This month we speak with Gabor Cselle, VP Engineering of innovative start up Xobni. Xobni focusses on helping people extract value from the information generated by how they use email.
Before joining Xobni, Gabor worked at Google and Yahoo! He has a masters degree in computer science from ETH Zurich. What intrigues us about Xobni is that they are helping people gain insight by extracting information from the mass of data in their email inbox (the name Xobni is inbox spelled backwards). In some ways this is similar to Nestoria’s own challenge of organizing masses of listings data. Our recent launch of acccess to property metadata is a similar attempt to gain higher level insights into the database. Gabor, thanks for talking with us 1. Explain the idea behind Xobni, and the power of making information visible. The only thing that has changed about email in the last ten years is that we’re all getting 100x more of it. Email clients are unchanged from a decade ago: You get three panes - folders, your list of emails, and a preview area. This user interface is the small hole in the wall through which peek out at vast seas of information: tens of thousands of archived messages that document our behaviors, buying habits, travel plans, relationships, and social networks. At Xobni, we want to make this data visible, useful, and actionable. We’ve built two products: Xobni Analytics, which shows you statistics about your email usage, and Xobni Insight, which is a sidebar for your email client. First, it offers lightning-fast email search. You start typing, we start searching. It shows you information about the people you communicate with. Imagine that I’m looking at an email from you. It shows me the profile for Ed Freyfogle which contains: |
|
All of this information has been hiding in your email archives. Xobni makes it visible.
2. What are some of the technical challenges you face? First, we’re on the desktop and are dealing with a lot of data. We have users with 250,000 emails in their archives, and we need to be able to show information quickly. We built some proprietary technology which is a database for email information and supports quick querying and aggregation. Second, we have to integrate with existing email clients. Some of these technologies and interfaces have been around for more than a decade. They come with lots of bugs and cruft. Also, as with any desktop application, we have to support a wide variety of configurations. There are almost as many unique Windows configurations as there are Windows machines, and we need to make sure we work on all of them. 3. How might some of these concepts be applicable to the vertical search space, such as property search with Nestoria? Email and property search are very different. We have all the data on the person’s harddrive or webmail account, but you need to reach out and grab it from the web. We can simply order email search results by date or contact popularity, while you need to take into account match for search criteria, physical distance, and result freshness. I use Craigslist for searching for real estate in San Francisco. I’m constantly frustrated by it. The search criteria are city, min and max price, and whether cats or dogs are allowed. This is not that useful if you’re looking for a specific number of bedrooms in a specific neighborhood. You guys kick ass. Gabor, thanks for the vote of confidence. You’re right the volume of information we all have at our disposal is rising rapidly. We share your belief that finding ways to relevantly and quickly present masses of data is a key technical challenge for the future. For those interested in tracking the progress of Xobni as they move towards launch please subscribe to the Xobni blog and/or Gabor’s personal blog past Nestoria interviews: Gregory Marler, Artem Pavlenko, Harvey Edgecombe.
Join the 