EOS Has No Strategy Layer. Here's How We Fixed It.

Three years into running EOS, I’m a genuine believer in what it does well. The accountability chart works. Rocks work. The L10 meeting rhythm works. If your company is operationally messy, EOS will clean it up — and it’ll do it faster than you expect.

But there’s a gap in it that took me longer than it should have to name clearly.

EOS is fundamentally an operations framework. The “vision” component — your 10-year target, 3-year picture, 1-year goals — is there, but it’s largely left for you to figure out. The system gets you aligned and executing. What it doesn’t give you is a mechanism to check whether you’re executing in the right direction. You can have green across your scorecard, tight L10s, rocks delivered on time — and still be drifting sideways relative to where you actually need to go. I’ve had those quarters.

Part of what makes this hard is that “vision” and “strategy” aren’t the same thing, and EOS mostly gives you the former. Getting your leadership team in a room to write down a 10-year target is useful, but it’s not strategy — and honestly, in 2026, a 10-year vision is of questionable value on its own. The pace of business has compressed. AI has accelerated that further. A fixed destination a decade out is less useful than a clear, honest view of where you’re headed over the next three years and what bets you’re making to get there.

The most painful version of this gap, for me, was in the connection between short-term execution and long-term direction. EOS doesn’t give you a mechanism for that. So we built one.

The first thing we changed was how we treat the 3-year picture. Rather than leaving it as a directional statement, we converted it into measurable outcomes — essentially OKRs. “Be the market leader” sounds good at an offsite but it doesn’t tell you anything when you’re setting priorities. “Own X% of the North American tuning market” actually guides decisions.

From there, we test our 1-year goals against those 3-year OKRs and convert them into 1-year OKRs as well. Then when we set rocks — which are effectively quarterly OKRs — we run them against the 1-year goals explicitly, using AI to pressure-test alignment before they get locked. The rough rule: about 80% of rocks should have a clear line-of-sight to the long-term trajectory. The other 20% can be operational necessities that just need to get done. If a team’s entire rock set is tactical with no connection upward, that’s a signal something’s off — either the rocks are wrong, or the longer-term picture isn’t specific enough to evaluate against.

The third piece is more of a mindset shift than a process change: treating strategy as a separate, ongoing discipline rather than something you do once a year at a retreat. Good strategy comes from deep industry knowledge translated into specific bets. You have to actually understand what’s happening in your market well enough to form a real point of view — not just a whiteboard full of possibilities.

EOS is still the operating system I run, and I wouldn’t swap it. But if you’re running it and you feel like you’re executing well but not really getting anywhere, this is probably the gap worth examining. The operating system and the strategy layer are two different jobs, and EOS only gives you one of them.