Great article, Andy ! For investigative analytics, don’t you think that even before we start thinking about measuring various ‘universal’ metrics such as Cohort Analysis, Funnel Analysis, etc. it is of outmost importance to first begin with the MOST important metric that actually matters to your product?
Basically, as Josh Elman from Greylock proposes, when you begin measuring you must start with a) Purpose b) Core Actions and c) Cycle time of core actions. In a way the minimum viable metric would be “How many times your users perform core actions within expected cycle time”. If you set up to measure this metric first and evaluate the product performance then I believe it is even Minimum of ‘Minimum Viable Analytics’.
What is your take on that ?
