Average to Distribution

When the average describes almost nobody, show the spread, subgroups, and tails instead.

Your money app announces you spend an average of $310 a month on dining out, so you set a $250 target and feel responsible. But the average is a fiction describing no actual month: eight months you spent under $150, and four months, the ones with birthdays, visitors, and one bad December, you spent over $600. Budgeting for the average means budgeting for a month that never happens.

Average to Distribution is a 30-second unmasking for exactly this cue: the average describes almost nobody, or in this case, almost no month. The procedure: instead of the single summary number, show the spread (the range from lowest to highest month), the subgroups (ordinary months versus event months), the tails (that $700 December), and the change over time. The picture that emerges suggests a completely different intervention: ordinary months are already frugal and need no rule, while event months need either a sinking fund or a per-event cap. It works when the single misleading number resolves into two or three distinct populations, each with its own sensible action. The spec's warning: small samples cannot support elaborate segmentation. With only a handful of data points, slicing into subgroups produces confident nonsense; keep the cut coarse until you have enough months to justify it.