Bundle.com Crowdsources Reviews, Without the Crowd Knowing They’re Reviewing Anything

By  |  Thursday, September 30, 2010 at 11:03 am

While trust in the pro media falls, faith in the views of regular folk keeps growing. In an April survey by PowerReviews, for example, over half of consumers said that they trust user reviews of products, and they are reading more of them.

But the more they read, the more confused they can get. According to users of Rotten Tomatoes, for example, the new movie Wall Street: Money Never Sleeps is either a “Kick Ass, well research, solid story” or “an embarrassment.” It just depends on what review you read. (And the overall average rating of 56 percent doesn’t clarify.)

To remedy the user-review confusion, a new company called Bundle is committing the social-media blasphemy of rejecting what people say. Instead, it looks at what they do–mainly by analyzing credit card data to see how much money people spend, and where. (They also look at government statistics and third-party surveys.) Through a deal with Citibank, Bundle culls anonymous credit card info from 20 million shoppers to analyze spending habits. “So what if five people swear really wildly that this restaurant is lousy, if 95 percent [of the customers] go back,” says the company’s founder, Jaidev Shergill.

Launched in January, Bundle started by providing financial advice–a lot like Intuit’s Mint.com money-management site, for example, you enable access to all of your banking and credit card accounts in order to monitor spending, savings and debt. But Bundle also lets your gage your thrift or profligacy, or learn money-saving tips, by seeing the spending habits of other people just like you–that is, in the same age range, income bracket, family situation and zip code.

So what about those user reviews? That’s the next project. By mid-October, Bundle plans to officially launch its Restaurant Recommender site, starting with 4,000 of New York City’s local eateries (excluding known-quantity chains like McDonalds). Seeing where people eat, how often they eat there and how much they spend, Bundle promises to pinpoint the hot spots. And by filtering for folks just like you, it aims to better match your taste, style, and budget.

A rough version of the site is already online. The setup is clean and simple, but the metaphor is odd. Instead of starting with a location or type of food, as on Yelp, Bundle asks for the name of a restaurant you already like, and recommends similar ones that should please. Never been to New York and don’t know the names of any restaurants? You’re stuck, for now. Shergill plans to later build the site out with filters based on location and other (for now, unspecified) criteria. It will also link to restaurant sites or possibly–and ironically–reviews sites, even crowd-sourced Yelp. Personally, I’d like to take action by go straight to sites like Menupages or OpenTable.

One longstanding glitch, however, is that Bundle knows nothing about how people spend in restaurants where they pay in cash or with Amex. In Manhattan, that rules out much of the city below 14th street – a virtual cash-only (or maybe Amex) zone.

One remedy to the problem is incorporating that mendacious social media. Bundle does plan to ultimately parse user reviews, but only after grinding them through algorithms that filter out the crazies and narrow results to the foodies who visit a lot of restaurants and are better able to compare.

The company hopes to go national in the future. Meanwhile, if you’re hungry and outside New York, get ready to page through Yelp reviews.


Read more: , ,

4 Comments For This Post

  1. @piplzchoice Says:

    Customer reviews are amazing source of a very rich content, but is very underutilized for a number of reasons. Consumers often complain about inadequacy of 5 stars (Leikert scale) in terms of usefulness in making purchasing decision and many manufacturers find it too expensive to convert raw text data into actionable knowledge. The biggest winners are online retailers who see their visit to conversion rates increase substantially when a sufficient number of customer reviews is available on the product page.

    Our software mines opinions expressed in customer reviews to measure a difference between their expectations and their experience to produce specific, consistent and unbiased metrics that allow to compare products based on attributes that are important to the most important market segment – people who buy your product or product of your direct competitor.

  2. @ChefCindySez Says:

    Wish I had software that could “mine opinions” in our hotel dining rooms! The Likert 5/7- point psychometric scale is common in the hospitality industry (cf. today’s Los Angeles Times 3.5 Star Restaurant Review of Tom Colicchio’s Craft Los Angeles). Since my desserts have a decided advantage over other dinner courses because they are last, such an unbiased metric of a guest’s expectations based on the menu compared to actual dining experience would assist me in controlling daily quality in the bakery as well as suggesting dining room menu items. As a Pastry Chef, I am always concerned with our product consistency (reliability); I secondarily focus on comparisons to other fine dining on the Las Vegas Strip. Does my over-the-top deep rich chocolate mousse rescue a mediocre entrée and thus make the credit card charge seem “worth it”? Conversely, if I “burn” my Crème Brulee, do I not only “torch” the server’s tip but the goodwill toward our restaurant’s brand? Kudos to Bundle.com for analyzing crowd wisdom not in terms of what people say but what they actually do; dare I say “putting their money where their mouth is”?

  3. water damge orange county Says:

    Here is the 2nd time relating to come across your blog in the last few weeks. Appears like I should notice it.

  4. weight loss surgeon Says:

    I’m impressed, I have to admit. Genuinely rarely do I encounter a weblog that’s both educative and entertaining, and without a doubt, you’ve hit the nail around the head. Your concept is outstanding; the dilemma is actually a thing that too small folks are speaking intelligently about. We are pleased that we stumbled across this in my seek out something with this.