One way of choosing what matters, across every part of the group.

What it was really about

On paper, this was an OKR programme. Underneath, it wasn't really about OKRs at all. It was about the whole group having one way to:

• choose the few things that matter most to growth and competitive advantage
• say them clearly, so every category and function hears the same thing
• agree what you'd actually see happen if they worked
• take calculated risks, knowing some won't play out, and say so out loud when they don't

Just as important was what it stopped. OKRs had become the way a category or division was seen and judged. When the score becomes the reputation, people write goals they know they'll hit, and the bold ones never make it onto the page. Taking that weight off the scores was what made the rest possible.

Which meant our job was never only to design the framework and train people on it. It was to talk to people, get to know them and what worried them, and help them see the better way of working as good for them rather than a risk. You can say that, and you should write it down. But nobody believes it until they've experienced it.

Where it started

OKRs had become the default container for nearly everything. Outcomes, outputs, discovery work and keep-the-lights-on activity all lived in the same place. Each one was there for a sensible local reason. Together they drowned out the signal good OKRs are supposed to give.

The categories with the longest OKR history had the hardest road. For them OKRs weren't just a planning tool. They were woven into how people managed the work, ran their rituals and reported upwards. Swapping in a new framework wasn't going to cut it.

The barrier was legacy, not capability

The teams that struggled most weren't the least talented or the least motivated. They were the teams where OKRs had been doing the most jobs for the longest time. Every habit had a history, and most of the histories were reasonable.

What worked was leading with the why. Teams didn't need telling what to do. They needed to understand, change by change, why doing it differently would make their goals clearer and their progress easier to see, and where the work that used to live inside OKRs would live instead. It meant explaining the same thing more than once. That's not failure. That's the job.

What we actually did

We planned ahead, then kept re-planning. We trained and coached the people who carry this day to day: leaders, OKR champions and team leads. We got to know them as people, not just as roles with OKRs attached, because that's where you learn what's really stopping someone.

And we changed the implementation many times. Not because the plan was wrong, but because we kept seeing things that nobody could have seen when the statement of work was written: a ritual that mattered more than it looked, a team that needed a different starting point, a meeting that was quietly doing a job nobody had named.

We built OKR quality measures, then spent a lot of time coaching team after team on why each one mattered, not just what it asked for. Quality went up across the organisation. The number of OKRs fell dramatically: fewer things, the ones that genuinely mattered, each with a clear line of sight to the company's goals. And because thousands of people were never all going to be in a room with us, we left behind a complete digital learning track, for everyone already there and every new starter who joined after us.

That's the honest shape of an OKR implementation at this scale. You need to know what great looks like, and you need to be able to adapt your way there.

The newest category went fastest

The category newest to OKRs had the least to unlearn. It arrived without embedded habits and with real curiosity about doing this well. It adopted the approach quickly, grew confident in its own judgement, and reached the point where it no longer needed coaching at all. That's the outcome we design for, and it told us a lot about the conditions under which this kind of change moves fastest.

Where it landed

Convergence was the goal, and in broad terms it happened. Not uniformly and not without effort, but across categories and functions there's now a much more consistent understanding of what OKRs are for. The processes, conversations and rituals that go with it are becoming normal rather than special.

Discovery already has its own home, and works better for it. The next step, and the one that will make all of this durable, is doing the same for every type of work, each with its own logic and its own reporting cadence.

Planning in a busy organisation

People were extremely busy, and getting time for planning was hard. So we adapted: a lighter, asynchronous planning process, big whiteboard sessions split into groups with several facilitators, and an idea submission and scoring workflow that fed the OKRs without needing everyone in one room. It worked. But planning squeezed into a forty-minute gap between meetings tends to produce incremental thinking, so we'd now treat time and attention for planning as a strategic problem in its own right, and protect it from the top.

What's next

• A shared definition of strategy. Much of what was called strategy was sensible operational improvement. Worth doing, but not a defensible position.
• Measures of winning, not just improving. Revenue and growth in usage say how you've done. A relative measure, such as share against named competitors, says whether you're winning.
• A living metric tree in every category/division/BU, understood by everyone, not just its leadership.
• Incentives aligned to the strategic tier, so leaders aren't rewarded only for what's easiest to count.

Why it matters

This is the version of the job most large organisations actually face. Not a blank sheet, but years of local practice, each part with good reasons for doing things its own way. The work is to keep what's good, give everything else a better home, and leave the organisation running it without us.

Not many people have done a project like this. We've done many. That's the part you can't get from a framework, a book or a template: having already met the legacy habits, the nervous teams and the plan that needs changing in month four, and knowing what to do next. It's that experience other organisations get to benefit from.