Locke & Latham's goal-setting theory — the science behind (and against) OKRs

60-Second Summary

Specific + difficult goals outperform vague or easy goals — by ~16% on average across studies.
Five moderators: commitment, feedback, task complexity, ability, goal conflict.
Goal-setting fails dangerously when the metric becomes the target (Goodhart) or when learning, not performance, is the right mode.
Use learning goals for new domains, performance goals for known ones. OKRs default to performance — that's the trap.

Edwin Locke and Gary Latham's goal-setting theory is the most validated theory in organisational psychology. Every OKR consultant unknowingly invokes it. Knowing the actual theory makes you faster than them at fixing broken goal systems.

The core finding

Specific and difficult goals lead to higher performance than vague ('do your best') goals or easy goals — assuming five moderators hold. The effect size has held up across 35+ years of replication.

Five moderators

Commitment — the person must own the goal; imposed goals decay.
Feedback — without feedback, goals can't be adjusted.
Task complexity — for highly complex tasks, learning goals beat performance goals.
Ability — beyond ability, harder goals depress performance.
Goal conflict — competing goals cancel each other.

The dark side of goals

Ordóñez, Schweitzer, Galinsky & Bazerman's 'Goals Gone Wild' (2009) catalogued the failure modes: narrowed focus, unethical behaviour, distorted risk preferences, reduced intrinsic motivation, and erosion of learning. Sears, Wells Fargo, and the Ford Pinto were all goal-setting disasters.

Implications for OKRs

For novel/ambiguous work — use learning OKRs, not stretch numeric ones.
Pair every OKR with explicit feedback cadence — weekly at minimum.
Cap to 3-5 OKRs to avoid goal conflict.
Never tie OKRs directly to comp — invites gaming.
Review the conditions every cycle; a goal valid in Q1 can be invalid in Q3.