More data isn’t worth gathering when the decision is reversible and cheap, when the data won’t actually change what you decide, or when the cost of delay exceeds the value of the certainty. Each of these is a reason to decide now and stop collecting. A reversible, low-cost decision can simply be made and adjusted later, so waiting for more information buys certainty you do not need at the price of time you do. If you can change course tomorrow at little cost, the fast path is to act, observe the real result, and correct, rather than study a choice you can undo.
The second case is the sharpest: data that will not change the decision is not worth gathering, however much of it you could collect. Before you go looking, ask what the data could possibly show and whether any plausible result would move you. If every likely outcome leads to the same choice, the analysis is theater. The third case is timing. When delay carries its own cost, a slipping window, a competitor moving, a problem compounding, the value of being more certain has to clear the cost of waiting, and often it does not. Certainty that arrives too late is worth less than a good-enough decision made on time.
This sits against two failures. Analysis paralysis insists on gathering more before every call, which stalls reversible, low-stakes decisions that should have been made and adjusted. The opposite error refuses to wait even when the decision is costly and irreversible and the data really would change it. The pivot is reversibility and decision-relevance: how hard is this to undo, and would more information actually change the answer. When the decision is costly or irreversible and the data would genuinely shift it, gathering is the right call. Otherwise, deciding is.
To locate the line, ask three questions before you collect anything. Can I reverse this cheaply if I am wrong? Would any realistic finding actually change my decision? Does waiting cost more than the certainty is worth? If the call is reversible, or the data would not change it, or delay costs too much, decide now and adjust. Gather only when the decision is costly and the data would truly move it.