>many people will possibly agree that this problem exists >a reliable way to measure whether it actually improves things After the process has been carried ou

>many people will possibly agree that this problem exists >a reliable way to measure whether it actually improves things After the process has been carried out, it is measured as successful if less people agree the problem exists. Very simple, very quantifiable. >Then do some thorough data analysis, trying to learn as much about the problem as you can. No, that's not necessary. That seems more like gatekeeping and wikilawyering, which is basically the actual problem - entrenched power networks refusing to let anybody consider changing things without first having produced ten man-years worth of proof. You don't solve the problem by doing what the problem tells you that you must do in order to be allowed to solve it.