I think two major things are understanding platform standards and that what users say they want is often not what they need or what will work best for them.

Innovation is fine, but IUE and usability are often better served by following common platform standards (for general apps on the OS, not necessarily trying to match existing and often shitty enterprise software).

Also, it’s important to look for root causes when people ask for a specific feature. Many users focus on specific UI elements when they’re having issues, and in a complex system, it’s easy to wind up with bloat by solving the wrong problem. This is often how enterprise software becomes so unwieldy, with options and elements seemingly vomited all over the UI. What users think they want and what they need can be very different things.

Sorry if this is too general or basic. It’s hard to design a system that works for beginners and experts, and A/B testing a mix of users early with lofi prototypes helps a lot. For existing software, just watching users interact with it is massively helpful.

Is this what you meant? Happy to be more specific if I missed the mark.