a stable process beats a high-performing one

Imagine two processes. One performs brilliantly on its good days, better than anything else you run, but it is unpredictable, swinging from excellent to poor without warning. The other performs modestly, nothing spectacular, but it is completely stable, producing the same result day after day. Asked to choose, most people instinctively want the brilliant one, because its ceiling is higher. In most situations that is the wrong choice, and understanding why gets at something fundamental about quality that often gets missed. A stable, predictable process is usually more valuable than a high-performing but erratic one, because you can manage, improve, and trust what you can predict, and you can do none of those things with what you cannot.

you cannot manage what you cannot predict

The core reason is that predictability is the foundation everything else in quality is built on. A stable process lets you plan, because you know what it will produce. It lets you make commitments to customers you can actually keep. It lets you build the rest of your operation on a known quantity. An erratic process, however high its peaks, gives you none of that, because you never know which version of it you are going to get. You cannot promise a customer a delivery when the process might have a bad week for no reason you can predict. The brilliance on the good days does not compensate for the inability to rely on any given day, because a business runs on reliability, not on peaks.

you cannot improve what is not stable

There is a technical reason a stable process is more valuable, and it is one of the most important ideas in quality. You cannot meaningfully improve a process that is not stable, because improvement requires knowing whether a change actually changed anything, and an erratic process swings so much on its own that you cannot tell the effect of your change from the noise. When a process is stable, you have a reliable baseline, and any change you make either shifts that baseline or it does not, visibly. When a process is unstable, you are trying to improve a moving target while blindfolded, unable to distinguish your improvements from the process's own random swings. Stability is the precondition for improvement, which means the modest stable process can actually get better over time while the erratic brilliant one is stuck.

erratic excellence hides its own problems

An unpredictable process that sometimes performs brilliantly also carries a hidden danger. Its good days mask its bad days, and its average can look acceptable while hiding wild swings that are doing real harm. A process that averages a good number by combining excellent and terrible is producing terrible output part of the time, and those bad periods have consequences, escaped defects, missed commitments, firefighting, that the flattering average conceals. The stable process at a modest level has no such hidden harm. What you see is what it does, all the time. The erratic process is lying to you through its average, and the brilliance on the good days is part of how it hides the cost of the bad ones.

stability first, then raise the level

The right sequence is the opposite of chasing peak performance. First make the process stable, even if that means giving up some of the spectacular good days to eliminate the terrible bad ones, because a predictable process at a modest level is worth more than an unpredictable one at a higher average. Then, from that stable foundation, work to raise the level, which you can now do reliably because you have a baseline to improve against. This is backwards from the instinct to chase the highest peak, and it is why disciplined operations obsess over reducing variation before they chase higher performance. Stability is not the boring alternative to excellence. It is the foundation that makes real, sustained excellence possible.

trust the process you can predict

Faced with the choice, take the stable process. Its modest, predictable performance is a foundation you can build on, plan around, improve from, and trust. The erratic high performer is a liability wearing the costume of a star, impressive on its good days and unreliable exactly when you need it most. The goal is not the highest peak you can occasionally reach. It is the reliable level you can always deliver, because that is what a customer actually experiences and what a business can actually be built on.

If you had to choose between a process that is brilliant but unpredictable and one that is modest but stable, which does your organization actually reward, and which one would your customers rather you ran?