We Ran EOS L10 Meetings for Months With a Professional Implementer. Here’s What the System Can’t Fix.

I want to be upfront about something before this article goes anywhere: I like EOS. The Entrepreneurial Operating System is one of the best frameworks out there for giving a growing company structure.

The Level 10 meeting format works. Rocks work. The accountability chart works. IDS works. I’m not here to trash any of it.

But I spent months watching a real company run L10 meetings the correct way, with a paid professional EOS implementer facilitating every single session and the same problems showed up week after week after week.

The facilitator did her job. The format was followed. The meetings happened on schedule. And execution still fell apart between Fridays.

That gap the space between the meeting and the work is what this article is about. Not because EOS is broken, but because EOS assumes something about human behavior that isn’t true at scale.

And if you’ve been running L10s for six months and still feel like your team is spinning, what I’m about to describe will sound painfully familiar.

The Setup

The company I observed operates across ecommerce and healthcare under one founder who manages a portfolio of 40+ entities.

Two of those businesses had adopted EOS and were running weekly L10 meetings with a professional implementer handling facilitation. She managed the agenda, tracked IDS items, reviewed rocks, assigned to-dos, and kept meetings within time.

She used ClickUp for task management between sessions. She tagged people, set deadlines, created subtasks. She followed up.

On paper, this is what good EOS execution looks like. In practice, here’s what actually happened. 

The Scorecard That Nobody Fixed for Five Weeks

In late January, the founder acknowledged during the L10 that the scorecard KPIs were wrong. Not slightly off fundamentally wrong.

The team was tracking metrics that had nothing to do with the business as it currently operated. New KPIs had been discussed weeks earlier but never implemented.

His exact words during that meeting: “This will be the last level 10 we do where the KPIs are incorrect.”

It wasn’t. The following week, the team reviewed the same broken scorecard. The fix was pushed to a Saturday session.

That session got rescheduled to Tuesday. Then to “tomorrow at 8 AM.” By mid-February, three L10 meetings later, the facilitator was still trying to schedule the scorecard reset.

Now here’s the thing. The EOS L10 format has a scorecard review built into every meeting. It’s literally step two of the agenda. And the team did review it every week.

They read the numbers out loud. They noted that 165 out of 168 accounts were suspended, a 98% suspension rate. They reported zero Amazon sales. They moved on.

The scorecard review happened. The scorecard just didn’t mean anything. And there was nothing in the system that stopped the meeting from proceeding with a scorecard everyone knew was garbage.

The format says review the scorecard. The format doesn’t say refuse to continue until the scorecard is accurate.

That distinction matters more than you think.

The Number That Fooled Everyone

Three weeks into the scorecard problem, something else surfaced that made it worse. The team had been reporting “net sales” as their primary revenue metric. In one meeting, they proudly reported $6,000 in weekly net sales for eBay.

The founder paused. “What does net sales mean again? Because when you guys say net, that’s not the same net in my mind.”

Turns out the team’s definition of “net sales” was revenue minus platform fees and refunds.

• Not the cost of goods. 

• Not shipping. 

• Not storage. 

• Not ad spend.

Just revenue with a couple of deductions. The actual net income that week was around $900. On $6,000 in “net sales.”

The founder’s response: “You could have net sales and have negative income.” He then mandated changing the scorecard to report net income going forward.

But think about how long this went on. The L10 scorecard review happened every single week.

A professional facilitator was on the call. And for weeks, maybe months, the team reported a metric that made the business look five to six times healthier than it actually was.

Nobody caught it until the founder asked a basic definitional question.

EOS tells you to review the scorecard. It doesn’t tell you to validate that the numbers on the scorecard are the right numbers to be reviewing.

That’s a human judgment call, and when nobody makes it, the meeting becomes a ritual with no teeth.

“On Track” With Nothing to Show

EOS uses “Rocks” quarterly priorities that should be the most important things a team accomplishes in 90 days. During the rock review section of the L10, each rock owner reports whether their rock is on track or off track. Simple.

In one February L10 I observed, the team had two active rocks: liquidation of all inventory stocks, and completion of P&L; reports for every business unit.

The facilitator asked the rock owners for their status. Both said “on track.”

Then the founder started asking questions. For the liquidation rock, the owner hadn’t obtained the SKU list from the warehouse. Hadn’t posted anything on Craigslist.

Hadn’t contacted liquidators. Hadn’t even set a deadline for the warehouse to hand over the inventory sheet. The founder had to say during the meeting:

“We need to set a deadline for this guy to give us the list. Then find the retail value. Put it on Craigslist. First person to come pick up everything wins. Done in 48 hours.”

That’s not a rock that’s on track. That’s a rock with zero measurable progress. But in the L10 format, “on track” is a self-reported binary.

There’s no mechanism that checks whether the person actually did anything since last week. The facilitator can ask, and this facilitator did ask, but the format itself doesn’t require evidence.

The facilitator’s response was to remind the rock owner to update their ClickUp task with progress.

She said: “Make it your notebook. So if we need to refer to it on where you are, that’s where the explanation would come from.”

Good advice. But the ClickUp task had existed for weeks already. It just wasn’t being updated. And next week, the same facilitator would ask the same question, and the same rock owner would say “on track” again.

Because nothing in the system prevents that cycle from repeating forever.

Where To-Dos Go to Die

EOS L10s generate to-dos. That’s by design. You discuss, you decide, you assign. The to-do gets a seven-day deadline and gets reviewed at the next meeting. Clean system.

