Do the autopsy on the winners too

Do the autopsy on the winners too
Nobody dissects success. You get a good result, you scale it, and when it eventually dies you have no idea which part was carrying it.
Run the same process on anything that works:
— Which stage outperformed your expectation? Success has a location too, and it is rarely everywhere.
— Was it repeatable or circumstantial? Timing, a competitor's absence, a seasonal effect you did not plan for.
— What did you do differently from the last attempt? Be specific. "It just worked" is how a lesson gets thrown away.
— What was the actual constraint on scaling: traffic availability, caps, approval, or cash?
— Which assumption turned out wrong in your favour? Those are the most valuable findings and the ones nobody records.
Write it up in the same format as your failures. Months later the file of winners is more useful than the file of losses, because it tells you what to try first.
Study what worked, or you can only repeat it by accident.