g.ho.st
ClickForensics

Dispatch From the Web 2.0 Launch Pad

Next Story

AT&T quietly launches Sierra 881 3G PCMCIA card

launchpad.jpegToday’ Web 2.0 Summit ended with a Launch Pad session where six startups each got six minutes to pitch their companies to the crowd and a panel of venture capitalists. Here’s a thumbnail sketch of each with my initial impressions (For a more thorough take on these startups from a real venture capitalist, read Christine Herron’s post):

CleverSet—Best of Show went to CleverSet, a Seattle-based company that takes a sophisticated statistical approach to product recommendations and personalization. This is not exactly an unknown company. It’s technology already powers 85 sites, including Sephora’s, Wine Enthusiast, and part of Overstock (I also wrote about them last summer in Business 2.0). CleverSet is applying some advanced math to improving recommendations, and claims to increase revenues for Websites that implement its technology by 18 to 30 percent, on average. If that’s true, they deserve to win. But then I ran into Netflix CEO Reed Hastings, who is offering a $1 million prize to anyone who can improve his movie recommendations, and he expressed some skepticism about how useful any statistical approach can be. Hastings has found that even within just the category of movies, knowing what horror films someone likes tells you nothing about what dramas they might like. So making statistical correlations across products would be even more difficult.

TripIt—A company that presented at TechCrunch40, TripIt builds a personalized itinerary starting from your airline confirmation. A useful travel organizer. See Mike’s previous post.

G.ho.st—All of our data and applications are moving online, why not the operating system? G.ho.st is a Web operating system of sorts that ties together all the data and applications you may be using across different Websites with one password and URL. Conceptually, I’m with them. But getting people to change their behavior and abandon everything on their desktops except for their browser is going to be tough. (G.ho.st was in the TechCrunch40 Demo Pit)

SpiceWorks—Ad-supported enterprise software. Already 160,000 IT professionals use SpiceWorks to help manage their computer networks. SpiceWorks then serves up news feeds and product deals targeted at the specific devices on the networks they manage. It’s a consumer approach to enterprise software. This will work—until the ad bubble pops.

ClickForensics—The CEO claims that the click fraud rate is nearly 16 percent (and over 25 percent on distributed advertising networks like AdSense or Yahoo Publishers Network). ClickForensics offers a neutral service to both advertisers and publishers that audits the quality of the click traffic generated by any given ad campaign. This is a community approach to solving a growing problem, although some argue that the click fraud rate is already priced into what advertisers are willing to pay per click, so it is already being taken care of by the markets.

Realius—Combine casual gaming and real estate porn and you get Realius. The fantasy real estate site, which will launch in beta in two weeks (and was also in the TechCrunch40 Demo Pit), will take real listings and let people guess how much each house is worth (using a slider that shows where other people have voted). Revenues will supposedly come from advertising, referral fees, and service fees from brokers who can use the game for training purposes. The game is based on actual real estate data. The CEO lost me, though, when he said that you don’t find out if your guess was right until later when they send you an e-mail (which is designed to drive you back to the site). Any game that does not generate instant feedback on how you’ve done is dead in the water, IMHO. Check your e-mail to see if you’ve won! That’s going straight to the junk folder.

Other startups that didn’t quite make the short list include Castfire, Kango, Footnote, Lemonade, Search-to-Phone, WooMe (another TC40 company), Sprigley, and GoXDML.

blog comments powered by Disqus