OKR Government Implementation: Steps and Pitfalls for Public Entities

What are OKRs?

OKRs (objectives and key results) are a way of stating what an organisation wants to achieve and how it will know it has done so. The objective is simply what is to be achieved, ideally concrete and inspiring. Key results measure progress towards it: they should be specific, time-bound, aggressive yet realistic, and measurable and verifiable, so it is clear whether each one was met. ^[1]

The usual formula is “I will (objective) as measured by (key results)”, with three to five key results under each objective and, typically, a quarterly cycle. The method began at Intel under Andy Grove, and John Doerr introduced it to Google’s founders in 1999. ^[1]

How do OKRs differ from government KPIs?

Government entities in the Gulf do not start from scratch. In the UAE, for example, the General Framework for Government Performance Management sets out linked levels of indicators: national indicators, strategic indicators that measure whether a federal entity achieves its strategic objectives, operational and service indicators, and government enabler indicators. It requires at least two strategic indicators per strategic objective. ^[4]

The framework calculates achievement as the actual result divided by the target, and the ADAA system colour-codes it: green above 95%, yellow from 75% to 95%, red below 75%. ^[4] These thresholds appear in the version of the framework written for the 2017-2021 strategic cycle, so confirm the ones your entity currently applies. ADAA has evolved through several versions since its launch in 2008, and on 9 July 2025 the Proactive Government Performance System was launched, using AI analytics to track targets and the implementation of national strategies. ^[5]

So OKRs should not be presented as a replacement for the official KPI system. A KPI tracks performance that must be sustained and for which the entity is held accountable; an OKR states what management wants to change over a short period. Other GCC states have their own national frameworks and bodies for measuring government performance, so check the one that applies in your country before designing an extra layer.

Why would a government entity need OKRs at all?

Government policies and directives are often written as actions rather than outcomes, leaving appointed and career officials to translate them into measurable targets, and the intended result can be lost in translation. ^[3] OKRs help turn a directive into a result the public can notice.

The same source distinguishes between OKRs written by policymakers alongside the policy itself and OKRs written by the implementing agency when the policy arrives without measurable results, which it says is often the case. ^[3] The second kind is what most government departments need: a team that knows the directive and has to decide what will change this quarter.

Steps for implementing OKRs in a government entity

1. Start from the approved plan: every OKR should serve an existing strategic objective or official indicator, so staff can see how their work fits the whole picture, which the UAE framework pursues by linking levels through cause and effect. ^[4]
2. Choose a few priorities: Google recommends three to five objectives with around three key results each. ^[2] In the first cycle, start with one or two departments and fewer objectives.
3. Classify every objective: a committed objective is expected to be delivered with a passing grade; a stretch objective may not be fully met and may span more than one cycle. ^[1] Label each one, because departments that depend on your work need to know they will get only about 70% of a stretch objective. ^[2]
4. Write results that describe impact, not activity: “launch a platform” is an activity; “70% of applications completed online” is a result.
5. Check in regularly: Google describes a mid-quarter check-in that shows each team where it stands and an end-of-quarter check-in that prepares for final grading. ^[2] Ask each key-result owner to state their confidence in achieving it, not just the percentage complete.
6. Grade and learn: Google scores each key result from 0 to 1 and takes a rough average for the objective. ^[2] More important than the score is recording why something slipped and carrying the lesson into the next cycle.

Hypothetical example: a municipal licensing department

The department already has an official strategic KPI on customer satisfaction, and this quarter it wants to improve trade-licence renewal for small businesses. The figures below are illustrative inputs:

If the quarter ends with 60% of renewals online and processing at two and a half days, that is good progress on a stretch objective, yet it would show as yellow, meaning behind, if entered unchanged into a traffic-light system designed for committed targets. That is why the two layers must be kept apart.

What are the main pitfalls?

Mixing stretch goals with traffic lights

Google treats a grade of 60% to 70% as the sweet spot for stretch OKRs ^[2], whereas the UAE framework’s colour scale marks anything below 75% as red. ^[4] Grade stretch objectives by the same standard and managers will quickly learn to write safe ones.

Tying OKRs directly to rewards

What Matters describes OKRs as divorced from compensation, and Google says they are not synonymous with employee evaluations. ^{[1, 2]} Some government systems, however, build them into individual appraisal. In the UAE, Cabinet Resolution No. (161) of 2025 on the Injazati system, in force since 1 January 2026, sets three to five OKRs for each federal employee, each with two to five KPIs. OKRs make up 80% of the appraisal for leadership roles, 70% for supervisory, expert and special-grade roles and 60% for technical and administrative roles, and achievement below 65% of target is rated “unacceptable”. ^[6] In that case, keep individual objectives committed and realistic, and keep stretch objectives at team level, outside individual appraisal.

Business-as-usual OKRs

Mistakes Google identifies include OKRs that restate current work rather than priorities, objectives that can be met without real effort, low-value objectives that would not matter even if achieved, and too few key results to reveal a miss early. ^[2]

Too many objectives and mechanical cascading

Cascading objectives mechanically from sector to department to section to employee produces hundreds of OKRs and destroys focus. Limit OKRs to the levels that genuinely need coordination.

Excessive secrecy

At Google, OKRs are public so that everyone can see what others are working on. ^[2] In a government entity it is enough for departmental OKRs to be visible internally to related departments; that exposes conflicts and dependencies early.

Checklist before launching the first cycle

• Every objective is linked to a named strategic objective or official indicator.
• Each objective has one owner, and each key result has a known data source.
• Every objective is labelled committed or stretch, and the label is visible.
• Key results describe impact on the customer or service, not activities.
• The way scores are read is documented and, for stretch objectives, differs from the traffic-light system.
• Mid-quarter and end-of-quarter reviews are in the calendar.
• The relationship with the employee appraisal system that applies in your country is clear and written down.

Conclusion

In a government entity, OKRs are a tool for deciding what must change this quarter, not a new measurement system competing with official KPIs. Link them to the plan, classify each objective, and keep stretch goals away from traffic lights and rewards, and you will get real improvement instead of safe numbers.