Most rollouts stall for the same reason. Someone reads a book, schedules a long offsite, produces a beautiful plan, and has nothing different happening by the following Monday. The order below reverses that, so you start the weekly habit early and refine the plan while it's running.
The plan assumes a company of 15 or fewer with one or more founders and no dedicated operations hire. It needs a document, a spreadsheet and a calendar. To make the steps concrete, we'll follow Harbor, an invented 11-person software company with three leaders and $40k in monthly revenue.
What can a team really get done in 30 days?
A team can install the habits in a month, but it can't finish the job. EOS® (a trademark of EOS Worldwide) says its full process takes approximately two years with an Implementer, so any promise of a complete system in a month should make you suspicious. What you can have by day 30 is a weekly routine people show up to, numbers they trust, and written rules for the few things that keep going wrong.
Here's the time budget, as our estimate for Harbor. Each leader spends about 11.5 hours across the month: 3 in week one, 2.5 in week two, 3 in week three and 3 in week four. Add an hour each for the other eight staff to read and comment. That comes to 42.5 hours, or about 2.2% of the company's 1,892 working hours in a month at 40 hours a week. Tell the team this before you begin, so nobody expects a big programme.
Week 1 (days 1 to 7): write down what's already true
Pick one person to run the rollout. In most companies this is a founder, or whoever already chases everyone. They own the calendar and the documents, and they get to nag.
Start with the last two months. Ask each leader to list the five things that went wrong or took far too long. These are your requirements, and you'll use them to judge the system later. At Harbor, the list includes a renewal lost because nobody owned it, two hires delayed by unclear approval, and a feature built for a customer who'd already left.
- Hold a two-hour session. Agree on one goal for the quarter and at most three priorities that serve it. Harbor chose "Reach $55k MRR by quarter-end", with priorities for renewals, a self-serve trial and hiring one account manager.
- Assign one owner to each priority. Where two names come up, pick one. Some project theories hold that only one person should be accountable for each deliverable, and at this size it's a rule worth adopting.
- Write it on one page and share it. Goal, priorities, owners, plus a line on what you've decided not to do this quarter.
- Tell the whole team on Friday. Explain the plan, the next four weeks and why you're doing it, in ten minutes.
End-of-week test: ask two people outside the leadership group to name the quarter's goal. If they can't, the page isn't clear yet.
Week 2 (days 8 to 14): start the meeting and the numbers
Now put a weekly meeting on the calendar: same day, same time, for 60 minutes. Don't wait until the numbers are perfect, because the first few meetings are how you find out what's wrong with them.
Pick five to seven weekly numbers. EOS's Scorecard allows 5 to 15, and we'd stay at the low end under 15 people, since someone has to gather them by hand at first. Each number needs one owner, a target and a source. If the owner can't get it in ten minutes, choose a simpler proxy for now.
| Number | Owner | Target | Week 2 actual | Source |
|---|---|---|---|---|
| New MRR added | CEO | $1,200 | $700 | Billing export |
| Renewals due in 60 days with no contact | Account lead | 0 | 4 | CRM filter |
| Trial signups | Marketing | 40 | 26 | Product analytics |
| Trial-to-paid rate, trailing 30 days | Product | 18% | 11% | Spreadsheet |
| Open support tickets over 2 days old | Support | Under 5 | 9 | Help desk |
| Weeks of cash at current burn | CEO | Over 52 | 61 | Bank, updated monthly |
Use a simple agenda: 10 minutes on numbers, 10 on priorities, 5 on last week's commitments, 30 on the top problems, 5 to recap who does what. Report numbers as on track or off track, without discussion. Anything off track goes on a problems list, and that's where the talking happens. This single rule is worth taking from EOS even if you never adopt the rest.
End-of-week test: the meeting ended on time, and at least one problem got an actual decision.
Week 3 (days 15 to 21): settle who decides, and fix the handoffs
Go back to the list of things that went wrong in week one. Most will trace to one of two causes: nobody knew who could decide, or work got lost moving between people. Deal with these next.
Write a one-page decision rule
Split decisions by how easily you can undo them. Jeff Bezos calls reversible ones two-way doors and suggests a lightweight process for them. Let the nearest person decide those, and tell others afterwards. For the expensive ones, such as hires, pricing changes or long contracts, use a small structure like Atlassian's DACI: one Driver who gathers input, one Approver who decides, Contributors who have "a voice, but not a vote", and people who are simply Informed.
Harbor set one spending threshold: anyone can commit up to $500 without asking, leaders up to $5,000, and anything above needs the CEO. Start a decision log with the date, the decision, the owner, what you expect to happen, and a date to check. Five lines is plenty.
Document three handoffs
Pick the three that fail most. Common ones are sale to delivery, support to product, and signed contract to invoice. Write each on one page: who hands over, what must be included, by when, and how the receiver confirms. Have the people who actually do the work write it, and then test it on the next real case. Harbor's first draft for sale-to-delivery lasted one week before the account lead added a field for what the salesperson had promised.
End-of-week test: take a recent decision that dragged on and ask whether the new rule would have named the decider. Take one handoff and run it from the page alone.
