Measuring coaching progress without reducing people to scores
Coaching progress is often easier to describe than to capture. Useful measurement blends reflection, behavioural evidence and agreed outcomes, while avoiding crude rankings. The aim is not perfect precision, but a fairer picture of change that respects context, trust and the person behind the metrics.
Coaching works best when people are seen in full, not flattened into a dashboard. Yet organisations still need some way to understand whether coaching is helping. The tension is familiar: leaders want evidence, coaches want nuance, and coachees want privacy and agency. The answer is not to abandon measurement, but to use it more carefully. Research on leadership, AI and organisational change suggests that performance can be observed in different ways, but also that measurement choices shape behaviour 23. A sensible coaching approach therefore treats data as one input among several, not as the whole story. The practical task is to track progress in ways that support learning, rather than turning development into a scorecard.
Why coaching progress resists simple scores
Coaching deals with people in motion. Goals shift, contexts change, and progress may show up first in judgement, confidence, or relationship quality rather than in neat output measures. That makes coaching different from many operational processes. A single number can miss important movement, especially when the work involves leadership presence, emotional regulation, stakeholder trust or decision-making under pressure. Harvard Business School’s discussion of leadership assessment shows that even in controlled tasks, outcomes are shaped by the interaction between the individual and the situation, which is a useful reminder for coaching measurement 2.
A score can also create the wrong incentives. If a coachee starts performing for the measure, the measure becomes part of the problem. Overly rigid tracking may narrow curiosity, discourage honesty and push people towards safe, visible actions rather than meaningful change. In practice, progress is better understood as a pattern: what is different, who notices, and in what situations the difference holds.
Use mixed evidence rather than one verdict
A robust coaching review can draw on a small set of evidence types. First, agree a few development themes at the outset: for example, clearer delegation, more decisive prioritisation, or stronger boundary-setting. Then collect evidence that sits close to those themes. That might include self-reflection, stakeholder observations, work samples, meeting behaviours, or short pulse conversations. MIT Sloan’s writing on generative AI and management emphasises that technology tools are useful when they are governed well and used with clear purpose, rather than treated as substitutes for judgement 3. The same logic applies here: measurement should be designed, not improvised.
Triangulation matters because each source has blind spots. Self-report captures insight, but not always impact. Stakeholder feedback shows effect, but may reflect politics or a single moment. Behavioural observation can be illuminating, but only if it focuses on specific situations. The practical interpretation is straightforward: ask whether several sources point in the same direction, and whether they tell a coherent story about change over time. That story is more valuable than any isolated metric.
Make progress visible through behaviour
The most useful coaching indicators are often behavioural and observable. Rather than asking whether someone has become a better leader in the abstract, ask what they do differently. Do they pause before reacting? Invite dissent? Set clearer expectations? Close conversations with agreed next steps? These kinds of markers are concrete enough to discuss, yet broad enough to capture real development. Harvard’s discussion of leaders working with AI agents notes that stronger leaders asked more questions, created more turn-taking and used more inclusive language 2. While that study is not coaching research, it supports a wider point: the quality of interaction is often more revealing than self-promotion or volume.
For coaching, that means building a behaviour-based progress log. Keep it light. A coachee might note one or two situations between sessions where they tried a new approach, what happened, and what they learned. A sponsor or line manager can contribute observations only where confidentiality and consent allow it. The goal is not surveillance, but reflection grounded in real work. Importantly, the language should remain descriptive rather than judgemental.
Protect context, confidentiality and trust
Measurement only helps if people trust the process. Coaching conversations often involve sensitive ambitions, tensions and blind spots, so data collection must be narrow, transparent and agreed in advance. Explain what will be tracked, who will see it, and how it will be used. Avoid turning coaching notes into performance management by default. If the coachee believes every reflection may be weaponised, they will edit themselves and the coaching will lose honesty. MIT Sloan’s guidance on AI adoption highlights the importance of clear usage rules, safe experimentation and responsible evaluation; coaching deserves the same discipline of boundaries and purpose 3.
A useful safeguard is to separate developmental evidence from formal assessment wherever possible. Use coaching data to inform the coaching relationship, not to create hidden rankings. If organisational leaders want aggregate insight, it should be high-level and anonymised, focused on themes rather than individuals. That preserves dignity and reduces the temptation to reduce complexity to a league table. Authenticating progress means asking not only what improved, but also what changed in a way that matters to the person and their context.
A practical way to review coaching progress
A simple review structure can keep the process humane and useful. Start by revisiting the original development intention in plain language. Next, gather three forms of evidence: the coachee’s reflection, one or two behavioural examples, and any relevant stakeholder observations. Then discuss what has changed, what remains difficult, and what the person now notices that they did not before. Finish with one or two next experiments rather than a final verdict. This keeps coaching future-focused and reduces the pressure to prove a permanent transformation.
At Authenticate, an understated view is that coaching measurement works best when it resembles good clinical note-taking in spirit, even though it is not therapy: careful, contextual and faithful to the person’s lived experience. The aim is to make progress visible without turning development into a score. Done well, that helps organisations learn while still leaving room for judgement, privacy and human complexity.
Sources
- The 2026 AI Index ReportStanford University — Human-Centered AI · University research · link checked 15 October 2026
- Think You’re a Good Leader? Try Leading an AI TeamHarvard Business School AI Institute · University research · link checked 15 October 2026
- How to manage two types of generative artificial intelligenceMIT Sloan School of Management · University research · link checked 15 October 2026
- Who should be the head of generative AI — and what they should doMIT Sloan School of Management · University research · link checked 15 October 2026
Further professional reading
External professional resources, separate from the numbered sources above. These are not university studies, and listing them does not indicate any collaboration, certification, affiliation or endorsement.
- Global Institute of Directors (opens in a new window)External professional resource · governance and directorship
- World Coaching Institute (opens in a new window)External professional resource · coaching practice