Coaching skills: the GROW model and powerful questions, explained

A direct report comes to you stuck on something. The generous instinct is to solve it, you know the answer, you have ten minutes, so you tell them what to do and you both move on. Coaching is the discipline of resisting that instinct: holding the problem open long enough for the other person to find their own way through it, because the answer they reach themselves is the one that teaches them something.

GROW is a container, not a script

The GROW model was developed in the United Kingdom in the late 1980s by Sir John Whitmore, Graham Alexander and Alan Fine, and first written up in Whitmore's Coaching for Performance in 1992, a book that went on to sell more than a million copies and became the default training framework for managers learning to coach ("GROW model," Wikipedia, summarising its origins). The four letters map a sensible arc for any developmental conversation:

• Goal, what does this person actually want from this conversation, and from the situation?
• Reality, what is true now? What have they already tried? What is really in the way?
• Options, what could they do? Plural, and theirs before yours.
• Will (or Way forward), what will they commit to, by when, and how confident are they?

Whitmore's deeper claim was that coaching is about raising someone's awareness and responsibility, helping them see their situation clearly and own the next move, and that GROW is simply scaffolding for doing that. In practice, that means treating GROW as a checklist you can loop through, not a railway timetable. A real conversation will jump back from Options to Reality when a new fact surfaces; that is the model working, not breaking.

flowchart LR
  G(["Goal, what do you want?"]) --> R(["Reality, what's true now?"])
  R --> O(["Options, what could you do?"])
  O --> W(["Will, what will you do?"])
  R -."new fact surfaces".-> G
  O -."hits a wall".-> R
					

The honest limitation: GROW is a structure, not evidence of effectiveness. It was built from practitioner experience, not a controlled trial, and on its own it can become a hollow ritual, a manager marching through four headings while still steering hard toward the answer they had in mind. The container only helps if you fill it with genuine questions.

Why "powerful questions" beat good advice

Here is the finding that should give every manager pause. When Julia Milner and Trenton Milner asked 98 participants on a leadership programme to coach someone on a simple topic, recorded it, and had the sessions assessed, the managers overwhelmingly defaulted to telling, offering solutions and instructions, while sincerely believing they were coaching (Milner & Milner, "Most Managers Don't Know How to Coach People," HBR, 2018). The blind spot is the problem: you cannot fix a habit you do not know you have.

"Managers tend to think they're coaching when they're actually just telling their employees what to do.", Milner & Milner, HBR (2018)

Herminia Ibarra and Anne Scoular make the same distinction the centre of their work on the manager-as-coach. They separate directive coaching (you have the answer and transfer it, closer to mentoring) from nondirective coaching (you draw the answer out of the other person), and argue most managers are stuck at the directive end and need to deliberately practise the nondirective half. Their recommended starting structure is GROW (Ibarra & Scoular, "The Leader as Coach," HBR, 2019). The practitioner who has done most to make this usable is Michael Bungay Stanier, whose The Coaching Habit (2016) names the enemy precisely: the "Advice Monster", the reflex to jump in with a solution. His antidote is almost comically simple, ask, then ask "And what else?" (his "AWE question"), because the first answer is rarely the best one.

So the move is to make your questions open, short and curious. "What's the real challenge here for you?" outperforms "Have you tried X?", the second is advice wearing a question mark. And when someone gives you their first answer, resist closing; ask "What else?" once or twice before you respond. A powerful question is one you do not know the answer to.

flowchart TD
  S(["Someone brings you a problem"]) --> Q{"Your instinct?"}
  Q -->|"Tell them the answer"| A(["Advice Monster: fast, feels helpful, builds dependence"])
  Q -->|"Ask a real question"| C(["Coaching: slower, builds their capability"])
  A -.-> D(["They come back next time"])
  C -.-> E(["They solve the next one themselves"])
					

The limitation worth stating: coaching is not always the right register. When a building is on fire, you give directions. Ibarra and Scoular are explicit that there are moments for directive guidance, for situational judgement, and sometimes for no coaching at all. The skill is not "always ask", it is knowing when the moment can carry a question and when it cannot.

The skill is learnable, and the bar is low to start

It would be easy to read the Milners' finding as a verdict on managers. It is the opposite. Their headline is that managers can learn, participants improved their coaching after a relatively short intervention, on the order of fifteen hours of training and practice. Coaching is not charisma; it is a small set of behaviours you can rehearse: ask before you tell, stay quiet a beat longer, follow up the first answer.

