OKR guide

    Why OKRs fail

    OKRs usually fail because of how they were implemented, not because the method is wrong. Nobody owns the process, nobody explained it, leadership exempted themselves, the meetings were never scheduled, or planning started too late.

    By Reuben Chircop, OKR coach and creator of OKRgen

    There is a sentence I hear often enough that I now have a practised response to it. "OKRs aren't working here."

    It sounds like a verdict on the method. It is almost always a description of an implementation that was never really put in place. The objectives were written, perhaps, and shared in a slide. But the things that make OKRs work, the ownership, the explanation, the rhythm of looking at them again, were left to happen on their own. They never do.

    When I look closely at an organisation where OKRs have stalled, I rarely find a problem with the idea of setting outcomes and measuring progress towards them. I find a gap between announcing OKRs and actually running them.

    The five reasons

    The same few causes come up again and again. Most stalled implementations have at least two of them.

    • 1. Nobody owns it. There is a sponsor, perhaps, but no one whose job it is to move the process forward week by week.
    • 2. Nobody explained it. OKRs were announced, not introduced, so people guessed what they meant and guessed badly.
    • 3. Leadership exempted themselves. Teams were asked to adopt a new way of working that the people at the top did not use.
    • 4. The cadence was never scheduled. Objectives were agreed once and then not looked at again until the quarter was over.
    • 5. It started too late in the year. Planning began in December, so January opened with drafts instead of direction.

    All five are fixable. None of them require a different framework.

    A story: "Why?"

    I was once brought into a company halfway through their first year of OKRs. The brief was simple: they were not working, and could I fix them.

    Leadership had communicated one objective, a revenue figure. There was no reasoning attached to it and no strategy behind it that anyone below the top had seen. OKRs had been announced, but nobody had said when they would be set, who would set them, or how they would be used.

    My first question was whether anyone had been given training. The answer was not no. The answer was "Why?"

    That one word told me most of what I needed to know. In the absence of any explanation, people had drawn their own conclusions. OKRs meant more meetings, more reporting, more admin. Some had decided it was a new way to be tracked.

    People fill the gap with the worst plausible interpretation.

    So we started with training, for everyone, leadership included. That last part mattered more than anything else we did.

    Alignment that skips the top is not alignment, it is compliance, and people can tell the difference immediately.

    Then we found OKR champions, one per department. They were mentors, not administrators. Their job was to help their colleagues write and use OKRs, not to chase updates. The champions met weekly, so the same message reached every team in the same words, and questions surfaced quickly instead of turning into rumours.

    None of this involved rewriting a single objective. What changed was that someone owned the process, people understood why it existed, and there was a rhythm that kept it alive.

    Not better objectives. Ownership.

    The three questions that decide whether it works

    Before any objective is drafted, I ask three questions. The answers tell me more about whether OKRs will work than the objectives themselves.

    Who is driving it?

    Not who sponsors it. Who moves it forward on a Tuesday afternoon when nobody is asking. "The leadership team" means nobody.

    Who decides?

    Someone named in advance to resolve it when two objectives overlap or a key result turns out to be unachievable. Without that person, every conflict becomes a negotiation.

    Who meets, and when?

    OKRs are not a document, they are a cadence. The meetings are in calendars before the quarter starts, with names against them.

    If those three do not have names and dates against them, you do not have a system. You have an announcement.

    When the objectives were written in the wrong room

    The second failure I see often is quieter. The objectives are written by leadership alone and handed to teams. It looks efficient, but it is the old top down method in new vocabulary, and the teams then have to be convinced of objectives they had no part in shaping.

    The split that works is different. Leadership sets the mission and the reasoning behind it. The teams closest to the work shape how it is measured and what gets done, because they are the only ones who know.

    For a practical way to agree that split, see how to balance top down and bottom up OKRs.

    Buy in is not something you secure once the objectives are written. It is what writing them together produces.

    Late planning makes everything worse

    Every one of the problems above gets harder when planning starts in December. There is no time to explain, no time to involve teams, and January begins with drafts rather than agreed direction. If you can, start planning earlier.

    Already behind? The order that works

    If you are partway through a year that is not going well, this is the order I would fix things in.

    1. Name the owner. One person who drives the process, not the outcomes, with the authority to put meetings in calendars.
    2. Explain before you ask. One session on what OKRs are and are not, for everyone, leadership included. It costs an afternoon and saves a quarter.
    3. Agree objectives at the top first, with the reasoning attached. People can execute a number they understand. They cannot execute a number they have only been shown.
    4. Find your champions. One per area, chosen for credibility rather than availability.
    5. Put the cadence in the calendar before the quarter starts.
    6. Then draft the key results. This is the part everyone starts with, and it is the last thing that should happen.

    Only the final step is about writing OKRs at all.

    Check the OKRs themselves

    If the owner is named, people understand why, the meetings happen, and results still stall, then look at the OKRs. The most common problem at that point is objectives and key results that are really tasks. Run each one through the tick off test: if it could simply be marked complete, it is work to be done, not an outcome to be measured.

    If targets have become safe because people fear what a miss will do to their rating, see how to keep OKRs and performance reviews apart.

    When the blank page is the problem

    Even with ownership and cadence sorted, drafting is where many teams stall. Turning a strategy into objectives, key results and initiatives from nothing takes time most teams do not have.

    The free generator drafts an objective with measurable key results. The full OKRgen builder then helps you add initiatives, check the set's health and track it. Treat the draft as a starting point worth arguing about, not the answer. It does not fix ownership or cadence for you. Those still need a name and a date.

    FAQ

    Usually because of how they were implemented, not because the method is wrong. The common causes are that nobody owns the process, nobody explained it, leadership exempted themselves, the review meetings were never scheduled, or planning started too late in the year.

    One named person who drives the process, not the outcomes. They need the authority to put meetings in calendars and the time to keep things moving when nobody is asking. A committee or "the leadership team" in practice means nobody owns it.

    Expect the first cycles to be a learning period. In my experience it takes a few quarters of setting, reviewing and adjusting before the habit holds, and the teams that get there fastest are the ones that keep the review rhythm going even when the first results disappoint.

    Not before you have checked the implementation. If nobody owns the process, people were never told why, or the reviews never happened, a different framework will fail for the same reasons. Fix ownership, explanation and cadence first, then judge the method.

    Make your next OKR cycle the one that works