5 mins read
Sooner or later a chief executive comes back from a conference and asks HR to "roll out OKRs". The request usually arrives with enthusiasm and little detail, leaving the HR leader to work out what the thing is, what it is for, and what their own part should be. This primer is for that moment.
OKR stands for objectives and key results. The method comes from Andy Grove at Intel, and John Doerr, who had worked under Grove, brought it to Google in 1999. Google kept it as it grew, which is why the term is now everywhere. The idea has not changed much: decide where you want to go, agree how you will know you got there, and check often.
An objective is a short, qualitative statement of direction, the kind of sentence a team could say aloud without notes. Under it sit two to four key results, each a measurable outcome with a starting value and a target. A guide to how objectives and key results fit together gives the full anatomy; in short, the objective supplies the meaning and the key results supply the proof.
Why the cadence matters more than the wording
Most first-time OKR programmes spend weeks polishing sentences and then leave the results untouched until the quarter ends. That is backwards. The value comes from the rhythm: a short cycle, usually a quarter, and a brief check-in every week or two in which each key result gets a current number, a status and a sentence on what changed. A polished OKR nobody looks at for eleven weeks achieves nothing; a clumsy one revisited every Monday will usually move.
This is also where OKRs part company with SMART goals. The SMART criteria, first set out by George Doran in 1981, ask for goals that are specific, measurable, achievable, relevant and time-bound. The "achievable" test pulls against ambition, and OKRs deliberately go the other way: a key result is meant to be a stretch, and landing at 70 per cent of an ambitious target is treated as a good quarter.
OKRs are not KPIs either
A key performance indicator, as KPI.org's basics describe it, measures how well an organisation or process is performing against its ongoing purpose. It is a health reading, watched and acted on when it drifts. An OKR is a decision to change a few specific things by a specific amount within a specific window. Turnover on a dashboard is a KPI. Cutting regretted turnover from 14 per cent to 9 per cent this quarter, with an owner and a weekly check-in, is a key result. Companies need both; trouble starts when every KPI is rewritten as an OKR.
What HR's role is, and is not
The instinct when handed a rollout is to own it: write everyone's goals, chase the updates, build the tracker. That makes OKRs feel like an HR process, which is fatal. Teams have to own their own objectives; a goal written for you by another department is a task, not a commitment.
HR's real contribution is to enable the rhythm. That means helping teams write outcomes rather than activities, coaching managers to run a twenty-minute check-in, protecting the cycle from being quietly abandoned in month two, and starting with a few willing teams rather than a company-wide mandate. It also means holding one line firmly: OKR attainment stays out of the pay review. The moment a key result feeds a bonus, people set targets they know they can hit, and the ambition disappears.
An HR team OKR, with numbers
It helps to practise on your own team. Suppose the objective is that new starters feel productive and welcome within their first month. The key results might be to raise the 30-day new-hire survey score from 3.7 to 4.4 out of 5, to increase the share of new hires who have a one-to-one with their manager in the first week from 60 per cent to 95 per cent, and to reduce the average time until a new starter's equipment and system access are complete from six working days to one. Three numbers, each checked weekly. After one quarter the team will understand the method better than any slide deck could teach it.
Where OKRs do not fit
OKRs are not for everything. Steady-state work such as payroll accuracy or compliance training completion is better served by a KPI with a threshold. Teams with no real autonomy cannot commit to outcomes they do not control, and asking them to write OKRs produces resentment. And an organisation that intends to link results to ratings or money should drop that plan or drop the OKRs, because it will not get both.
Within those limits, OKRs are a modest, useful discipline: a few hard things per quarter, a number on each, and a short weekly conversation about whether they are moving. HR is well placed to make that habit stick, provided it resists the urge to run it.
