Exploring Bundle’s data
I’ve been playing around a bit with Bundle, a clever new tool which lets you scope out merchants — shops, bars, restaurants, hotels — based on actual aggregated consumer behavior, rather than reviews and other qualitative information.
Bundle’s based on a huge anonymized database of consumer credit-card and debit-card spending. So when you call up a merchant on Bundle, it’ll show you reviews, but it’ll also show you data like how often people who shop there go back.
The raw data is fascinating. For instance, the average spend at the Breslin is $70-80; at the Spotted Pig it’s $80-90; at Momofuku Ssam Bar it’s $90-100; at Lure Fishbar it’s $120-130, at Annisa it’s $200-210, and at Gramercy Tavern it’s $220-230. (At Per Se, it’s $890-900.)
And you can learn a lot just by seeing that Spotted Pig customers also tend to spend money at the Starbucks on the Montauk Highway in Hampton Bays.
Then there’s Bundle’s proprietary loyalty score, which is promising but which needs work. It’s based on three variables: purchase frequency, share of wallet, and popularity. And it’s meant to spit out a score in a range between 1 and 100. But I’ve been plugging in the restaurants I know with the greatest customer loyalty — the ones above, as well my own favorite, Oyster Bar — and so far the only place I’ve found with a score over 50 is Szechuan Gourmet, whose score of 60 still only counts as “average”. Poor Annisa only gets a 32, which might be related to its being closed for a large chunk of the database’s time period.
So I’m definitely looking forward to the future of Bundle, where the loyalty algorithm will get improved. But the technology here is really promising: we vote with our wallets every time we spend money, and now those votes are being counted and analyzed.
How does Bundle separate out places where people tend to split the check from places where they don’t? Or places where they tend to go as couples from places where larger groups are more common? It doesn’t, yet, and I’m not sure it will ever be able to. But they do say that they’re looking to put together subsegments of the customer base — foodies, for instance, or deal-hunters.
And in the meantime, we can just marvel at some of the data which is already there. Here, for instance, is the data for Robert’s Steakhouse, an excellent restaurant which just happens to be located inside a strip club. Now that, it turns out, is a great way to build loyalty. And an average check of over $500.