TLDR: The redesign flop was the version of this you can investigate. This is the version you cannot, because by definition you cannot point at it. When a business is built to ship output and never to measure outcome, its own decline goes invisible, undiagnosable and impossible to steer. Every business run this way has the problem to some degree. The only thing that separates them is whether they are willing to admit it, and that willingness is exactly what starting to experiment requires.
Click to discuss this article with your favourite AI
The last article had a clean ending. The number dropped, someone went looking, and the culprit was caught. This one does not have that, and I want to be upfront about why. There is no tidy example to show you here, because the businesses that have this problem almost never find out they have it. That is not a hole in the argument. It is the argument.
Twenty suspects and no evidence
The decline does not land on a launch day you can circle in the calendar. It builds quietly, under a constant stream of small changes, none of them tested on its own. So when someone finally asks what is dragging the numbers down, there is no answer in the building. Every change of the last two quarters is a suspect but where to start?
You are looking at twenty possible causes and no evidence that points at any of them. You cannot remove the thing that is hurting you, because you cannot find it. The loss does not show up as one bad call. It shows up as an inability to steer.
When things are going well you can live with this. Not knowing why you are winning is a luxury. The bill only lands when the numbers turn and you reach for the wheel and find it is not connected to anything.
It starts with what you reward
This is not really a data problem. It is a priorities problem, and you can read a business’s priorities straight off what it counts.
Most businesses count output. Features shipped. Tickets closed. Edicts to deliver two hundred CX fixes this quarter. Speed, throughput, volume, all of it a measure of how much left the building. Almost none of it asks the only question that matters afterwards, which is whether any of it worked. So the whole machine gets tuned to produce rather than to learn. Work ships, the quarter looks productive, everyone is busy, and nobody is accountable for the outcome, only for the output. Ship two hundred fixes and you will be congratulated for two hundred fixes, whether the numbers went up, down or nowhere.
In every business I have worked in that ran this way, I could feel this problem long before I could point at any single instance of it. That is exactly what makes it dangerous. There is no one broken thing to fix. There is a way of working that guarantees you will never know.
Speed is not the same as velocity
There is a word for this that has quietly lost its meaning. Teams say velocity when they mean speed, how many things we shipped and how fast. But the two are not the same. Speed is just a number. Velocity is speed with a direction attached. You can have enormous speed and no velocity at all, and if the direction is wrong, all that speed is doing is getting you to the wrong place faster.
Output over outcome is exactly that. A team shipping two hundred fixes a quarter has plenty of speed. Whether it has any velocity depends on a question nobody is measuring: are these changes taking us where we want to go? You cannot answer that without checking, and speed with no fix on the direction is not progress. It is just motion.
Which is why you cannot prove you do not have it
Here is the uncomfortable part. If you are not measuring outcomes at a per output level, you cannot claim all your outputs are working. A quiet dashboard is not evidence of health, it is just an absence of measurement. Every business run on output has some version of this, including the ones that look like they are winning. They might be winning despite half of what they ship, not because of it, and they would have no way of telling the difference. You cannot test your way out of a problem you will not look for.
It’s time to be honest with yourself
The useful question is not whether your business has this blind spot. If you ship without testing, it does, to some degree. The question is whether you are willing to admit it.
That admission is harder than it sounds, especially for teams who work hard and take pride in how much they get out the door. Saying “we do not actually know if our work is working” feels like an insult to all that effort. It is not. It is the first honest thing, and it is the doorway to everything after it. The businesses willing to say it are the same businesses that are ready to start experimenting or already are, because experimentation is just the discipline of finally checking. Everyone else keeps shipping into the fog and calling it progress.
Which leaves one more question. If nobody can tell what is actually working, who has been deciding what to do all along? Usually the loudest or the most senior person in the room. That is the next article.
Are you ready to make a change?
If your team is shipping hard and nobody can honestly say what any of it did to the numbers, that is the problem I help businesses fix, starting with the willingness to look.