Most redesigns are answers to a question nobody wrote down. Here's the question we make every client write down first.
A redesign fails long before the first screen. It fails in the meeting where someone says the product looks dated and everyone nods, because at that moment the project acquires a goal that cannot be measured and therefore cannot be finished.
We ask for one sentence before we open a file: what should be measurably different in six months? Not prettier. Different, in a way a number can show.
The dated-looking trap
Visual staleness is real, and it is almost never the expensive problem. When we audited Nova Pay's signup, the interface did look ten years old. It also had a document upload step in position three that a third of merchants physically could not complete where they were standing.
Restyling that flow would have produced a beautiful screen with the same drop-off. The fix was structural, and we only found it because we watched people fail in the room they fail in.
Redesign as deletion
Our best projects ship with less. Kindred Market's checkout lost 34% of its fields and gained 23 points of completion. Terrace Health's vitals form went from fourteen inputs to six.
Addition is easy to justify in a roadmap review. Deletion needs evidence, which is exactly why deletion is where the leverage sits.
Ship in increments or don't ship
A redesign delivered as one detonation gives you no way to attribute what worked. We ship in two- to four-week increments with a measurement window between them, so by the time the last piece lands we already know which decisions earned their place.
It is slower on a slide and faster in reality.
The Musebit Agency team
Research, product design and design systems.