InsightsInsights · Development, coaching & careers

How to know whether coaching worked: ask the people who would notice.

When managers learn to coach, the evidence is not on the feedback forms. It is in what their team and their own manager see them doing differently, two months later.

By Ron EldridgeNovember 2026 · 7 min read · Source: great{with}talent Coaching for Performance programmes · Fact-checked October 2026

Decide firstName the behaviours that should change before anyone attends, in words a team member could judge.
Ask the observersAsk each participant’s manager and the people they manage, not only the participants themselves.
Time itCheck behaviour at about eight weeks, and look for results some months later.
On this page
  1. Four levels, and the one that matters
  2. Step 1: name the behaviours before the programme starts
  3. Step 2: take a starting point
  4. Step 3: ask the observers at about eight weeks
  5. Step 4: ask the participants four questions
  6. Step 5: look for results later
  7. Mistakes to avoid
In short

Judge a coaching skills programme on behaviour, not reaction: name the behaviours first, take a starting point, ask participants' managers and team members about change at about eight weeks, ask participants four reflection questions, and look for results some months later.

Organisations can spend a good deal on teaching managers to coach, and then judge the investment on how the workshop felt. The day went well, the trainer was good, people said they would use it. None of that tells you whether anyone in the participants’ teams is being managed differently, which is the only reason to run the programme.

This guide is about evaluating a programme that teaches managers coaching skills. If you are judging coaching bought for one person, from an external coach, see Buying coaching: seven questions to answer before you start, which covers agreeing goals and reviewing them with the sponsor.

Four levels, and the one that matters

Donald Kirkpatrick’s four levels, first set out in 1959, are still the clearest way to think about it: reaction (did they like it?), learning (do they know more?), behaviour (do they do anything differently at work?) and results (did that make a difference to the organisation?). Reaction is the easiest level to collect, and it is tempting to stop there. For coaching, behaviour is the level that matters, and it is also the one you can measure without much difficulty, because coaching happens in front of other people.

Step 1: name the behaviours before the programme starts

Write down eight to ten things a coaching manager does that someone working with them could see. Keep each one to a single behaviour, in plain words. An illustrative list:

  • Gives praise without adding a “but”
  • Gives clear feedback about how to develop
  • Asks the team for feedback on their own management
  • Asks open questions to explore a problem before offering a solution
  • Listens until they understand what someone needs
  • Explores a problem fully before jumping to a fix
  • Asks how people in the team are really feeling, not just “How are you?”
  • Talks with people about what motivates them
  • Does something deliberate to build people’s capability

Use the same list in the programme itself, so participants know what they are aiming at, and in the follow-up, so you measure what you taught.

Step 2: take a starting point

We suggest asking participants, before the programme, which of the behaviours they have shown in the last two weeks, and if you can, asking their teams the same. It gives you something to compare against, and it can be the moment managers realise how much they tell.

Step 3: ask the observers at about eight weeks

When one client evaluated a Coaching for Performance programme we had delivered, it asked participants’ own managers, and the people the participants managed, about eight weeks after the workshop. Eight weeks is long enough for new habits to show and short enough for people to connect them to the programme. Ask about change, not competence. A change scale like this one asks the question you actually want answered, and it avoids asking a team to grade its manager:

ScoreSince the programme, this behaviour has shown…
1No noticeable change
2Slightly different
3Noticeably different
4Very different
5Transformed

Add two open questions: what is this person doing differently, and what difference has it made to you or the team? Show the managers’ and the team members’ views side by side. Where they agree, you have evidence. Where they disagree, you have a conversation worth having with the participant.

Keep team members’ answers anonymous, and do not show results for a manager where so few people replied that an individual could be identified. Anonymity and confidentiality in employee surveys explains how to set that up.

A happy sheet tells you the day went well. Only the people who work with a manager can tell you whether the manager changed.

Step 4: ask the participants four questions

Their own view matters too, and writing it down helps the learning stick. Ask: what have been your main lessons from the programme? How have you used the skills and the mindset since? What difference has that made, to you, to the people you manage and to the organisation? What will you do to keep using them? Some organisations make these four answers part of a simple internal accreditation.

Step 5: look for results later

Some months on, we suggest six or so, look at measures you already collect for participants’ teams: engagement survey scores on management and development, early leaving, internal moves and promotions, absence. Be careful with what you conclude. Many things move these numbers, so compare participants’ teams with similar teams where you can, and treat the results as supporting evidence rather than proof.

Mistakes to avoid

  • Asking only the participants. How managers see their own behaviour can differ from what their team sees, and participants have a stake in the programme having worked.
  • Measuring the day after. At that point you are measuring reaction again.
  • Changing the list. If the behaviours move between the programme and the follow-up, you cannot tell what changed.
  • Treating one number as the verdict. An average change score hides the participants who changed a great deal and those who did not change at all; look at the spread and talk to both groups.

Asking the people around a manager about specific, forward-looking behaviours is what a good 360 does. FeedForward is built that way, and it can be run before and after a programme.

Drawn from great{with}talent’s Coaching for Performance programmes for managers, including their reflection questions and an eight-week follow-up that one client ran with participants’ managers and team members. The four levels are Donald Kirkpatrick’s, first set out in 1959. See also Coaching for performance: the manager as coach.

Next step

The clearest thirty minutes in your week.

Send us your current 360 framework and we’ll show you how it looks on FeedForward – no charge, no deck, no obligation. Or start with a pilot or a 30-minute conversation.