Performance Coaching Models: Pick One Per Client, Not Per Coach (2026)

TL;DR: Performance coaching models are structured frameworks that move a session from a goal to a committed next step. Four are worth knowing: GROW, OSKAR, CLEAR, and FUEL. GROW fits a client who knows the goal and needs a plan. OSKAR fits a client who is stuck. CLEAR fits a client who does not trust the process yet. FUEL fits a manager coaching a direct report. Pick the model from the client’s blocker, not from your own preference.

I run a career coaching practice with 33+ active clients. In a normal week I use two or three different performance coaching models, and sometimes two of them inside the same hour. That is the part most guides skip. They teach a framework as though you are supposed to adopt one and become “a GROW coach.” You are not. The model is the shape of the conversation, and the right shape depends on what is actually blocking the person sitting in front of you.

The four performance coaching models worth knowing

GROW – Goal, Reality, Options, Way forward. The default, and the one most coach training starts with. It came out of Sir John Whitmore’s work in the 1980s and is still the most widely taught framework in the business (Performance Consultants, Whitmore’s own firm, publishes the canonical version). You name the goal, get honest about where things actually stand, generate options, then commit to a next step. I wrote a longer breakdown of where GROW breaks and how I run it.

OSKAR – Outcome, Scale, Know-how, Affirm and action, Review. Solution-focused rather than problem-focused. Mark McKergow and Paul Z Jackson built it around 2000 and published it in The Solutions Focus (sfwork is McKergow’s site). The engine is the scaling question: “on a scale of one to ten, where are you right now?” followed by “what would move you half a point?” It refuses to spend the session excavating the problem.

CLEAR – Contract, Listen, Explore, Action, Review. Developed by Peter Hawkins, the British coaching and systemic team-coaching author. It is relationship-first. The whole front of the session is spent agreeing what the session is even for, which sounds like overhead until you have watched a client spend forty minutes answering a question they never actually wanted to answer.

FUEL – Frame the conversation, Understand the current state, Explore the desired state, Lay out a plan. From John Zenger and Kathleen Stinnett’s The Extraordinary Coach. Built for a manager coaching a direct report, where the power dynamic is real and pretending otherwise makes the conversation worse.

Pick the model from the client’s blocker, not your preference

This is the part that actually changed my sessions. Match the model to the diagnosis:

• The client knows the goal and has no plan. GROW. It is built for exactly this and nothing else is faster.
• The client is stuck, flat, and says “nothing works.” OSKAR. The scaling question gets a number out of someone who has already decided the situation is hopeless, and a number can move.
• A new client, a guarded one, or a genuinely hard topic. CLEAR. Contracting first buys you the permission you need to push later.
• The client is a manager who has to coach their own team. FUEL. Teach them a model they can run without pretending they are neutral.

The failure I see most in other coaches is not picking the wrong model. It is picking one model in training and running it on every client for a decade. Two clients with identical goals and different blockers need different sessions.

Where each of these performance coaching models breaks

Every framework has a failure mode, and knowing it is more useful than knowing the acronym.

• GROW assumes the goal is real. If the client names a goal they think they are supposed to want, the whole session executes cleanly against the wrong target. Bad goal in, tidy useless plan out.
• OSKAR gets gamed when the relationship is thin. A client who wants to look like they are progressing will just say a higher number. It works best after trust exists, not before.
• CLEAR is slow. In a 45-minute session, contracting can eat a third of your time. Worth it for session one, expensive for session nine.
• FUEL collapses when the manager has already decided. If the “coaching conversation” is really a decision being delivered, the model becomes theater and the direct report can tell.

How to run a model without sounding like a script

The client should never be able to name the model you are using. If they can, you are performing the framework instead of coaching.

Three things keep it invisible. Hold the sequence loosely – if a client walks in already at Options, do not drag them back to Goal for the sake of tidiness. Ask the model’s question in your own words rather than the training-manual version. And let a session change models mid-flight: I open in GROW, hit a wall where the client insists nothing has ever worked, switch to a scaling question, and finish in OSKAR. Nobody announces the handoff.

What I do keep rigid is the ending. Whatever model ran, the session closes with a written, specific next step and a date. I keep that in a one-page coaching plan per client so the commitment survives contact with the following week.

What no performance coaching model covers

Here is the honest limit. All four of these frameworks describe what happens inside the hour. None of them describe the six days afterward, which is where the actual result is won or lost.

This is the problem I built Sage to solve. Before it, my session structure was excellent and my follow-through was held together with Notion pages, a calendar, and a spreadsheet I had stopped trusting. A perfect GROW session that produces a commitment nobody tracks is a nice conversation. The model gets you to the commitment; a system for managing clients between sessions is what turns the commitment into an outcome.

FAQ

What is the best performance coaching model?

There is no single best one. GROW is the best default because it is fast and clients follow it intuitively. Beyond that, match the model to the client’s blocker: OSKAR when they are stuck, CLEAR when trust is the constraint, FUEL when the coach is also the boss.

Do I need to tell my client which model I am using?

No, and usually you should not. Naming the framework mid-session pulls the client out of their own thinking and into watching your technique. Structure should be felt, not announced.

Can I combine performance coaching models?

Yes, and in practice most experienced coaches do. Switching from GROW to an OSKAR scaling question when a client stalls is common and works well. The risk is drifting between frameworks with no destination, so keep the ending fixed: a specific next step with a date.

How is OSKAR different from GROW?

GROW examines the current reality before generating options. OSKAR deliberately does not – it moves from the desired outcome straight to a scale and asks what would raise it slightly. When a client is fixated on why things are broken, that difference is the whole point.

Are these models only for workplace or performance coaching?

They started in business coaching but the structure is general. I use all four in career coaching, and the same logic applies in life and health coaching. The framework organizes the conversation; it does not dictate the subject.

What to do next

See how Sage fits your practice. Check the Sage pricing page and see whether the between-session side of your practice is worth systematizing.

Read next: Accountability coaching: what it actually is, and why it works – the mechanics of making a session commitment survive the following week.

Try it on your own clients. Sage has a 14-day free trial, no credit card. Run your next month of sessions in it and see what the follow-through looks like when it is not held together by a spreadsheet.