Coaching Metrics That Lead to Decisions: A Dashboard Guide

Which decisions should coaching metrics support?

Every metric on a coaching dashboard should answer a question you face repeatedly. Before a session: which commitment needs discussion? At a review: has the agreed outcome changed? Across your practice: which client needs an administrative response? These questions need different views, even when they use the same records.

The Government Digital Service's 2017 performance-metric guidance starts with service purpose and hypotheses before selecting measures. Apply that sequence to coaching: write the decision in plain language, then identify the evidence needed. Do not begin with every field your software happens to collect.

For an independent practice, begin with a client view and a private operations view. The client view holds accepted actions, goal evidence and questions for review. The operations view holds upcoming sessions, unanswered requests and missing records. Revenue and renewal forecasting belong to practice management, not a client-facing judgement of engagement.

A useful dashboard tells you what needs a decision, not merely what has been counted.

Which coaching KPIs make a minimum viable set?

The following is a proposed dashboard specification, not a universal coaching benchmark or a list of automatic MentPass calculations. Select the rows that fit your agreement. If no action was due, show 'not applicable' for completion rather than congratulating the client on a fictional 100%.

A percentage becomes interpretable when its population is visible. The 2016 Government Digital Service guide to measuring completion rate explicitly separates completed transactions from all started transactions. Coaching needs its own eligibility rules, but the same numerator-and-denominator discipline applies.

Add a short metric dictionary beneath the dashboard: name, purpose, formula, exclusions, source, cutoff and owner. For action completion, state how you handle withdrawn work and revised dates. Preserve the original cohort for historical reporting, while showing the current action status alongside it. Otherwise, moving a due date can quietly rewrite last month's result.

A percentage without its eligible population is an incomplete statement.

How do the metrics work with real records?

Consider this illustrative monthly review, not a MentPass client result. A client accepted 8 actions due during the month. At the cutoff, 5 were complete, 1 was blocked by a supplier and 2 remained open. Completion by the reporting cutoff is 5/8, or 62.5%; the blocked item stays visible within that original cohort. This does not establish whether completed actions met their individual deadlines.

The useful next question is not 'How do we get to 80%?' It is whether the supplier decision is preventing a valuable launch, whether the 2 open tasks still matter and whether the completed work changed anything. A low completion percentage can coexist with an important milestone; an immaculate task list can coexist with a stalled goal.

Do not average percentages across clients without deciding what the result means. If one client completes 1/1 actions and another completes 1/9, the mean client rate is about 55.6%, while the pooled action rate is 2/10, or 20%. Both calculations are valid answers to different questions. Neither should be labelled simply 'completion' in a practice summary.

Counts explain the movement that percentages can hide.

How should marketing metrics sit beside coaching activity?

Keep coaching engagement metrics distinct from the client's marketing results. An accepted action is evidence of commitment; a qualified enquiry is a business observation. Place them in adjacent sections with dates, then write the plausible connection and the competing explanations. Do not merge them into one 'impact score'.

Google's Google Analytics 4 (GA4) engagement-rate definition, checked in September 2026, uses engaged sessions as a proportion of sessions. That is website behaviour, not engagement with coaching. Keep the full metric label and property context so a client cannot mistake one meaning for the other.

Google's Search Console report guide, also checked in September 2026, defines click-through rate as clicks divided by impressions. If you show search clicks, add impressions and the applied search filters. More clicks after a mentoring session are not evidence that the session caused them.

As of September 2026, MentPass can place read-only reports from connected marketing accounts beside client records. Use that proximity for discussion, not automatic attribution. The broader mentoring analytics definition explains why business measures need interpretation as well as collection.

Related records can support a question without proving a causal answer.

How do you keep metric reviews honest?

Use explicit states: current, incomplete, stale and unavailable. A disconnected account is unavailable, not a collapse to zero. Mark a changed event definition as a comparability break. Use equal review windows where appropriate and explain holidays, launches or other changes in operating conditions.

1. Check the population. Confirm which clients, goals and actions qualify for this period.
2. Check the evidence. Open the underlying record for an exception before interpreting it.
3. Discuss the meaning. Ask the client what changed outside the recorded plan.
4. Record one decision. Name its owner, review date and the evidence that will change your mind.

Choose attention thresholds with the client, not from an unsourced industry target. For example, a supplier dependency unresolved by an agreed launch checkpoint can trigger a review. That is an operational rule, not a prediction of churn. Revisit thresholds when the work or the client's capacity changes.

Unknown data deserves a label, not a reassuring colour.