Government Data Unbound

WeMakeItSafer.com is a platform where businesses and consumers can search for recall data on food, toys, electronics and other items. They can also register products they own and be alerted if they’re recalled.

Founder and CEO Jennifer Toney launched WeMakeItSafer in 2006, before any major push to put raw government data online. Toney had previously worked as a litigation consultant for companies facing product liability suits. One day she realized a recalled food product she’d been brought in to discuss was sitting at home in her refrigerator and it occurred to her that there’s no way for most companies to alert consumers about recalls. Methods that do exist, such as FDA press releases, tend to use a megaphone approach that is unlikely to reach most consumers.

At first, Toney’s team laboriously translated recall information from government press releases into in alert system and used computer learning technology so the system could transfer recall information itself and correct for common government errors. Now, most of WeMakeItSafer’s data is culled directly from websites managed by the Consumer Product Safety Commission, the Food and Drug Administration, the Agriculture Department and the National Highway Transportation Safety Administration.

WeMakeItSafer offers free searches to consumers for product recalls stretching back to 1999 and allows them to register for alerts when products they own are recalled. The company also helps companies register consumers for recall alerts through their own websites. WeMakeItSafer runs a paid service for retailers and secondary sellers to monitor their entire inventory for recalls. Those plans begin at $29 per month.

The difference between recall data available today and in 2006 is “night and day,” Toney said. But government hasn’t found a happy medium between simply making data available to outside developers and building its own public information portals. There’s also not an easy, standard way for data consumers to ask questions of government data producers so they’re sometimes left guessing what particular data entries mean or whether something is a mistake.

• Make sure you understand the data you’re working with inside and out. Don’t make assumptions about its quality, its history or how it was gathered.
• Know the problem you want to solve before you set out. Don’t just try to build an app around some cool data you found.

Panjiva offers companies with global supply chains intelligence on which international suppliers are most popular among their peers.

Panjiva co-founders Jim Psota and Josh Green wanted to help companies make smarter decisions about global sourcing. Their original plan was to do this through eBay style reviews. When they went looking for official data to validate those reviews, however, they realized they had stumbled upon an even better gauge for which companies are the best suppliers of a given product—the ones other companies are buying from.

Panjiva is built with import and export data from the U.S. Customs and Border Protection bureau, which isn’t available online. Panjiva pays to have that data overnighted on CD-ROMs. It also purchases data from some other U.S. government agencies through resellers and export data from other nations, including China and parts of Latin America.

Panjiva customers pay a monthly fee beginning at $99 to view profiles of product buyers and suppliers around the world, including the most popular suppliers of a given product and which companies buy from them.

The bedrock of Panjiva’s business—U.S. customs data—is only open in the broadest sense of the term. That lack of transparency, however, is also the company’s competitive advantage, because it’s willing to do the hard work of acquiring and crunching through that data when other companies aren’t. One example of the value the company adds is that the U.S. Customs and Border Protection bureau is now a Panjiva customer, effectively buying back its own data in a cleaned up form.

• Start with a problem you want to solve and find the best way to go about it. The data should be secondary.
• Keep it simple. Don’t over-engineer a data solution just because you can.
• Start small and iterate, getting feedback along the way. Otherwise you may spend months or years building something that doesn’t speak to what other people are interested in.