Except here’s what I watched happen across multiple meetings:

The facilitator would assign a task during the meeting. She’d create it in ClickUp, tag the person, set a deadline, add the founder as a follower. Textbook.

Then the next week, the task wasn’t done. Not because the person refused. They just... didn’t open ClickUp.

One team member across a different division under the same company admitted it outright: “We don’t really have a way to store them.

Normally what I do is when I’m working on one, I’ll just get it done and send it over.” Their ClickUp was full of tasks that had expired months ago.

The facilitator reminded people to send end-of-day reports a basic practice where you write down what you accomplished before logging off.

She brought it up in one meeting:

“I haven’t received anything from your team yet. If you could send it starting next Monday, please. I don’t want to really tag you every now and then, but just do it before you log off.”

That phrasing tells you everything.

She’s been asking. They haven’t been doing it. And her only tool is to ask again, more gently this time. There’s no automated reminder.

No flag that says “Mariam didn’t submit an EOD for three consecutive days.” No escalation path. Just a human being reminding other human beings to do what they already agreed to do.

This is the EOS blind spot. The framework gives you a meeting structure that generates accountability agreements.

But it provides zero infrastructure for enforcing those agreements between meetings. The assumption is that adults will do what they said they’d do.

And in my experience watching these teams, that assumption breaks down within about two weeks of implementation.

The Facilitator Did Everything Right. It Still Wasn’t Enough.

I want to be clear about something: the EOS implementer on these calls was good at her job.

• She kept meetings on time.

• She moved discussions to IDS  when they went off track.

• She created tasks in ClickUp with proper owners and deadlines.

• She followed up between meetings.

• She nudged people to update their tasks.

When a conversation got too big for the L10 format, she correctly called for a separate offline meeting instead of letting it derail the session.

But she’s one person. And her enforcement tool is conversation. She can remind, suggest, tag, and follow up. What she can’t do is make the system itself enforce compliance.

She can’t make ClickUp ping someone’s phone at 4 PM if their task is due tomorrow and hasn’t been started.

She can’t auto-escalate a rock that’s been reported “on track” for three weeks with no subtasks completed. She can’t generate a report that shows the founder exactly which to-dos have been recycled across consecutive meetings.

In one meeting, she highlighted a team member’s progress on P&L; reports in ClickUp as a “company win.”

She literally celebrated one person using the tool correctly because nobody else was. She told the rest of the team: “That’s the practice that you guys need to do also within ClickUp.”

When using your project management tool is noteworthy enough to be called a company win, you don’t have a tools problem. You have an enforcement problem.

What EOS Assumes (And When That Assumption Breaks)

EOS works beautifully in the meeting room. Scorecards create visibility. IDS creates decision-making discipline. Rocks create a quarterly focus.

To-dos create weekly commitment. The L10 format ties it all together in 90 minutes.

But the entire framework rests on one assumption: that the people in the room will execute what they committed to before the next meeting.

EOS gives you the structure to make commitments. It doesn’t give you the system to enforce them.

For small, tight teams of five or six people where the founder is in every meeting, this works.

Social pressure does the enforcement. The founder rememers what you said last week andasks you about it on Tuesday morning in the hallway.

But the company I observed had multiple divisions, remote teams across time zones, and a founder running 40+ entities who physically could not follow up on every to-do from every L10 across every team.

He said it himself in one meeting: “I can’t solve any problems if I don’t know they exist because nobody’s giving me the information.”

He wasn’t wrong. His private label ecommerce operation had been running for six years without generating a single dollar of net profit and nobody surfaced it because the P&L; reports excluded months with zero activity.

The good months showed up. The bad months disappeared.

The L10 meetings continued. The rocks were “on track.” And underneath all of it, the business was bleeding money that nobody could see.

That’s not an EOS failure. EOS did what it was designed to do: it gave the team a meeting structure and a language for accountability. But it didn’t give them an enforcement mechanism that would catch a six-year-old reporting gap.

No meeting format can do that.

The Enforcement Layer That EOS Needs

After watching these meetings for months, I started asking a different question.

Not “why isn’t EOS working?” but “what would need to exist between meetings to make the commitments stick?”

The answers were surprisingly concrete. If a scorecard hasn’t been updated in 14 days, something should flag it automatically not wait for a human to notice during a meeting.

If a rock has been reported “on track” for three consecutive check-ins with zero subtask completion, the system should escalate it to the rock owner’s manager before the next L10, not during it.

If a to-do appears in two consecutive meetings without being completed, it should trigger a different workflow maybe it needs to be re-scoped, reassigned, or killed.

And if a team member hasn’t submitted a daily check-in for three days running, their manager should know about it without having to count the days manually.

None of this exists in EOS. It’s not supposed to. EOS is a framework, not a software platform. It tells you what meetings to have and how to structure them. It doesn’t automate what happens after the meeting ends.

That gap is exactly why I’m building ShiftFocus OS.

Not as a replacement for EOS. If you’re running L10s, keep running them. The meeting structure is excellent.

But between those meetings, you need a system that enforces the commitments your team makes automatically, with escalation paths, with visibility, and without relying on one human being to remember everything and remind everyone.

Because here’s what I learned watching those teams: the meetings aren’t where execution fails. Execution fails on Wednesday afternoon when nobody’s watching and the task from Friday’s L10 gets buried under 40 new messages in WhatsApp.

EOS gives you the meeting. ShiftFocus gives you Wednesday.