49.5% of companies running OKRs have stopped using them on at least one team, and 19% have stopped company-wide at some point. Our survey of 420 operations and strategy leaders found stalls peak two or three cycles in, half end without anyone deciding, and teams running four core habits hit their goals very consistently 80% of the time, against 10% for teams running none.
Ask the operations lead at a 120-person company whether OKRs are working, and the answer is usually yes: 69% of the leaders in our new study said their program is working well. Ask whether any team has ever stopped using them, and half say yes. Those two answers sit side by side in the same companies, because OKR programs tend to break down one team and one cycle at a time while the rest of the business carries on.
To measure how often that happens, when and why, we surveyed 420 operations and strategy leaders at technology companies of 50–200 people, all running OKRs today and none of them OKRs Tool customers. The full findings are in our new report, The Death of OKRs (and How to Revive Them). The short version is that stalls cluster two or three cycles in, half end without a decision, and the programs that survive share four habits that can each be in place within a single cycle.
Half of Companies Have Lost OKRs Somewhere
49.5% of companies have stopped using OKRs on at least one team after trying them. For 30% it was a single team or department, and for 19% it was the whole company. Every company in the sample runs OKRs today, so these are programs that broke down in part of the business, or stopped and restarted, while OKRs carried on elsewhere.
Partial stops usually begin with a team that ran OKRs because it was asked to, without a reason of its own. Updates slow down, nobody follows up and the rest of the company keeps going. A stalled team also signals to every other team that OKRs are optional, which is why the first team to stop is worth treating as an early warning of why OKRs die more widely.
The programs still running aren't all healthy either. Asked to describe their program right now, 25% of leaders called it working but fragile, 5% said it was struggling and 1% said it was barely used. A fragile program is still running, which makes it the easiest kind to rescue.
Stalls Peak Two or Three Cycles In
Only 18% of stalls happen during rollout. Leaders place the most common stalling point after two or three cycles (35%), with a further 25% after the first full cycle, so six in ten stalls happen once the program is live. Stop rates by program age show the same shape.
Among companies three to eight cycles in, 56% have stopped OKRs on at least one team. Past nine cycles, the figure falls to 22%, even though those companies have had the longest time to stop. The first OKR cycle runs on launch energy. By the second or third, the novelty has gone, OKRs compete with day-to-day work for attention, and if the weekly habit hasn't formed, updates slip without anyone noticing.
Cycles two and three deserve the same planning as the launch: shorter check-ins, fixes for whatever felt clunky in cycle one and progress that's easy to see. Programs that make it through the middle tend to keep going, which matches the shape of the OKR maturity curve.
Half of Programs End Without a Decision
When a team stops using OKRs, 39% of leaders say it happens through a clear decision. The rest end in ways nobody chose.
32% say OKRs are dropped with nobody deciding, and 18% say they stay officially in place while nobody uses them, so half of all endings are silent. A silent ending is harder to recover from, because nobody reviews what went wrong or decides what to keep. Individual goals usually have owners, but the program as a whole often doesn't, so nobody asks whether it still works. An end-of-cycle reflection with one person responsible for deciding, team by team, what continues, changes or stops closes that gap.
The 18% of zombie programs, still on the books with nobody using them, share one factor. Where progress is rebuilt just before reviews, 35% of leaders describe this pattern. Where progress is updated manually before check-ins, it's 20%, and where progress is tracked as work happens, it falls to 14%. A program whose OKR tracking only happens the night before a review can look alive for quarters after the teams stopped using it.
What Leaders Say Kills OKRs
Asked what causes OKRs to stall, leaders point first to lack of clarity (52%), followed by low team engagement (41%), too many goals (38%) and lack of leadership buy-in (35%). OKRs not connected to real work came next at 24%, then priorities shifting mid-cycle at 21%, with tooling last at 19%.
Among companies that have already stopped OKRs on a team, leadership buy-in climbs to 41%, the largest rise of any cause, so teams that have been through a stall are more likely to trace it back to the top and to leadership accountability.
Lack of clarity usually means one of three things: teams don't know which goal matters most, can't tell how a key result is measured, or can't see how their goals connect to the company's. Each one can be settled before the cycle starts with one number per key result, one owner and a visible link to a company objective.
The causes also change with the age of the program.
Early stalls, during rollout or after the first cycle, come down to confusion, with lack of clarity (47%) and low engagement (42%) leading the list. Late stalls, after four or more cycles or a slow fade, come down to overload and drift. Too many goals rises from 37% to 56% and becomes the top cause, because goals pile up cycle after cycle unless someone removes them.
Priorities shifting mid-cycle doubles from 16% to 32%, and manual tooling doubles from 14% to 28%. A new program needs fewer, simpler goals and a clear reason to care. A mature one needs the right number of OKRs reviewed every cycle and an agreed way to adjust OKRs when priorities change.
Programs Die Where Nobody Is Looking
Silent endings follow how visible OKRs are between reviews. They run at 64–67% where some OKRs lack an owner, retros happen only occasionally or progress is rebuilt before reviews. With each of those habits in place, they fall to 41%.
