> no one can effectively quantify employee performance in "knowledge work" That's certainly not universally true. You can look at engineer commits, bug counts,

> no one can effectively quantify employee performance in "knowledge work" That's certainly not universally true. You can look at engineer commits, bug counts, speed of execution, all sort of metrics that would be directly impacted by distractions. If these were plummeting under open plan it would be pretty clear to management that it was costing them more money, not less.