OKRs are one of the most cargo-culted frameworks in startups. Founders hear that Google uses them, implement them badly, watch everyone ignore them, conclude they don't work, and go back to running the business on instinct. Then they raise a Series A, an investor suggests they try OKRs again, and the cycle repeats.
OKRs do work. But most implementations fail for the same predictable reasons. Understanding those reasons is the difference between a framework that drives clarity and one that creates administrative overhead with no payoff.
"OKRs should make trade-offs obvious. If they don't, you're doing them wrong."
Why Most OKR Implementations Fail
Goals are too easy
OKRs were designed to be stretchy — Andy Grove at Intel intended them to be set at a level where achieving 70% was good. When founders set OKRs at levels they're confident they'll hit 100% of, they're not OKRs. They're a to-do list with extra branding.
No accountability mechanism
Setting OKRs without a regular review rhythm is like setting a fitness goal with no gym membership. The goals exist on paper but nothing in the operating system of the business drives people toward them. Weekly check-ins where OKR progress is explicitly discussed are not optional — they're the mechanism that makes the system work.
Disconnected from daily work
If your OKRs are reviewed quarterly and not referred to between reviews, they're not shaping behaviour. People work on what they're asked about. If nobody asks about OKR progress until the end of the quarter, people work on whatever's most urgent today — which is almost never what the OKR was tracking.
Too many
More than three OKRs per team per quarter is too many. If everything is a priority, nothing is. Five OKRs means a team is trying to optimise for five things at once. They'll progress on all of them a bit and make decisive progress on none of them.
The Right Structure
One objective. Three to four key results. That's it.
Theobjectiveis qualitative and inspiring. It answers: "What are we trying to achieve this quarter?" It should be ambitious enough to require real trade-offs, and clear enough that everyone on the team could repeat it without looking it up.
Good: "Establish a repeatable outbound sales motion that doesn't depend on the founder."
Bad: "Continue to grow the business and develop our sales capabilities."
Thekey resultsare quantitative and measurable. They answer: "How will we know if we achieved the objective?" Each one should be a specific number with a current state and a target state.
Good: "Increase SDR-sourced opportunities from 8 to 24 per month."
Bad: "Improve our outbound process."
Cascading Without Bureaucracy
Company-level OKRs set the direction. Team-level OKRs translate that direction into specific work. Individual OKRs — in most early-stage companies — are usually unnecessary and create overhead. The question to ask at each level: "If we hit these key results, does it directly contribute to the company objective?"
Cascading should take one workshop, not a month of internal politics. If alignment on OKRs requires extensive negotiation, the organisation either has too many people involved or doesn't have a clear company-level strategy to cascade from.
The Quarterly Rhythm That Works
- Week before quarter starts:OKR-setting session. 2 hours. Draft OKRs, debate them, lock them.
- Every week:15-minute OKR check-in in the team meeting. What's the current score? What's the blocker?
- Mid-quarter:One hour review. Are we on track? Do we need to adjust tactics (not the OKRs themselves)?
- End of quarter:Score the OKRs (0–1.0). Retrospective: what worked, what didn't. Feed into next quarter's OKRs.
What Good OKRs Actually Do
The best OKR implementations I've seen share one quality: they force the organisation to say no. When you have one objective and three key results, it becomes clear what doesn't belong in the quarter. The request that would take three weeks but doesn't move any key result gets deprioritised. The nice-to-have feature that nobody's asking for gets pushed back.
That clarity — not the goal-setting itself — is the real value of OKRs. A well-designed set of OKRs is a forcing function for prioritisation. And in an early-stage company where every person's time matters, consistent prioritisation is one of the most valuable things you can build.