The three habits work as a set. Without an owner, nobody notices a goal slipping. Without live progress, it looks fine until the next review. Without a retrospective, nobody decides what to do about it. Leadership attention moves the number too: where leaders review OKRs weekly, 43% describe silent endings, against 55% where leaders review them monthly.
Who sets the strategy matters as well. Where strategy is set collaboratively across layers, 40% of companies have stopped OKRs on a team. Where the exec team sets it, that rises to 53%, and where one leader sets it alone, 74%, though that group is small (34 companies).
79% of companies that set strategy across layers describe their OKR program as working well, against 62% where the exec team sets it. People defend goals they helped shape, and a mix of top-down and bottom-up OKRs lets leaders set the direction while teams draft the objectives and key results that get them there.
The Spreadsheet Graveyard
Tooling was the least cited cause of stalling, at 19%. Outcomes by tracking method show a much larger effect.
Companies tracking OKRs in spreadsheets with no dedicated software have stopped OKRs on a team 56% of the time, and only 54% describe their program as working well. Companies on dedicated software without spreadsheets stop at 37%, and 76% say their program is working well. Spreadsheets also make endings quieter: 60% of spreadsheet-only companies describe silent endings, against 45% on dedicated OKR software.
The mechanism is simple. A spreadsheet has no reminders and no roll-up, so someone has to chase every update, and that job usually lands on the person who launched OKRs. When they get busy, the program goes quiet. Manual upkeep is easy to miss as a cause because it slips between meetings, where it's hardest to see, and that's the core weakness of OKR tracking spreadsheets once more than a couple of teams depend on them.
Goals You Can't Measure Don't Survive
The type of key result predicts survival. Companies whose teams write quantitative key results have stopped OKRs on a team 43% of the time, and 84% describe their program as working well.
Teams writing milestones stop at 57%, with 56% working well, and teams writing done/not-done key results stop at 66% (a small group of 32). Numeric key results hit more often too: 58% of teams writing them hit their goals very consistently, against 33% of teams writing milestones.
A milestone records whether a task got done, so once it ships there's nothing left to track, and the weekly check-in has nothing to discuss. Keep the project as an initiative and write the key result as the change it should cause, such as cutting onboarding time from 14 days to 7. Our guide on how to write OKRs shows that pattern across every function.
The Revival Staircase
We compared the 289 programs leaders describe as working well with the 131 that are fragile, struggling or barely used. Healthy programs are twice as likely to score every key result (65% against 31%) and to give every OKR an owner (60% against 30%). They're more than twice as likely to have leaders reviewing OKRs weekly, and nearly three times as likely to have almost all team goals linked to company goals.
Four of those habits do most of the work: a weekly check-in, progress tracked as work happens, an owner on every OKR and consistent retros.
[INSERT IMAGE: death-of-okrs-revival-staircase.png]Alt text: Teams hitting their goals very consistently rise from 10% with none of the four core habits to 20% with one, 50% with two, 71% with three and 80% with all four.
Teams running none of the four hit their goals very consistently 10% of the time. Each habit added moves the number up, to 20% with one, 50% with two, 71% with three and 80% with all four. Program health climbs alongside it, from 34% working well with no habits to 88% with three.
The habits make each other easier: an owner makes the weekly check-in someone's job, live progress gives it something to look at, and a retro turns it into a decision. Starting with one or two is enough to see the change, since teams running two already hit their goals very consistently five times as often as teams running none.
Long-running programs show the same pattern, in line with what our OKR Benchmark Report found about mature teams. Of the 72 companies past nine cycles, 72% give every OKR an owner, 76% run a retro at the end of every cycle and 81% describe their program as working well. They kept the same basics going through cycles three to eight, the stretch where other programs let them slip.
Eight Warning Signs a Program Is Stalling
Each sign below is linked in the data to a higher stop rate, more silent endings or a weaker program. Two or three together describe a program in the danger zone.
The report includes an eight-question self-assessment built on these signs. For a broader read on where a program stands, the free OKR Maturity Index scores it across each stage of adoption.
Keep Goals Owned, Visible and Measured Between Reviews
OKR programs stall in the middle, once the launch energy has gone and before the habit has formed. The common thread across all 15 findings is what happens between reviews: goals lose their owners, progress stops being updated and key results drift into to-do lists, until a team has OKRs on paper that nobody opens.
For the person accountable for OKRs, the revival plan starts with two moves that everything else depends on: put one named owner on every objective and key result, and make progress update as work happens so every review starts from a picture that's already current.
Teams with every OKR owned hit their goals very consistently 65% of the time, against 17% without. From there, give every key result a weekly confidence score, end every cycle with a decision on each OKR, and build goals with the teams who run them. Our playbook on bringing OKRs back after they've died walks through a restart step by step.
None of these habits depends on one person's energy if a platform carries the routine work. Automated check-in reminders, a live view of progress and owners required on every key result keep a program running through cycles three to eight, the stretch where half of programs are lost.
Data: The Death of OKRs (and How to Revive Them).
