The performance objective says 'improve cross-functional communication.' Six months later the manager and the employee sit in a review and neither can say whether it happened, because nothing in the sentence commits anyone to anything observable. Multiply that by every review in the company.

The SMART acronym is usually credited to George Doran's 1981 article, and the acronym is the least interesting part. The part where goals live or die is the M, and the best treatment of it is Douglas Hubbard's How to Measure Anything. Hubbard's claim is that nothing a manager cares about is genuinely immeasurable, because measurement is uncertainty reduction, not exact quantification. His clarification chain does the work: if a thing matters, it has consequences; if it has consequences, they are detectable; if they are detectable, they are measurable. Most 'immeasurable' goals turn out to be unclarified ones — nobody has said what would look different if the goal were met.

Hubbard also names the trap on the other side, which he calls the measurement inversion: organizations systematically measure what is easy rather than what matters. That is exactly how 'measurable' and 'achievable' get corrupted in goal-setting — the team picks the countable proxy and quietly swaps it for the aim. David Hand's Measurement: A Very Short Introduction explains why the swap is dangerous. Most management measures are what Hand calls pragmatic: they partly define the attribute they claim to quantify, and once a measure becomes a target, people optimize the measure. Hand's warnings about reification and gaming are the standing caveat on every SMART rewrite — a proxy always misses something, and the discipline is to name what it misses rather than pretend it is the thing itself.

Instead of a worksheet, the rewrite runs as a service: fuzzy aim in, SMART objective out with one metric, target, and review cadence — plus honest flags for what made the original un-measurable and what the chosen proxy misses, and the matching OKR if you cascade upward.