Edition-over-edition data

A single edition's numbers are almost uninterpretable on their own. Is 14,000 visitors good? It depends entirely on whether last edition drew 11,000 or 18,000, and whether the same kinds of people came back. Edition-over-edition data is what gives every metric its meaning: registration pacing against the same weeks-out last cycle, exhibitor retention year to year, whether your "new visitor" growth is real growth or just churn replacement. It's also the raw material for everything predictive — no history, no forecasts. Building it takes two disciplines that are easy to skip. First, consistent definitions: if "attendee" included exhibitor staff in one edition and excluded them the next, every comparison built on it lies. Second, entity matching across editions: the same company registering as "Acme", "Acme Group", and "ACME SARL" looks like three companies and wrecks your retention math. The honest nuance is that clean comparison sometimes requires normalizing for things that changed — a bigger hall, different dates, a free-ticket push — before concluding the show grew or shrank. The common mistake is comparing raw totals across editions that weren't run under comparable conditions, then building strategy on the artifact.