Here's what doesn't work: hacking around and ignoring what doesn't work.
Here's what also doesn't work: doing your best with your work and then dismissing the elements that don't work as experiments.
The best experiments are experiments on purpose. They are done with rigor and intent. They are measured. They are designed to either fail or create an approach that can be scaled.
Great experimenters measure their results. They probe. They fail on purpose. And when they find something that works, they hand the knowledge over to operators and executors who can scale their work.
You don't get to call it an experiment after it fails.