Goal Setting That Survives Contact With Reality

Coaching goals are set in the least representative hour of a client's fortnight. They are focused, unhurried, talking to someone whose entire attention is on them. Of course the goal feels achievable.

Then Tuesday happens.

The problem is not that clients are unrealistic. It is that a goal set in ideal conditions encodes those conditions as an assumption, and no ordinary week supplies them.

Where SMART Goes Wrong in Coaching

SMART is fine for objectives someone controls. Most coaching goals are not — they involve other people, internal states, or outcomes that arrive on their own schedule.

"Be promoted by Q3" is specific, measurable, and time-bound, and the client does not control it. When it fails, the client experiences personal failure over something that was never theirs to determine. Meanwhile the behaviours that would actually make promotion likely were never named.

The fix is to split every goal in two.

Split the Outcome From the Behaviour

• The outcome is what they want: the promotion, the launched business, the repaired relationship. Keep it. It supplies the motivation.
• The behaviour is what they control: the conversations, the hours, the specific avoided thing they will now do weekly.

Then hold the client accountable only to the behaviour, and track the outcome as information. This sounds like a small distinction and changes the emotional texture of the entire engagement, because the client can succeed at their part every week regardless of what the world does.

Build In the Bad Week

At the point of setting a goal, ask: "What does this look like in a week where everything goes wrong?"

Most goals have no reduced form, so a bad week is a total break, and a total break is where abandonment begins — one missed week is not a problem, but the belief that the run is spoiled reliably is.

Define the minimum version in advance. If the goal is three deep-work sessions a week, the bad-week version is one. Named ahead of time, it is a plan; invented afterwards, it feels like an excuse.

Make the First Step Absurdly Small

Whatever the client proposes as a first step, it is too big — for the reasons covered in designing homework clients complete. Starting is the expensive part; once started, continuing is comparatively cheap.

Attach a Review Date, Not Just a Deadline

Deadlines create a binary. Review dates create a decision point.

Set a date — six weeks is usually right — where the goal is explicitly reconsidered rather than judged. Three outcomes are legitimate: continue, adjust, or drop. Making "drop" a named option in advance is what stops clients dragging dead goals through months of sessions out of embarrassment.

Knowing When to Change a Goal

Coaches are often too slow here, because changing a goal can feel like conceding failure. Signals that the goal, not the client, is the problem:

• The same commitment has been missed three times with a different reason each time.
• The client's language about the goal has gone flat — they describe it dutifully rather than with any charge.
• They have stopped mentioning it unprompted.
• Progress on the behaviour is fine and the client still feels no better.

That last one usually means the outcome was a proxy for something they have not named — the most productive conversation available in the whole engagement, and one you will only reach if you notice.

All four signals are patterns across weeks rather than events, which makes them nearly invisible in memory. They are obvious in a record: the same missed commitment in structured notes, the flattening language in weekly check-ins. That is the practical case for Interlude — not more data, but the few patterns that tell you when to change course.