Roman augurs — the priests tasked with reading omens before major decisions — were not permitted to begin a consultation by asking 'what do I want to find?' They had to fix the boundary of their observation field before looking, so the outcome couldn't quietly reshape the question. That constraint is almost entirely absent from modern product management, where we define success metrics after we see the results. The philosopher of science Paul Feyerabend noticed something adjacent: that researchers unconsciously adjust what counts as confirmation based on what they're hoping to confirm, not through dishonesty but through the ordinary human pressure to make meaning cohere. The augurs' ritual formalism wasn't superstition — it was an institutional solution to the problem of motivated interpretation. The practical move is anterior: before you look at any product metric, sprint outcome, or discovery readout, write down in one sentence what would constitute a genuine failure of your hypothesis. Not 'what we'll measure' — what would change your mind. That sentence, written before you see the data, is the boundary marker the augurs drew in the sky.

What is the last product hypothesis you tested where you wrote down, in advance, the specific result that would have made you abandon the idea — and did you actually do it, or did you write the post-mortem after you saw the numbers?

Drawing from Roman Religious Practice / Philosophy of Science (Feyerabendian) — Paul Feyerabend — Against Method (1975), particularly his critique of the retrospective rationalization of scientific discovery and the blurring of confirmation and disconfirmation criteria