Since the promotion was targeted at their existing audience, the bit that annoyed me was they didn't estimate
the amount that audience would have bought over that period even if they hadn't had a promotion to respond to. I have no doubt it would have been lower, but it looks from the graph at the bottom like JewelScent's weekly revenues were pretty substantial anyway, and the run up to Valentine's Day should have been a bumper week for them.
The flip side, of course, is they also don't guesstimate future revenue from new customers attracted by likes generated by the promotion, or from increased engagement (or revenue lost because people made purchases they were going to make anyway but at a 10% discount, or because they were annoyed by the promotion and unsubscribed). Even direct marketing often is more data alchemy than data science :)
