An OKR implementation starts with strategy. We write the first objective about halfway through

After more than twenty implementations, a recurring pattern shows up in these conversations. The problems companies come to us with are almost never problems with goals. They are management problems, and goals only expose them. Hence an order of work that is the reverse of what the client expects: first strategy and areas of responsibility, then annual and quarterly goals, and real work with the teams only at the end.

This pattern doesn't depend on company size, and that was the first surprise in the history of this format. Coffee with OKRs began because two people were implementing OKRs in parallel in completely different organisations, one in a company of eighty to a hundred people, the other in a large corporation, and they agreed to compare their conclusions. It turned out the problems were similar. Publishing those conversations grew into a company.

In this episode we name the four challenges that come back most often, and add a fifth during the conversation. A lack of clear priorities, meaning a list from which nothing can be chosen. Working on tasks instead of results. Being swamped by the day-to-day, after which nobody asks any more why they are doing something, because the more tired a team is, the less often it thinks about it. A lack of hard management skills: contracting a team, planning work, bringing meetings to a decision. The fifth, added on the spot, is a board that wants results but doesn't want to make decisions.

None of them takes weeks to diagnose. These patterns come out in the first two or three meetings, and after a number of implementations they are easy to recognise. Usually it is enough to sit in on one team meeting. The meeting doesn't end with decisions, it is followed by an email with a long essay from which it is unclear what was agreed, and tasks are managed in the inbox. Then one person phones everyone in turn and asks who has done what. That is why an implementation can be broken down into stages at all. We don't guess where to start, because it shows very early.

Stage by stage: from the company's picture of itself to its first quarter of work

The first stage is building the context, and the context is the strategy. We check whether the company has a named problem it is facing, whether it has policies, the guiding principles for dealing with that problem, and whether it has strategic goals. On top of that comes the picture of the market: how the company sees itself, its competitors, its target groups, and the products and services aimed at those groups. We aren't looking for objective data here. We are looking for how the board and the leaders see the company, because it is that picture that drives decisions, even when it is incomplete.

The second stage is writing it down. The strategy, and next to it the mission, vision and values, goes onto one page, and we check whether it is all consistent. The simplest exercise at this stage goes like this: several people fill in the same document separately, and then we compare the versions. The gap shows immediately, and that gap is the real result of the exercise, not the filled-in sheet itself.

At the same point we do something that is easy to overlook: a sheet of areas of responsibility. Who is responsible for what, where responsibility is shared, and which area has no owner. Very often one board member is convinced that another is handling a given area, and vice versa. Until that is settled, priorities have nothing to refer to. Prioritisation methods then stop working, because without a goal everyone compares from their own perspective and everything seems equally important.

Only now do the goals come: first annual, then quarterly. And only once the goals are written does real work with the teams begin, the part the client had in mind when they said at the start that they wanted to implement OKRs. We don't reverse this order, because a goal written without a point of reference is an epic under a new name, and the company will only find that out after several quarters.

What to do about it this quarter

The first step can be done in one afternoon and you don't need anyone from outside for it. Fill one page with everything you know about your organisation: where it is going, what problem it is solving, how it will know it has succeeded. Then give the same thing to several people, separately, and compare the versions. If the versions diverge, you have the answer to why priorities in your company are contested.

The second step is separating strategy from operations. For every larger task and every item on the roadmap, ask one question: if we had no strategy at all, would we still be doing the same thing? A yes means it is the day-to-day, not the execution of the strategy. It sounds trivial, but in practice it can split a list that nobody had been able to split before.

The third step concerns meetings, because inefficiency scales: the same topics come back at several meetings with the same people. Make a simple table: what is this meeting for, what result should it produce, who takes part, how often does it happen. Then merge or cancel. One well-known startup does a radical version and once a year deletes all recurring meetings from calendars, bringing back only those people themselves ask to attend. While you are at it, set up two mechanical things that are easy to put off: one place for tasks, where you can see who is doing what and by when, and all meetings in the calendar, with an agenda and a clear cadence. This isn't working with goals yet, but without it working with goals has nothing to hold on to.

The fourth step is a habit that won't appear on its own. Someone in the team has to ask, with every idea for a task or project, what effect it will have and which metric it will move. This reverses the order of planning: we don't plan tasks and then check whether they fit the goal, we derive what to do this week from the goal and a measurable effect. And if you have a team you think doesn't take responsibility, then before you change its tools, first sit in on its meeting with a good-meeting checklist and check whether a result was reached, what the atmosphere was like and whether the roles were clear. A lack of responsibility is an effect, and the cause is usually simply that nobody is managing that team.

What this order won't take care of

Not having a strategy doesn't block the start, and this matters for anyone who reads the above and decides they have nothing to start from. It is enough to focus on the most pressing problem, or to get one shared idea for the coming months out of the board. With a written strategy it is easier, but you can get moving without one too.

Nor will it stand in for the board's presence, and that is a condition no stage can get around. It happens that the board sends a representative to the work on goals, usually one of the managers or leaders, and then the decisions made with that representative go back to the board through the back door and are overturned there or made differently, because the representative had no real authority. A lot of work, a lot of people involved, zero decisions. The same mechanism works in a milder version: the board agrees to the change in words, but shows through its actions that it wants to keep things as they are. That is doubly demotivating, because it demotivates both the board and the employees, who see the difference between what has been called important and what time is actually spent on.

What this order won't do for anyone is decide which part of the organisation to start with and what structure of goals to build. That is decided by the goal of the implementation itself, a written and measurable answer to the question of what exactly this tool is supposed to help with. It isn't an empty shell for the start, but a signpost you keep coming back to along the way.

There is also something in this episode that is hard to say when you like this method. In one organisation the board announced that Product Owners were now working on OKRs, without first checking whether they were working on business value at all. The objectives were epics written in different words. This went on for several quarters, until someone said plainly that it added nothing. The team was right, not the board. If OKRs bring no value to the people who are supposed to work with them, that isn't a question of their attitude, but a sign that someone skipped the stages described above.

So before you start persuading a team to talk about goals, look for its real pain, and check whether the bonus system isn't paying for exactly the opposite. When a question came up in this episode about how to encourage a sales department to talk about strategy when it treats it as a waste of time, the answer wasn't that it's worth it. Strategy says which markets we sell in and to whom, so sharpening the persona lowers the cost of prospecting and raises conversion. That is an argument sales will buy. An implementation that goes well looks fairly boring, by the way: a few conversations about where the company really is, one document everyone refers to, and a quarter after which you can see we are doing fewer things at once.