Website custom search engines suck, so these guys quit their jobs to try to fix it
Search-within-site engines suck. You know...when you type a word into a site's search bar and no results pop up? Or none of the results that would be useful to you? If you're like me you avoid site search boxes like they've got internet STD's. Case in point, while writing this article I just tried out Pando's search box for the first time ever. It was worthless.
Somehow, individual website search got stuck in the late nineties with AOL and Internet Explorer. Quin Hoxie and Matt Riley are trying to change this with their year and a half old startup Swiftype. It gives website owners an interface to curate their search engine results, and detailed analytics to show what consumers are searching for. It has 70,000 users and growing.
Hoxie and Riley encountered the problem of sucky site search engines while working at Scribd. They were tasked with buying a good internal search engine for the publisher, but there were none to be found. As an alternative, Hoxie and Riley built one from scratch and it didn't take much for them to see the market opportunity. The masses wanted customizable site search bars. Or at least, Scribd did and that was enough to pique Hoxie and Riley's entrepreneurial spirit. After a bit of hemming and hawing, "We quit our jobs as a means of convincing ourselves that we were serious about this," Hoxie says.
The duo applied to Y Combinator, bluffing in their application that they had built a bunch of things they hadn't, and then landed an interview and promptly started freaking out. They booked a hotel room in Mountain View and pulled a two day/night hackathon to create their demo prototype right before the big interview.
Over the last year and a half that prototype has evolved into Swiftype, which is announcing its $1.7 million seed round today from big investors like Andreesen-Horowitz and Kleiner Perkins.
Nowadays, it's a fully-functioning service that lets website creators choose the order of results that appear through their site's search bar. More importantly, these creators get a range of analytics about what their users are searching, broken down into detailed chunks of information like "top unsatisfied queries." Those occur when a reader searches for something that doesn't match any of the site's content.
"Users come and tell you exactly why they're there and what they want," Hoxie says. "You should capture that intelligence so you can use it." In essence, Swiftype aims to give website providers a way to learn about what their readers are interested in and want more of.
Google's got its custom search engine for websites that does a better job of indexing a site's pages, but Hoxie says it doesn't offer curatable results or detailed search analytics like Swiftype's.
Paul Buchheit, creator of Gmail, screened Hoxie and Riley himself during their Y Combinator interview. "He was the scariest person we could have had in there," Hoxie says. "Google is Google and we were effectively presenting a search engine to them." But the story ends well. The duo got accepted to Y Combinator, and Buchheit eventually came on as an angel investor. The backing of a major Google figure bodes well for a startup trying to reinvent search capabilities within sites.
There's other competition out there, although Hoxie thinks they're claiming a different search vertical than Swiftype. Endeca services higher tier companies like Ford or Toshiba. Amazon cloud search requires a technical mind to navigate the code.
Swiftype is trying to go for the little guy in between, who wants a good, curatable search engine with analytics for his business/blog, but doesn't want to pay big bucks for it or spend time on code. When I entered my first name through Pando's search bar, none of my articles come up. Guess we could use Swiftype too...
[Image courtesy Flickr]