Most teams running EOS do not fail at Rocks because they lack ambition. They fail because the Rocks are vague, owned by committee, or quietly measuring the same thing the Scorecard already tracks. By week eight, half the list is "on track" in a way nobody can verify, and the Level 10 meeting turns into a debate about what "on track" even means.
Rocks are the EOS answer to the priority problem. Each quarter you pick the few things that matter most, assign them, and judge them at the end as done or not done. The discipline is in the constraints. Get those right and the quarter mostly runs itself.
Keep the number between three and seven
The classic EOS guidance is three to seven Rocks per person per quarter, and most teams should sit near the low end. Three is not a sign of low ambition. It is a sign that you understand a quarter is thirteen weeks and that real work, the kind that changes how the business operates, takes most of that time.
When someone has nine Rocks, what they actually have is a to-do list wearing a Rock costume. Nothing on a list that long gets the focus a Rock is supposed to get. Cut it. Ask which two or three outcomes, if achieved, would make the quarter a clear success. Those are the Rocks. The rest is regular work that lives in your to-do list and weekly check-ins.
One owner, always
Every Rock has exactly one owner. Not a team, not a department, not two co-leads who will "figure out the split." One name.
Shared ownership is the most common way a Rock dies. When two people own it, each assumes the other is driving, and the Rock stalls in the gap between them. The owner does not have to do all the work. They are accountable for the outcome and for reporting the honest status each week. If a Rock genuinely needs several people, the owner coordinates them, but the line of accountability stays singular.
Make the finish line binary
A Rock is done or it is not done. There is no 80 percent. This is the rule teams resist most, and it is the one that makes the system work.
A binary finish line forces you to define "done" before the quarter starts. "Improve onboarding" cannot be scored. "Launch the new onboarding flow to all new customers" can. You either launched it or you did not. Writing that crisp definition is half the work of setting a good Rock, because it surfaces the vagueness while you can still fix it. If you cannot state the done condition in one sentence, the Rock is not ready.
Company Rocks and individual Rocks are different things
Company Rocks are the three to seven priorities for the whole organization this quarter. They belong to the leadership team and set the direction everyone pulls toward. Individual Rocks are what each person commits to, and some of them support a company Rock while others handle work specific to that role.
Keep the two layers distinct. A company Rock owned by "the team" is a company Rock with no owner, which is the failure mode again at a larger scale. Assign each company Rock to one leader, then let that leader's individual Rocks ladder up to it. The V/TO is where the company Rocks connect to the one-year plan and the longer vision, so the quarter is not invented in isolation each time.
Do not let Rocks become Scorecard measures
This is the distinction that trips up experienced EOS teams. The Scorecard tracks the handful of weekly numbers that tell you the business is healthy: leads, cash, error rate, whatever your leading indicators are. Those numbers run continuously. A Rock is a one-time outcome you complete and then move past.
"Hit 50 demos a week" is a Scorecard measure, not a Rock. It is an ongoing target with no finish line. "Build the demo-booking system that makes 50 a week possible" is a Rock. It ends. When a Rock is really a recurring metric, you can never call it done, and it clogs your list quarter after quarter. Test every Rock against this: does it finish, or does it just keep running? If it keeps running, it belongs on the Scorecard.
Weak versus strong, side by side
A few rewrites make the pattern concrete:
- Weak: "Improve sales process." Strong: "Document and roll out the new five-stage sales process to the full sales team, with every rep trained by quarter end."
- Weak: "Better customer retention." Strong: "Launch the quarterly business review program for all accounts over $50k."
- Weak: "Hire engineers." Strong: "Fill the two open senior engineering roles with signed offers."
Each strong version names one outcome, implies a single owner, and has an unmistakable done bar.
Once your Rocks are set, they get reviewed every week in the Level 10 meeting, where each owner reports on track or off track and off-track Rocks drop into the Issues list to be solved. That weekly rhythm is what keeps a Rock from drifting until the final week. If you want help drafting and pressure-testing Rocks before the quarter begins, the EOS Rocks skill walks through ownership, the done condition, and the Scorecard test for each one.
The harder problem is connecting Rocks to the actual work that proves them, so status reflects reality instead of a weekly self-report. That is what Vindaris is built for: it ties each goal to the work delivering it and derives progress from that work, so a Rock marked done is done because the work is finished, not because someone said so in the meeting.