Craft compounds quietly
Taste looks like luck from the outside. Up close it is a decade of small corrections nobody applauded.
The gap between a good product and an excellent one is rarely a single decision. It is several hundred decisions, each of which could have gone the acceptable way and instead went the correct way. From the outside that reads as instinct. Inside, it is a habit of refusing the first answer.
Taste is trained, not issued
Nobody arrives with it. What separates the people who develop it is the volume of finished work they have shipped and then looked at honestly afterward. Shipping without review builds speed. Review without shipping builds theory. Only the pair builds taste.
You cannot critique your way to craft, and you cannot ship your way there either. It takes both, repeatedly.
This is also why craft is difficult to buy. You can hire a person who has it, but the standard does not transfer unless the organization is willing to slow down at the exact moments it feels expensive to slow down.
The practical version
- Keep a record of decisions you regret. Patterns emerge faster than you expect.
- Put unfinished work in front of people whose standard is higher than yours.
- Give the last ten percent a named owner, or it will not happen.
None of this is fast, which is precisely why it holds. Speed is easy to copy. A standard is not.