Ask any founder whether their startup has goals and the answer is yes. Ask to see them, and things get awkward. There is a slide from the January offsite, a Notion page nobody has opened since March, and a general sense of “grow faster” hanging over everything. The startup does not lack ambition. It lacks a working connection between the ambition and what anyone is doing this week.
This failure has a predictable anatomy, and it is worth understanding before reaching for a fix.
The three ways startup goal-setting breaks
The first break is vagueness. “Become the market leader” and “delight our customers” are directions, not goals, and understanding what OKRs are is largely about learning to turn one into the other, a craft most founders have never been shown. Nobody can tell whether this week’s work moved a vague goal, so it stops informing decisions within days of being written.
The second break is volume. Startups are opportunity-rich and discipline-poor, so the goal list grows to ten or fifteen items, which is another way of having none. When everything is a priority, the actual priorities get set ad hoc.
The third break, and the most common, is set-and-forget. Goals get written with real energy and then never revisited, because no recurring rhythm exists to force the question “how is this going?”. The goals simply drift out of memory until the next planning session rediscovers them, usually with some embarrassment.
There is a fourth, subtler failure worth naming: goals set badly can be worse than no goals at all. The Harvard Business School working paper Goals Gone Wild catalogues how poorly designed targets narrow focus, distort risk-taking and invite corner-cutting. For a startup, where one bad quarter can be fatal, that is not a theoretical concern.
What OKRs change
OKRs, short for Objectives and Key Results, attack each break directly. The format is minimal: a qualitative objective that states what you want to be true by the end of the quarter, paired with two to four key results, each a number that proves you got there.
Against vagueness, the key results are the fix. “Delight our customers” becomes “raise NPS from 31 to 45” and “cut median support response from 12 hours to 2”. Suddenly the goal can be checked against reality, and this week’s work either moved a number or it did not. If writing measurable goals feels unnatural at first, that is normal; most founders have never been shown the craft, and working through a practical guide to setting goals shortens the learning curve considerably.
Against volume, the format’s constraint does the work. One to three objectives per cycle, no more. For a ten-person startup, that limit feels brutal in the planning meeting and liberating every week after, because it pre-answers the question “should we also do this?” with a default no.
Against set-and-forget, OKRs bring a cadence. Cycles run a quarter, and inside each cycle sits a short weekly check-in: what does each number say now, are we confident, what is the one thing we do next. Fifteen minutes. This rhythm, more than the goal format itself, is what separates teams that execute from teams that plan. A key result that has been flat for three weeks becomes a conversation in week three, while there is still time to act.
And against goals gone wild, OKRs carry a built-in safeguard: scores are decoupled from compensation. Targets are set deliberately high, hitting 70 % of a stretch target counts as success, and because nobody’s bonus rides on the number, nobody needs to game it. Ambition stays honest.
Where the fix does not apply
Fairness requires a caveat. If your startup is pre-product-market fit and the strategy changes fortnightly, quarterly OKRs will mostly generate rewriting. At that stage, one or two guiding metrics serve better than a full framework. OKRs earn their keep once there is a repeatable thing to grow and a team large enough that alignment stops being automatic, often somewhere around eight to ten people.
The real fix is the habit
It is tempting to read about OKRs, write a beautiful set over a weekend, and consider the problem solved. It is not. The startups that fail at goal-setting rarely fail at writing goals; they fail at returning to them. The framework’s real content is the boring part: a small number of measurable commitments, reviewed every week, scored honestly at the end, and replaced by a slightly better set next quarter. Any founder can install that loop in a month. Keeping it running through a rough quarter is harder, and it is precisely then, when everything is on fire and focus is scarcest, that the loop pays for itself.
