The GROW coaching model

GROW came out of the UK coaching scene in the 1980s. Graham Alexander, Alan Fine and Sir John Whitmore all shaped it, and Whitmore's Coaching for Performance, first published in 1992, carried it into corporate training worldwide. Its parent is Timothy Gallwey's Inner Game work, which held that awareness raises performance faster than instruction does. That inheritance is the point of the model, and the first thing most managers lose.

The four stages at a glance

Goal: agree what the conversation is for

The Goal stage does two jobs and most people do one. There is the goal for the conversation, what the person wants to walk out holding, and the goal beyond it, what they are chasing over the quarter. Agree the first out loud, or the talk drifts until somebody else needs the room.

Vague goals produce vague coaching. "I want to get better at delegating" is not a goal, it is a mood. Push until you have something you could watch happening: "by the end of the month my deputy runs the client review and I am not in the room."

Coaching questions for the Goal stage

• What do you want to be different by the end of this conversation?
• If this were solved, what would you be doing differently on a Monday morning?
• How will you know you have got there? What would someone watching see?
• Is that goal yours, or one you have inherited from somebody else?
• What is the smallest version of this that would still count as progress?

If the person cannot say what they want, you are not coaching yet. Stay in Goal until there is something to aim at.

Reality: the stage everyone rushes

Reality is where the conversation earns its value, and where almost everyone hurries. The temptation is to accept the first account and move on. That first account is usually a summary the person has told themselves several times already, with the awkward parts sanded off.

Stay on facts before interpretation: what happened, before why they think it happened. Ask for numbers, dates, sentences that were actually said. Ask what they have already tried, because half the options you were about to offer have failed already, for reasons you do not know.

Coaching questions for the Reality stage

• What is happening now, specifically? Who is involved, and how often?
• What have you already tried, and what happened when you did?
• What is your part in this? Where are you contributing to it?
• What is stopping you? Is that obstacle real, or is it a prediction?
• If I asked your team the same question, what would they say?

That last question does more work than any other in the model. It is the shortest route to the gap between how somebody thinks they land and how they actually land. Active listening matters more here than clever questions: people tell you where the problem is inside two minutes, if you let them finish.

Options: quantity before quality

Ask for at least three options, and evaluate none of them until the list is finished. The moment you judge the second idea, the person stops producing and starts defending.

The common failure here is your own solution dressed as a question. "Have you thought about just escalating it?" is advice with a question mark on the end, and everyone can hear it. If you hold information the person lacks, label the switch: "I am stepping out of coaching to tell you something." Then step back in. Advice does not corrode trust. Advice pretending to be curiosity does.

Coaching questions for the Options stage

• What options do you have? Give me three before we look at any of them.
• What would you advise a colleague in exactly this position?
• What would you try if that constraint disappeared? Which part of that could you still do with it in place?
• Which of these would you regret not trying?

Will: make it dated or it did not happen

Will, sometimes written as Way forward, is where the conversation becomes real or evaporates. The test is blunt: can you write one sentence containing an action, an owner and a date? If not, nothing has been agreed, however good the discussion felt.

Then check the commitment rather than assume it. Ask them to rate, one to ten, how likely they are to do it. Under eight deserves one more question: what would make that a nine? The answer usually exposes an obstacle Reality missed, and you loop back for two minutes instead of failing quietly for a fortnight.

Coaching questions for the Will stage

• What will you do? Say it as an action, not an intention.
• By when, exactly? What is the first step, and when this week does it happen?
• On a scale of one to ten, how likely are you to do it? What would raise that number?
• What support do you need from me, and what do you not need from me?

GROW is not a feedback tool. Coaching after feedback works. Coaching instead of feedback is avoidance dressed up as development.

If somebody's work is not good enough, say so plainly, in its own conversation, using a feedback structure built for it. Asking gentle questions until they arrive unaided at "my work is poor" is avoidance.

Running GROW in twenty minutes

A workable default split for a short one-to-one:

1. Goal, three minutes. The outcome of the conversation, and the outcome beyond it.
2. Reality, eight minutes. The longest stage. Resist every urge to solve.
3. Options, six minutes. Three or more, unjudged, before any comparison.
4. Will, three minutes. Action, owner, date, confidence rating.

Reality gets the most time and Options less than you expect. Managers invert this, spending thirty seconds on the situation and the rest of the meeting solving a problem nobody described.

Where the model breaks down

GROW is a good structure and a poor religion. Four honest limits:

1. It assumes the person can find the answer. Coaching works on a will problem or an awareness problem, not a knowledge gap. If someone has never seen a decent forecast, no number of questions will produce one. Teach first, then coach.
2. In unpractised hands it becomes an interrogation. Run as a checklist, GROW feels like being processed. The stages are a shape for your attention, not a script, and the person opposite can tell which one you are doing.
3. The authority gap does not vanish because you asked a question. When a manager coaches a direct report, Options degrades into guess what the boss is thinking. Without psychological safety you get compliant answers and read them as insight.
4. It is built for goals, not feelings. When the real subject is exhaustion, a bereavement or a conflict gone personal, marching toward an action point reads as dismissal. Some conversations should end without a W.

Variants patch some of this. T-GROW adds a Topic stage at the front, for people who arrive with a tangle rather than a goal. Solutions-focused models such as OSKAR spend less time on the problem and more on what already works. None removes the underlying requirement: sitting inside somebody else's uncertainty without rushing to fill it.

How to tell whether you actually coached

• Who talked more? If it was you, that was a briefing.
• Whose idea won? If it was yours, expect it to be executed at half power.
• Is there a date? No date, no commitment.
• Did anything surprise you? If you learned nothing new about their reality, you were confirming, not asking.

None of that is taught by the framework. Anyone can memorise four letters; what takes practice is holding the silence after a question, and catching the moment your own solution starts pushing to get out. That is why coaching is trained in rooms rather than on slides: put a manager into a live conversation about a real problem, mark the second they jumped to advice, run it again. The change shows up in the second attempt, never in the first explanation. It is the same reason experiential learning beats theory on anything with another person sitting in front of you.

Tour De Force runs this as live, experiential training for teams worldwide — online and in person. Talk to us, or play Gamified learning appsThe Weekly Challenge to see the method in ten minutes.