute rehearsals, on 4 February and 5 March. Ivo accepts a written review of one short excerpt from each. Zoe declines weekly observation but agrees to the baseline review and a final discussion on 5 April. Dara reads those responses and confirms the smaller plan. None is enrolled as an unlimited evaluator or asked to supply real client material.
They use original staged scenarios and participant-checked written excerpts. There is no session recording, AI transcript or private customer information. The purpose, four readers and agreed handling remain visible when a new excerpt is selected. This arrangement supports this practice case; it does not automatically authorize reviewing any future client session.
Keep the first rehearsal's remaining problem
At the actual 4 February rehearsal, Nell again says the owner has not decided whether the rush service belongs in the business. Dara asks one short question: “How would you delegate it without interrupting your other work?” She has reduced the number of questions but still assumed the service will continue.
Nell says that the question skips the decision she wants to make. Dara checks that understanding instead of counting the single question as success. They include the exchange in the agreed written excerpt. Ivo reads it and returns this complete comment: “Your question is shorter, but delegation is still the answer inside it. Nell had not decided to keep the service. Practice establishing that decision first; one question alone is not the criterion.”
Dara reads the comment, replies that she recognizes the assumption and sends the revised practice focus to Nell. Nell reads and agrees that it reflects the difficulty she experienced. The feedback has actually returned to its recipient; an internal reviewed flag would not establish that.
Notice a useful change without erasing another weakness
During the actual 5 March rehearsal, Dara asks, “What do you want to decide before considering delivery options?” Nell chooses whether the service remains part of the business. Dara then asks what information would help that decision, listens to the answer and checks her understanding. Nell confirms that this opening gave her room to choose the focus.
The closing is less useful. Dara moves directly into asking for next week's action before checking whether Nell wants an action from this rehearsal. Nell declines. Dara stops the action-planning question and records the objection rather than describing it as low commitment.
Ivo reads the selected excerpt and returns two specific findings. The opening supports the agreed inquiry focus in this sample; the closing assumes an action is wanted. Dara reads the feedback and keeps both observations. A supported opening does not cancel the later assumption, and neither finding determines a credential result.
Close the ninety-day review with a bounded conclusion
On 5 April, Dara, Ivo and Zoe hold the agreed final discussion. They compare the original excerpt and the two later selected practice accounts. The period from 6 January to 5 April contains ninety days when both dates count. The 4 February check was day thirty; 5 March was day fifty-nine, chosen for the around-day-sixty review rather than relabeled as an exact sixtieth day.
Dara's first draft says, “My coaching is now consistently client-led.” Zoe challenges the scope: these are selected simulations, not a sample of all client work. Dara replaces it with this full final account: “In the second rehearsal, the partner chose the decision before we considered delivery options. The first rehearsal still contained my preferred answer, and the later closing assumed an action the partner did not want. My next focus is checking how the person wants to end. These practice examples do not establish consistent performance across client sessions or formal competence.”
Dara sends the corrected development record to the agreed four-person audience. Ivo, Zoe and Nell read it and confirm on 6 April that it reflects their permitted contributions. The review administration is complete; none agrees to further observation. A later real-client sample would require its own appropriate arrangements. The forty rehearsal minutes exclude baseline preparation, reviewer reading, the final discussion and administration, so they are not a whole-work cost or efficiency result.
Interpret talk time with the denominator stated
Talk time can highlight a question worth exploring. It is not a universal quality standard. A contract discussion, an accessible communication arrangement, a client’s preference or an explicitly agreed advisory segment can change the pattern.
In an original thirty-minute example, the coach speaks for nine minutes, the client for eighteen and three minutes are silence. Coach speech is 9 ÷ 30 = 30% of the whole period, or 9 ÷ 27 = approximately 33.3% of spoken time. Those are different denominators.
Review the actual content: Was the coach clarifying, interrupting, offering a long anecdote or responding to a request? Did the client have the time and communication support they needed? A target such as seventy to eighty percent client talk is not established here as an ICF rule or a validated optimum.
Automated speaker labels and transcripts can misattribute voices, omit pauses or mishandle overlap, accents and names. Check the relevant material against the source where authorized. A number reported by a tool does not prove attentive listening or reveal anyone’s nervous-system state.
Practice difficult moments without manufacturing a crisis
Simulation can let you examine a boundary, a disagreement or a sudden change of topic without sharing a real client’s material. Agree roles, a realistic scenario, accessibility, an option to stop and a debrief before starting.
For the design-studio example, a practice partner could say, “I hired you to tell me what to do.” The coach practices clarifying whether the person wants coaching, a competent separate advisory role or another service. The observer records the actual response and the partner’s feedback.
Avoid surprise humiliation, compulsory disclosure or acting exercises that imitate a clinical emergency without appropriate competence. Remaining physically still does not prove emotional regulation, and visible discomfort does not establish poor coaching. Ask what happened and what support would help.
