How do you measure whether a promotion worked?
Measure it against the goal you set before it launched, not the number of entries it pulled. Decide up front whether the job was winning new customers, keeping existing ones, collecting first-party data or lifting sales, then track the one metric that maps to that job. A promotion worked when it moved the number it was built to move, and the surest sign is what people do after it ends, not how big the crowd was on day one.
What is the difference between a vanity metric and a real one?
A vanity metric counts attention; a real metric counts a change in behaviour. Entries, impressions and total prizes handed out feel good and prove little on their own, because a large enough prize will pull a crowd that never comes back. The numbers worth watching persist after the promotion closes: repeat purchases, retention rate, members who stay opted in, and revenue you can tie to people who took part.
How do you know the promotion caused the result, not something else?
Compare against a baseline, and hold something back where you can. Know your normal repeat rate, sales and opt-in numbers before you start, so any lift is measured against that line rather than guessed at, and where the setup allows, keep a control group of comparable customers who were not exposed and measure the difference. Attribution is never perfect, but a result you can read against a known baseline tells you far more than a big entry count and a hopeful feeling.
What does a promotion that worked actually leave behind?
Something you still have after the prize is gone: customers who came back, and first-party data you own and are allowed to use again. A one-off campaign ends with a winner and a spike that fades, while a promotion built as a repeatable loop leaves a base of members who keep taking part, which is what turns a single result into the loyalty that lasts. Motor Culture Australia, which runs on Sota, keeps around 90% repeat customers for exactly that reason.