"Measure what you can" isn't the right approach. You can measure many, many things. Instead, you should measure things whenever they are good proxies for the o

"Measure what you can" isn't the right approach. You can measure many, many things. Instead, you should measure things whenever they are good proxies for the outcomes you want to drive AND there is good cost benefit in measuring+analyzing them. Surprisingly few things meet these criteria.