[speedybird]

> Science itself is a non-holistic way of looking at the world, blind to what is not measured.

This is unfortunately a broad class of error; qualitative metrics which are difficult to measure and quantify get brushed aside by people who want to make rational data-driven decisions. Robert McNamara became infamous for this; the 'data driven' way he tried to manage the Vietnam War focused on hard quantitative metrics, like bodycounts, and de-emphasized or ignored qualitative metrics like popular opinion in Vietnam and America: https://en.wikipedia.org/wiki/McNamara_fallacy

In the software industry, those that would rely on instrumentation and telemetry to guide product development often repeat the same mistakes. Many times I've seen user feedback ignored, derided, and demeaned. "Users are dumb, they don't know what they want. Ask users what they want and they'll ask for a faster horse." These ostensibly rational data-driven designers are ironically irrational because they ignore the well established limitations of their data-driven approach.