Creating Meaning from Data

Startup Sortable, the team behind CarSort.com, aims to help people find the right vehicle for their needs, including vehicles that fit into snug parking spots.

According to CarSort.com, there are 9 cars below USD 15,000 that are easy to park. Their length is less than 15 feet. How do I know this, by simply selecting the “price” and “easy to park” buttons on CarSort.com. Christopher Reid, co-founder of Sortable.com, of which CarSort.com is one of many sites, believes his vision is different from Google’s. “While Google’s goal is to help people find what exists on the web, Sortable’s goal is to help people find what doesn’t exist.” Reid wants to create meaning from data, to extract conclusions on what products might be best for consumers.

The team, schooled at the University of Waterloo in Canada, is trying to solve the very hard problem of helping people make informed purchase decisions, by turning unstructured data from disparate sources, into structured data and more importantly, meaning. This is what Reid believes is their “secret sauce.” According to their web site, the problems they’re solving include aggregating all the data necessary to make informed decisions, allowing users to discover the right product without being an expert, making all the data understandable, turning all the data into objective measurements and personalizing the whole experience to the user.

But along with their expertise in data mining, I found CartSort.com’s user interface to be nifty. A user can select a vehicle along with snapshot information and “pin” it on the left-hand side, making it easy to compare other vehicles that match the search criteria. And, based on which vehicle is pinned, the information about the other cars displayed on the right-hand side changes dynamically, displaying advantages and tradeoffs vis-à-vis the pinned vehicle. While the pinning feature is a bit hard to discover, it still makes the site easy to use.

CarSort also provides a simple way to visualize the pros and cons about cars and users can also “shortlist” their picks and share them through social networks.

In addition to CarSort.com, other web sites under Sortable.com include SnapSort.com for camera’s, GeekAPhone.com for mobile phones and LensHero.com for camera lenses. Reid plans to be in 20 product and service categories by the end of 2011 to “provide answers to anything.” Reid also plans to increase the number of sources of data to enhance the quality of information, including soliciting user-generated content. As a reward for contributing to the community, Reid is considering offering bounties, badges, leaderboards and other game mechanic features.

While Sortable is not focused on revenue, they currently make money through display advertisements and affiliate partnerships on a growing user base of 1.5 million visitors per month.

• E-commerce
• E-commerce
• artificial intelligence
• data mining
• startup
• startups