The practical takeaway: lower the stakes. You do not need a certification to coach better this week. Pick one conversation, a regular 1:1 is ideal, and run it as an experiment: more questions, fewer answers, one deliberate silence. That is the entire on-ramp. This connects directly to self-awareness and reflective practice, because the only way to know whether you are coaching or advising is to review your own conversations honestly afterwards.

A worked example

Priya, an engineering lead, has a senior developer, Tom, who keeps escalating decisions to her that he is perfectly capable of making. The directive instinct, and the fast one, is to keep answering. Each answer trains Tom to come back. (Names and details here are illustrative.)

Instead she runs a GROW conversation in their next 1:1. Goal: "What would 'good' look like for you on these calls in three months?" Tom says he'd like to make them without checking. Reality: "What stops you now?" After a pause, and after Priya asks "What else?", the real answer surfaces: he is not unsure of the decision, he is afraid of being blamed if it goes wrong. That is a completely different problem from the one she thought she was solving. Options: "What could give you cover to decide?" Tom proposes a lightweight rule: anything reversible he just does and notes; anything one-way he flags. Will: "What will you try before we next meet?" He commits to the rule for two weeks.

Priya gave no answer and solved a bigger problem than the one in front of her, because the question opened up the fear she would never have diagnosed by handing out instructions. That is the whole case for coaching in one conversation: the advice would have fixed today's decision; the question fixed the pattern.

Frequently asked questions

Isn't coaching just slow? I don't have time to ask twenty questions.

The time objection is real but mismeasured. Telling is faster this conversation and slower over a quarter, because every answer you hand over trains the person to return for the next one. Coaching front-loads a few minutes now to remove the recurring tax of dependence later. You also don't need twenty questions, Bungay Stanier's case is that two or three good ones, used consistently, do most of the work.

What's the difference between coaching and mentoring?

Mentoring is largely directive, you share your experience and answers. Coaching is largely nondirective, you draw out theirs. Good developers of people move fluidly between the two, but most managers over-use mentoring and under-use coaching. See mentoring & sponsorship for where the directive mode genuinely fits.

Do I have to use GROW exactly?

No. GROW is one structure among several, and its strength is that it is easy to remember under pressure. Treat it as a loop you can enter at any letter and double back through, not a four-step form to complete in order.

What makes a question "powerful"?

It is open (can't be answered yes/no), short (one clause, not a buried suggestion), and genuinely curious (you don't know the answer). "What's the real challenge here for you?" qualifies. "Don't you think you should just talk to finance?" does not, that's advice in disguise.

How do I know if I'm actually coaching or just advising?

Record or replay a conversation in your head and count: how many of your sentences were questions, and how many were instructions? The Milners' research suggests most managers are shocked by the ratio. That honest count is the fastest feedback loop you have.

Related in the Toolkit

• Mentoring & sponsorship, the directive, experience-sharing mode coaching deliberately sets aside, and when to switch back to it.
• Developing future leaders & successors, coaching is the day-to-day engine of building the bench beneath you.
• Delegation as a development tool, coaching and delegation are the same instinct: resist solving it yourself so someone else can grow.
• Building team capability, individual coaching conversations are how a team's overall capability actually compounds.
• Creating a learning culture, a team where asking beats telling is the cultural ground coaching grows in.
• Self-awareness & reflective practice, the only reliable way to catch your own Advice Monster in the act.
• Self-awareness & emotional self-regulation, staying quiet through a silence is regulation work before it is a technique.
• Building coalitions & securing buy-in, the same open, curious questioning that develops a report also moves a stakeholder.

Where to go next

• "Most Managers Don't Know How to Coach People. But They Can Learn", Milner & Milner, HBR (2018), the research behind the uncomfortable "I thought I was coaching" finding.
• "The Leader as Coach", Ibarra & Scoular, HBR (2019), the clearest short treatment of directive vs nondirective coaching, with GROW as the starting structure.
• Coaching for Performance by Sir John Whitmore (Nicholas Brealey, 1992; later editions current), the original source for GROW and the awareness-and-responsibility philosophy underneath it. Widely available; see the GROW model overview for the short version.
• The Coaching Habit, Michael Bungay Stanier, the most practical question-by-question manual for taming the urge to advise.
• "How to tame your advice monster", Michael Bungay Stanier (TED), a short, sharp talk on the exact reflex that stops managers from coaching.