The research against open plan offices is real, but it measures a particular kind of open plan: long rows of identical desks under flat fluorescent light with nothing to absorb sound.
Change three variables and the results change with them.
Acoustics do most of the work
Noise is not the problem — intelligible speech is. A room at 50 decibels of undifferentiated hum is easier to concentrate in than a silent room punctuated by one audible phone call.
We use felt baffles overhead, heavy curtains at the window line, planting as a soft diffuser and a low background of ventilation noise. The measured level barely moves. The perceived distraction drops enormously.
Zone by behaviour, not by department
Our studio has four zones: the library, where conversation is not permitted; the studio floor, where quiet talk is fine; the lounge, where calls and collaboration happen; and the booths, for anything private.
Members self-select constantly through the day. The zoning is only successful because moving between zones takes fifteen seconds and requires no permission.
Light people, not surfaces
Ceiling grids light the floor. Task lamps light the work. Warm, layered, dimmable lighting at desk level makes a 6:00 PM session feel civilised rather than clinical.