If you’re in an industry where you depend on suppliers from across the globe, searching the Web for a part or a collaborator isn’t exactly as easy as trying to figure out where you should eat dinner. Josh Green discovered this around 2005 when he was a Harvard Business School student interning at E Ink, the maker of the screen behind Amazon’s Kindle, and needed to find a supplier for an electronic component.
It wasn’t easy. So he started talking to MIT computer science student Jim Psota, who he had already been working with on putting together a startup. In 2006 they incorporated Panjiva, a Web company that aims to simplify the supplier search process for businesses of all sizes. Their product is out to tell manufacturers and retailers everything from who they can order parts from, to whether or not a potential supplier is stable enough to work with, to what their top competitors are shipping.
“We’ve built a Google-like product, but we’re organizing information specific to the global trading space,” says Psota, the co-founder and chief technology officer, who I spoke with at the company’s Cambridge, MA, headquarters (CEO Green leads the operations in New York).
Panjiva (a play on the name of the supercontinent Pangea) is based on a system of data mining algorithms, machine learning, and natural language processing that culls numbers from more than 10 sources, like U.S. Customs. After Green and Psota found their initial inspiration for the business, they spent a few years developing the product, and raising angel and venture financing. Its backers include Battery Ventures and angels like eBay veteran and Stanford professor Michael Dearing, fashion designer Diane von Furstenberg, and David Frankel of the Founder Collective, Psota says. Panjiva’s Web-based, industry-comprehensive search tool launched in April 2009. Home Depot was its first customer.
The core of the Panjiva product is its search engine, where users can find suppliers by searching for parts, or get more information on a particular supplier they have come across by searching their name. Panjiva’s database has more than 1.5 million companies (including suppliers and customers) from more than 190 countries. Its interface is made to allow users to sift through that deep set of data to find a short list of potential partners. Panjiva customers use drilldown menus based on variables such as country, certifications, shipping date, or the health of the supplier’s business.
“The raw data that we get in is very messy, it provides disparate data points that are not that interesting in and of themselves to helping the user,” explains Psota. “The secret sauce is taking the disparate and messy data sources and boiling them down and cleaning them up.”
Once users get a manageable list of suppliers they could potentially purchase from, they can select a particular company to view a complex profile, which includes information such as
