Y Combinator’s Summer 2011 Demo Day: The Definitive Debrief, Part 1
(Page 4 of 4)
Jason Zucchetto, Chris Auer, Mike Shafrir
“Instant financial news.”
When a public company sneezes, it has to report it to the SEC. So jokes the MarketBrief team, which has devised a way to make the treasure trove of SEC documents—a collection that grows by 8.5 million pages a year—more accessible. The company’s software scans the SEC database for new filings on 4,000 public companies; grabs quarterly and annual reports, IPO filings, and other documents the moment they’re published; and transforms them into easy-to-understand news articles. It all happens in under a second, which means MarketBrief can scoop Bloomberg, Thomson Reuters, Dow Jones, and all the other usual suspects in financial news. The company plans to partner with news organizations and also sell the information to investment banks, law firms, and other interested parties.
My take: I want this. Xconomy likes to inform readers when local startups raise money, so I spend part of every weekday scanning and rewriting the information in Form Ds, the SEC disclosures that private companies must file when they sell equity. I’d be elated to turn this part of my job over to software.
Scott Milliken, Ilya Lichtenstein
“Helping advertisers spy on competitors.”
If Company X decides to test the content of their online display ads in the real world before settling on the most effective ad copy or design, as many companies do, then its competitor Company Y could theoretically watch, learn what’s working best, and copy it. MixRank is turning theory into practice. Its software watches display ads running on Google’s AdSense platform across 93,000 sites. “As an advertiser, you can know exactly what’s working in the marketplace, and use that data to make more money, take over traffic, and steal campaigns,” says MixRank co-founder Ilya Lichtenstein. More than 2,200 businesses, including some big ad agencies, are already using the tool, he says.
My take: Disturbingly devious, yet completely legal and, in retrospect, obvious. But if everyone is running around stealing data about each other’s campaigns, who will be left to actually test and run them?
Anand Kulkarni, David Rolnitzky, Philipp Gutheim
“Better than Mechanical Turk.”
Amazon Mechanical Turk is the online crowdsourcing system that lets Amazon and its customers assign tasks such as speech recognition and image recognition that are too hard for computers to a pool of humans. It was revolutionary when it debuted, but Amazon has left a lot of room for improvement, according to the co-founders of MobileWorks. By imposing some “crowd curation” techniques such as analyzing users’ reputations and steering jobs away from users with high error rates, MobileWorks claims it can get higher accuracy rates on crowdsourced jobs. The company, which got its start as project for a technology-and-development course at UC Berkeley, is putting its platform to work in India and Pakistan, which have millions of highly educated but underemployed workers. Today’s Web companies have access to software, storage, and computation as a service, but they don’t have access to human intelligence, the company argues; “Mobile Works will be the place software goes to meet people.”
My take: Mechanical Turk is an amazing platform, but Amazon doesn’t have much incentive to improve it. It’s great to see a dedicated startup like MobileWorks rethinking the concept and applying it in the part of the world where it can do the most good.
Jason McCay, Ben Wyrosdick
“A fast and powerful hosting platform for MongoDB.”
MongoDB, in case you’re not up to speed on the world of databases, is the most popular “noSQL” database these days among Web developers. It’s an open source, non-relational database that’s considered to be cloud-friendly since it can be distributed across many hosts. MongoHQ calls itself a “Database as a Service,” meaning it specializes in running instances of MongoDB on cloud-based servers on behalf of other companies, so that they don’t have to manage their own database software. It’s proving popular; the service is already home to 14,000 databases owned by companies like Gilt Groupe, Second Life, Snapjoy, Grubwithus, and Earbits, and the company says it’s already profitable.
Now Continue to Part 2, including Opez, PageLever, Paperlinks, Parse, Picplum, Proxino, Quartzy, Science Exchange, Snapjoy, Stypi, Tagstand, Verbling, Vidyard, Vimessa, and ZigFu.