For my team, we use a rule of two naive implementations then the third is a refactor. Cleaning code is often wasted effort when you don’t have the full picture

For my team, we use a rule of two naive implementations then the third is a refactor. Cleaning code is often wasted effort when you don’t have the full picture or understand future variations. Like you mentioned with ripping out your refactor later. This improves the speed of implementation one and two because we don’t have to think as much. Then third iteration is faster as well because it’s obvious how to abstract.