Start planning next year's OKRs in September or early October. Run a retrospective on this year's OKRs at the end of Q3, draft next year's objectives in mid Q4, agree and communicate them by late Q4, and start January with direction rather than a blank page.
By Reuben Chircop, OKR coach and creator of OKRgen
Ask most teams when they will start planning next year and the answer is some version of the same thing. "We'll look at that in November. Maybe December."
December is too late. By then half the organisation is mentally on a ski lift and the other half is trying to close out the year. Nobody does serious strategic thinking in December, whatever the calendar invite says.
I watched this play out with a team I was brought in to help. I joined at the end of August. They began planning in November, at exactly the same time as they were building their December holiday campaigns. Campaigns that should have been tested and locked before November even started.
Every other week I gathered them to work on objectives and key results, and every other week we could not get there. Not because they were incapable. Because delivery was days away and their attention was, reasonably, on the thing in front of them.
Asking someone to think twelve months ahead while they are firefighting next week is not a planning problem. It is a mental shift they do not have room to make.
The campaign went out in December and underperformed. So January was spent doing two jobs at once: working out why the campaign had failed, and finally building the key results and initiatives that should have existed months earlier.
That is the real cost of late planning. It does not just delay the plan by a few weeks. It collides with everything else happening at the same time and makes both things worse. The plan is rushed because the failure needs attention, and the failure takes longer to fix because the plan is competing for the same hours.
Start in September or early October and none of that happens. It is far enough from year end that people can lift their heads. The pressure is not fully on yet. The mental shift is still available.
Annual planning works best as four phases spread across the second half of the year.
(end of Q3, early Q4)
(mid Q4)
(late Q4)
(January)
Organisations that only touch OKRs in Q4 are the ones for whom OKRs feel like an annual bureaucratic event, because for them that is exactly what it is.
Monthly. A short sync per team. Not a status report. A check on whether the key results are still the right ones and whether anything has changed.
Quarterly. A proper review. What did we achieve, what did we learn, what changes next quarter.
Q4. Next year's objectives are drafted, agreed and communicated.
If you meet monthly all year, the annual planning conversation is short, because everyone already knows where things stand. If you meet once a year, it is enormous, exhausting, and mostly spent reconstructing context that should never have been lost.
OKRs are not a document, they are a cadence.
Most planning stalls on things that have nothing to do with the objectives themselves. Put a name and a date against each of these first.
Not who sponsors it. Who moves it forward on a Tuesday afternoon when nobody is asking. If the answer is "the leadership team", it means nobody.
When two teams write overlapping objectives, someone needs the authority to resolve it. Name them in advance, or the conflict simply sits there.
If the meetings are not in calendars before the quarter starts, they will not happen, because the quarter fills up.
If those three do not have names and dates against them, you do not have a system. You have an announcement.
When you sit down to draft, skip the template and ask this first:
It is December 2027. What do we want to have achieved?
Revenue? Customer satisfaction? Employee engagement? Whatever the answer, make it something you can measure, and keep the list short. People should be able to remember next year's objectives without looking them up.
Here is a quick test for every objective you draft for next year. If it could be marked complete, it is not an objective. It is a task wearing the word.
Once initiatives have their own layer, the other two can finally be what they are supposed to be. The objectives stop being a to do list, and the key results stop being one too.
The best version of this I have seen took a team four quarters to learn. Their first year with OKRs began with sixteen objectives, most of which turned out to be task lists. By the end of it they had six objectives, meetings in the calendar, and a quarterly session that had stopped being a status update and become a place to ask for help and adjust.
Then they aimed to have the following year drafted by September. They missed it by a few days. January opened with a clear mission and a plan of action. They still adjusted in January and February, of course, but they were adjusting a plan rather than inventing one.
It is normal to be later than you would like, and it is recoverable.
Even with the owner named and the meetings booked, the drafting itself is where many teams stall. Staring at a blank page, trying to turn a strategy into objectives, is usually the moment people decide OKRs are bureaucracy.
OKRgen takes the friction out of that step. Describe what you are trying to achieve and it drafts objectives, key results and initiatives for you to argue about. The draft is the starting point, not the answer. The real work is the conversation it starts.
January stops being a scramble when people come back knowing what matters.