Turn findings into a ninety-day practice plan
Choose one or two priorities that you can practice and review. The calendar below is an original example, with a sixty-day check inside a ninety-day plan. Adjust it for the work available, consent, access needs and the seriousness of any concern.
| When | Activity | What to keep |
|---|---|---|
| Days 1–7 | Agree purpose, permissions and the baseline sample; choose the practice focus. | Evidence, limitations and one or two specific development questions. |
| Weeks 2–4 | Rehearse one question at a time in an agreed practice setting; apply thoughtfully in suitable work. | Brief notes on the choice, response and what needs changing. |
| Around day 30 | Discuss a permitted example with a suitable mentor, supervisor or peer according to the actual need. | Feedback and the revised practice choice. |
| Around day 60 | Review another suitable sample using the same narrow anchors. | What is comparable, what differs and what remains uncertain. |
| Around day 90 | Consider evidence across the period, participant feedback where appropriate and next support needs. | Continue, adapt, seek different training or choose the next focus. |
Improvement may be uneven and context-specific. A later example can show a more useful choice without proving a permanent habit or world-class competence. Do not force a preferred technique into every conversation just to produce more favorable evidence.
Use supervision for wider questions about practice and relationships, and appropriate mentor coaching for skill development within the agreed framework. A peer circle can offer feedback, but attendance alone does not meet every formal supervision or credential requirement. See the supervision models guide for that distinction.
Use organizational findings with appropriate boundaries
For a leadership program, first clarify whether the review supports development, accreditation, selection or performance management. Participants should understand the actual use, audience and consequences. A manager cannot promise independence if they are also deciding promotion.
Aggregate only where the process and group size support it. A short anonymous quotation or a distinctive scenario can identify someone. Consider whether a broad theme, such as difficulty agreeing a session purpose, can inform training without circulating individual client material.
Evaluate several layers separately: participation, the quality of the review process, observed choices in suitable samples, participant experience and any wider outcomes. Retention, engagement or project delivery also depend on staffing, leadership, resources and other changes.
An improvement in a business measure after training is not proof that the skills audit caused it. Avoid promising a financial return or describing a coach as a fiduciary merely because they followed a review rubric.
Check six digital capabilities against the actual workflow
| Capability to consider | What to verify |
|---|---|
| 1. Original review forms | Suitable fields, record ownership, completion route and who can review responses. |
| 2. Authorized recording or transcript review | Actual consent, tool availability, accuracy, recipients and retention. It may require a separate service. |
| 3. Appropriate feedback collection | Audience, access, notifications and the real limits of confidentiality or anonymity. |
| 4. Development records | A proportionate history of practice choices and review dates, without unsupported competence scores. |
| 5. Learning and follow-up | Resources and agreed ordinary actions matched to the person’s actual development plan. |
| 6. Information protection | Actual permissions, staff and technical access, exports, deletion and retention arrangements. |
We build Coachful. Coachful is our #1 recommendation for organizing permitted development reflections, resources and agreed follow-up in an appropriate coaching workspace. This is our editorial workflow recommendation, not a measured market ranking or a validated skills assessment. Do not assume native anonymous 360-degree scoring, automatic question-quality judgments or accreditation evidence.
Keep a permitted development record within its agreed audience. Where the participant is an authorized client of the mentor, deliberately assigned resources and a private client conversation can support the agreed process. A copied resource URL does not grant access; a coach-team form review does not deliver feedback to the participant. Return any agreed comment through the actual client-facing route. Explore Coachful’s supported features against the intended roles. The screens are demos, not Dara’s records or a customer UI test.
Frequently asked questions
Does this audit certify my coaching competence?
No. It is an original development process using limited evidence. Formal qualifications and credential evaluations have their own current requirements.
Must I record every coaching session?
No. Recording requires an appropriate process and permission. A permitted observation, limited note or simulation may be more suitable, with its limitations stated.
What if two reviewers disagree?
Return to the same evidence and agreed purpose. Record what each person noticed, clarify the sample and retain unresolved differences rather than manufacture precision.
Is a low coach talk-time percentage always better?
No. State the denominator and examine what happened in context. A ratio alone does not show whether the conversation served the client.
Can the sponsor see the client’s review material?
Only within the actual agreement, permissions and relevant obligations. Payment does not automatically authorize access to private session information.
How often should I repeat a review?
Use a cadence suited to your purpose and work. The sixty-day check and ninety-day review here are examples, not universal professional requirements.
Begin with one development question, gather a suitable sample and finish with a specific choice to practice. A useful audit makes the next conversation more thoughtful while remaining honest about what the evidence can establish.
Primary framework context: the 2025 ICF Core Competencies and current ICF PCC performance evaluation guidance. The five-dimension rubric, rating labels, fictional dialogue, evidence record, arithmetic and practice calendar are original learning aids. Product descriptions are limited to current Coachful source.