Week 4 (days 22 to 30): add cash, then review the whole thing
Add a monthly 30-minute cash review with cash, burn and runway, and compare them with last month. Ask whether you'd reach profitability on current money, which Paul Graham calls being default alive. If you don't have the number yet, this runway calculator will give you a first estimate. Decide what would trigger action: Harbor agreed to review its hiring plan if runway fell below 12 months.
On about day 28, run a retrospective for 45 minutes. The structure comes straight from plan-do-check-act, the improvement cycle that Walter Shewhart originated at Bell Labs and W. Edwards Deming later modified: you planned and did, so now check against what you expected and adjust. Bring the original list from week one. For each item, ask whether it would be caught now. Then make three lists: keep, change and drop.
- Drop something. If everything survives, you probably added too much. Harbor dropped its trial signups number, since the rate told it more.
- Set the next 90 days. Choose the priorities for the next quarter, using the same format, and add the next part only when a problem calls for it.
- Look at tools last. After a month you know which reports eat the most time, so now you can decide whether new software is worth it. The tool consolidation playbook can help.
Your day-30 checklist
Mark each line honestly. Four or more passes is a good first month, by our rough standard. Below that, repeat the weak weeks instead of moving on.
| Check | Pass looks like | Not yet looks like |
|---|---|---|
| Goal | Most of the team can name it unprompted | Only leaders can, or it changed twice |
| Owners | Every priority and number has one name | Shared ownership, or "the team" |
| Weekly meeting | Three in a row, on time, with decisions | Status updates that overrun |
| Numbers | Available before the meeting, and trusted | Argued over during the meeting |
| Decisions | Rule written and used at least once | Founders still get asked everything |
| Handoffs | Three written, one tested on a real case | Written, and never used |
| Cash | Runway known, with an agreed trigger | Only the CEO knows the number |
How this plan maps to a named method
You don't have to choose a framework now. These parts translate, so a later switch won't waste the work. The mapping below uses each method's published terms, and it's our reading. None of their owners has reviewed it.
| This plan | EOS | Scaling Up | OKRs | Shape Up |
|---|---|---|---|---|
| Quarterly goal and priorities | V/TO and Rocks | One-Page Strategic Plan | Objectives and key results | Bets, set per cycle |
| Weekly numbers | Scorecard | Critical Numbers | Progress on key results | Not covered |
| Weekly meeting | Level 10 Meeting | Huddles | Not prescribed | Not covered |
| Problems list | Issues List and IDS | Not stated on the pages we read | Not prescribed | Not covered |
| Owners | Accountability Chart | Who What When | Not prescribed | Teams own projects |
| Cash review | The 8 Cash Flow Drivers | Cash Acceleration Strategies | Not covered | Not covered |
For the tradeoffs among the first three, read EOS vs Scaling Up vs OKRs. If your company is very small, Shape Up's advice is that two or three people can drop most formal structure, and you should too. Run only weeks one and two, and revisit the rest as you grow. There's more on that in business operating system for startups.
If you'd rather hold all this in one product, Nautis is one option. Task boards take both outcome and KPI tasks, the Decision Loop logs calls with a review date 30, 60 or 90 days out, and Startup Health tracks cash, MRR and burn from a monthly check-in. It doesn't include an EOS or OKR method, so you'd still run the meeting and rules yourself, and a document and spreadsheet can do the same job for free.
FAQ
Can you implement a business operating system in 30 days?
You can install a working first version in 30 days: goals, owners, a weekly meeting, numbers, decision rules and a cash review. You can't finish it in that time. EOS Worldwide says its full process takes approximately two years with an Implementer, and the habits keep developing.
How much time does the rollout take?
For an 11-person company with three leaders, we estimate about 11.5 hours per leader across the month, plus an hour for each other person. That's about 2.2% of the company's working time. The weekly meeting then continues at about an hour a week.
Do we need software to run a business operating system?
No. A document, a spreadsheet and a recurring calendar invite are enough for a first version. Consider software after a month or two, once you can see which reports eat the most time. Buying first often means configuring a system before you know what you need.
What should we do first if we only have a week?
Pick one quarterly goal, name an owner for each of up to three priorities, and hold one weekly meeting with a short list of numbers. That covers the parts that change behaviour soonest. Add decision rules and written handoffs when problems point to them.
What if the team resists more meetings?
Cancel something in exchange, and keep the new meeting to 60 minutes with a fixed agenda. Report numbers without discussion, so the time goes to real problems. If it still feels like a waste after a month, use the retrospective to change it.
Sources
- EOS Worldwide: What is EOS (checked Oct 31, 2026)
- EOS Worldwide: Tools (checked Oct 31, 2026)
- Scaling Up: growth tools (checked Oct 31, 2026)
- Scaling Up: Scoreboard software (checked Oct 31, 2026)
- Amazon: 2016 letter to shareholders (checked Oct 31, 2026)
- Atlassian Team Playbook: DACI (checked Oct 31, 2026)
- Wikipedia: Responsibility assignment matrix (checked Oct 31, 2026)
- Wikipedia: PDCA (checked Oct 31, 2026)
- Paul Graham: Default Alive or Default Dead? (checked Oct 31, 2026)
- Basecamp: Shape Up (checked Oct 31, 2026)
- Basecamp, Shape Up, Appendix 2: Adjust to Your Size (checked Oct 31, 2026)