The 99/1 Rule: Why One Point Decides Your Output

Why I Stopped Working Harder - And Why We Call It 99/1

Reading time: about 4 minutes

In 2005, I was working harder than everyone I managed. And it wasn't working.

I had started as a developer at 1&1 Internet AG and ended up running the PMO. Hundreds of projects at once - the kind of internet infrastructure people use every day without thinking about it. From the outside, it looked like a career going well.

From the inside, it looked like this: demand climbed, so we hired. More people meant more coordination, so I worked later. Output stayed flat. So I worked harder.

I had quietly started to believe that this was simply what responsibility costs.

Then a colleague put a book on my desk.

One idea

The book was The Goal by Eliyahu Goldratt. It contains one idea, and the idea is almost annoyingly simple:

In a connected system, a single point sets the pace of everything.

Not the busiest point. Not the loudest one. Not the one with the angriest emails. One specific point - and everything else in the system, no matter how hard it works, cannot go faster than that point allows.

We went looking for ours. It took two weeks to find.

It was a four-person team. And they were setting the speed for 1,500 developers.

Nobody had put them there on purpose. No org chart said, "These four people control the company's output." They had simply become the place everything had to pass through.

Once we knew that, the fix was almost boring. Within a few months, we had tripled the output. Same people. Same budget. No new tools.

But that's not the part that changed my life

Tripling output was good. It is not why I still talk about this twenty years later.

What changed my life was that I stopped being needed everywhere.

Here is the thing nobody tells you about a system with an unfixed constraint: somebody is holding it together personally. Usually, that somebody is you. All the chasing, the checking, the escalations, the "quick calls" - most of that work exists only because the system has a bottleneck and you are compensating for it with your own hours.

Fix the point, and that work doesn't get delegated. It disappears.

I worked less. The results got better. For the first time, those two facts stopped contradicting each other.

Why this matters more now than it did in 2005

Back then, when output was too low, you could hire. That escape route is mostly closed.

Skilled people are hard to find and expensive to keep. Budgets are tight. And most companies have already tried the obvious answers - a new tool, a new process, a reorganisation, a transformation programme with four workstreams running in parallel. Then AI arrived, and a lot of companies rolled it out everywhere and found remarkably little value at the end of it.

All of those approaches share one assumption: that improvement is spread out - that if you make many things a bit better, the whole gets a bit better.

In a connected system, that assumption is simply false. If you improve something that is not the constraint, you have not improved the system at all. You have made a fast part faster while it waits for the slow part. You spent money and effort and got nothing - and then you concluded that the method doesn't work, or that your people don't execute.

This is why so many capable executives are exhausted. They are not lazy and they are not bad at their jobs. They are running a system with one unfixed constraint and paying for it personally with their evenings.

So why "99/1"?

Most managers know the Pareto principle: 80/20. Twenty percent of the causes create eighty percent of the effects.

Pareto is true - for independent things. Twenty percent of your customers really do generate eighty percent of your revenue. Twenty percent of your products, your defects, your complaints. Those items don't depend on each other, so the effect spreads out.

But a company is not a pile of independent items. It is a chain. Sales depends on delivery; delivery depends on procurement; procurement depends on a decision that depends on you. Everything is coupled.

And a chain doesn't have an 80/20 distribution. It has a weakest link.

That's the whole difference. In a coupled system, the output is not "mostly" determined by a few things. It is determined at one point. Roughly one percent of your business governs roughly ninety-nine percent of your result.

That's the name. 99/1.

The practical difference between the two ideas is bigger than it sounds:

80/20 says: find the important twenty percent and work on all of it. 99/1 says: find the one point and work on that. Everything else can wait - and most of it doesn't matter at all.

The first sounds sensible and keeps you busy. The second sounds arrogant until the first time you see it work, and then it is impossible to unsee.

What it looks like in practice

Since 2005, I have brought this into more than 48 companies around the world - manufacturers, service firms, software teams, family businesses.

The pattern repeats. Output goes up without new headcount. Delivery reliability finally holds; one client moved to 100% on-time delivery and gained six EBITDA points without hiring a single person.

But here is what the owners actually tell me afterwards. They rarely lead with the EBITDA number. They say the business no longer runs through their inbox. They say their people take responsibility and act, instead of waiting for a decision from above.

That's the same thing that happened to me in 2005. It just happens faster now, because we no longer have to discover it from scratch - we've condensed it into a handful of principles a manager can use on a Monday morning.

The uncomfortable good news

If you are working harder every year for the same result, that is not a character flaw. It is not a discipline problem, and it is not because your team lacks ambition.

It is a system with one unfixed constraint.

This is genuinely good news for one reason: a character flaw takes years to fix. A constraint usually takes weeks.

There is one point in your business that explains your current result. My work is helping you find it - and then getting out of your way.

Want to know where your one point is? Start here: find your constraint - one-lever.com