The team that debates speed versus craft is debating the wrong thing.
Every design week has a fixed budget. The honest question is not whether to spend it on speed or on craft. The question is what gets to be revisited later and what does not. Some decisions, once shipped, are nearly impossible to walk back. Others can be tuned in place for years. Treating them with the same care is the actual mistake.
Decisions you cannot revisit: information model, terminology, table schemas exposed in the URL, the shape of the customer’s mental map of the product. Once the customer has a noun for a concept, you do not get to rename the concept without bleeding trust. Once a table column exists in a CSV export, you do not get to remove it. Once the second pricing tier is named, you do not get to retitle it. These decisions deserve all the time. Slow down. Get them right. The “craft” framing makes sense here, but it is not really craft. It is permanence.
Decisions you can revisit: nearly everything visual. Padding, color, icon style, microcopy, illustration. These get tuned over years and the customer barely notices. Burning a week on the perfect empty-state illustration when the empty-state copy is wrong is the most common form of craft theatre. The empty state copy is the permanent decision. The illustration is the tunable one. Most teams have this exactly backwards because the illustration looks like craft and the copy looks like writing.
The signal I look for in a review is what the designer protected versus what she rushed. A designer who slowed down for the structural decisions and shipped fast on the visual layer is operating well. A designer who slowed down for the visual layer and rushed the structural decisions is doing the inverse of the work. This is independent of how the final screens look.
The counter-view is that fast visual decisions sometimes are the strategic decisions, because the visual layer is the brand. This is true at consumer scale and not true at most other scales.
The shift is to stop saying “we need to move faster” or “we need to invest more in craft” and start saying “which of these decisions are permanent and which are tunable.” Treat them differently. The argument disappears.