Bundle, a startup that offers users a data-driven view of local businesses’ popularity based on credit card spending data, launched a new filtering feature at Finovate 2011. How well does it work?
Unlike Yelp!, which lets users write and review restaurants of their choosing, Bundle uses only credit/debit card data to determine which local businesses people actually frequent, and rates them based on loyalty, amount spent, and other personal metrics like your age and marital status.
They gather their data from 20 million Citigroup credit card users, as well as some government sources. The data comes to them thoroughly anonymized and aggregated, so there are no privacy concerns — at least not on the personal scale, anyway — just raw data demonstrating what type of person spends what kind of money, and where. At Finovate 2011, they announced a new feature that finally takes advantage of all this data they collect: filtering. Now, you can filter by neighborhood and demographics to find out where people most like you — or least like you — are spending their money. This yields interesting results.
A Yelp! Killer?
On the face of things, a data-driven local business listing seems like a brilliant idea. Yelp! is so full of spam reviews and terrible, whiny writing that cutting the human element out of knowing what humans do with their disposable income seems like an excellent innovation.
So, filtering out all the noise, looking just at the hard data, what is the number one restaurant for customer loyalty, in the whole New York City metro area, according to Bundle?
With no filters on, it’s Inside Track, the restaurant for gambling addicts, conveniently located right above the defunct Off Track Betting parlor in Murray Hill. An addiction to gambling combined with short times between posts, and maybe alcoholism, can make Inside Track appear to have the most loyal customer base in Manhattan — in a sense, they do — but not for the right reasons.
The people who frequent Inside Track don’t need Bundle to find out where to get a 7&7 between the 1:08 post at Tampa Downs and the 1:15 at the Aqueduct — they already know. (And they can’t anymore, because New York’s OTB went out of business a year ago).
Playing with the filters can yield funny results, at times. Let’s say you’re a young lady looking for a fun night out. Using Bundle you filter for 18 to 25-year-olds, and bachelors — hoping to find a good singles bar, or nightclub, let’s say — and you probably expect to see The Jane Hotel or some other rad downtown place pop up. Nope: your top two results are strip clubs. The page for the top result with these filters is 59 Murray Enterprises, better known as New York Dolls, the strip club right by Ground Zero.
The page for 59 Murray Enterprises — Anna Lindow, Bundle’s Audience Development Director, assured me they were working on converting the LLC names to business names — explains what brought it to the top: the purchase frequency at New York Dolls is high, 62% higher than most, and visitors tend to spend “a larger share of their budget at this location than they do elsewhere.”
Bundle is Working to Make Results More Useful
I spoke with Lindow last week, who acknowledged that funny search results like this “can be difficult to pull out.”
“We need to reassess how valuable it is to include [strip clubs] alongside bars and restaurants,” she explained. Strip clubs serve drinks and food, she acknowledged, but “is that the only thing the credit card is being used for? Maybe not!” But, she went on, they are looking into suppressing results of this sort, to make Bundle a more useful tool.
An example of how this might work is the most useful filter: the “foodie” filter. The button for the filter, strangely, depicts a piece of white bread, but it elevates the quality of the options considerably. How? This is where Bundle’s process gets a bit muddy, but also demonstrates how Bundle might need to tinker with their data, and not rely only on spending data, to make it useful.
Speaking with Lindow, she explained that the foodie filter works by lifting up certain results, based on consulting experts in the field, and anecdotal evidence. Going outside of the data pool sort of goes against their stated mission, but this way, they’re able to cut through the noise that shows how loyal people are to their corner deli, and show more quality eateries.
But the website aims to show people’s day-to-day spending habits, not where they sit down once a month. “No one’s going to say the food at Daniel isn’t good,” explained Lindow (who really underestimates how petty Yelp! users can be) but with the foodie filter you might be able to find out where someone who eats at Daniel goes to get a burger when they’re downtown, or where they go to the deli.
Foodie Filter Works Wonders
Interestingly, the foodie filter works for things completely unrelated to food. You can apply it to the “Clothing, Shoes, and Accessories” category, for instance. Normally 18 to 25-year-olds are loyal to Marshall’s, according to Bundle’s results, but 18-25-year-old foodies shop at places like APC, Paul Smith, and Top Shop.
And if you’re wondering where foodies shop for cars, you only get two results: Audi Manhattan and Mercedes Manhattan (and put it on their credit card!?!). Unintentionally, the foodie filter filters for class, but at least it lets us know something other than: gamblers like bars that are close to where they can gamble, or young men like strip clubs.
Bundle aims to tell us where people like us swipe their credit/debit cards on a day-to-day basis. With some tweaking of the filters, this will be an interesting tool, and a useful way to discover new businesses.
Right now, it’s a fascinating social experiment in human compulsion and social media. We have these personae on the Internet that, one hopes, are improvements on our actual selves. We can curate our Twitter account or Facebook postings or Tumblog to put our best self forward online. But we are far less perfect. Sometimes we — please note that this is not an “editorial we” — want to get drunk and throw our money away at horses, or strippers, or both. And right now, Bundle is an interesting window into our imperfections, a social media site that cuts through the way we present ourselves online, and shows us who we really are.
I sort of hope it stays this way, but based on what they tell me, it’s only going to get better.