Goal-Setting Frameworks: OKRs vs MBOs vs SMART — Which Fits Your Team

Three goal-setting frameworks dominate the modern workplace: Objectives and Key Results (OKRs), Management by Objectives (MBOs), and SMART goals. They are often discussed as if interchangeable; they are not. Each was designed for a different organisational context, and the most common reason a goal-setting program fails is that a team adopted the language of one framework while operating under the assumptions of another.

The three frames

Where each came from

MBOs were popularised by Peter Drucker in The Practice of Management (1954). They assumed a stable, cascadable corporate structure where the CEO's goals could be decomposed into divisional, departmental, and individual goals — and where annual performance reviews would naturally evaluate progress against them.

OKRs were developed at Intel by Andy Grove (drawing on Drucker), then carried to Google by John Doerr in 1999. The crucial innovation: OKRs were deliberately decoupled from compensation. Grove's view was that linking ambitious goals to pay incentivises sandbagging the targets — the opposite of what an ambitious goal system needs.

SMART originated in a 1981 Management Review article by George T. Doran. It is a quality criterion for individual goals, not a system for aligning an organisation. Treating SMART as your goal-setting framework is like treating spell-check as your writing process.

Comparison matrix

The OKR misuse pattern

1. Symptom: every team scores 0.9–1.0 every quarter. Diagnosis: goals are sandbagged because they're tied to ratings.
2. Symptom: nobody writes ambitious KRs. Diagnosis: psychological safety around 'missing' a stretch goal is absent.
3. Symptom: 15 KRs per team. Diagnosis: it's an output list, not a goal set. 3–5 is the design.
4. Symptom: KRs are tasks, not outcomes ('ship feature X'). Diagnosis: misunderstanding of KR design — KRs measure impact, not work.

Choosing by stage and culture

Implementation checklist

• Decide whether goals are tied to pay BEFORE choosing a framework.
• Train managers on KR vs task distinction — the most common writing error.
• Publish goals across the org (assuming OKRs) — visibility is half the value.
• Run a mid-quarter check-in, not just an end-quarter score.
• Score honestly. Sandbagged 1.0s destroy the system within two cycles.
• Commit to the chosen framework for at least 4 quarters before changing.
• Separate development goals from delivery goals — they have different cadences.

Frequently asked questions

Should OKRs be tied to compensation?

No — at least not directly. Andy Grove and John Doerr were both explicit on this. Tying OKR scores to bonuses incentivises sandbagging Key Results so that achieving them is easy, which destroys the ambitious-goal-setting muscle the framework is designed to build. Pay can reflect overall performance and contribution, of which OKR delivery is one input among many.

How many OKRs should a team have?

Three to five Objectives at most, each with 3–5 Key Results. More than that and the framework becomes a backlog, not a prioritisation tool. The point of OKRs is to surface what matters; a long list says nothing matters more than anything else.

What is the difference between a Key Result and a task?

A Key Result measures an outcome ('Reduce checkout abandonment from 28% to 18%'). A task is the work to achieve it ('Redesign the checkout page'). Writing tasks as KRs is the most common goal-setting error — it converts the framework from an outcome system into a project tracker.

Are SMART goals outdated?

Not outdated, but often misapplied. SMART is a quality checklist for writing a goal — Specific, Measurable, Achievable, Relevant, Time-bound. It is not a framework for setting goals across an organisation. Use SMART to check that an OKR or MBO is well-formed; do not use it as your goal-setting system.

How often should we revisit goals?

OKRs: quarterly cycle with a mid-quarter check-in. MBOs: typically annual with a mid-year review. The mid-cycle check is essential — without it, teams discover misalignment too late to act on it.

Where to read further