Spent some time on the phone today grokking Wondir with its founder, Matthew Koll. Matt has a long and distinguished history in search, starting back in the non-web days (he created a text search engine in the early 90s which he sold to AOL in 1998) and running up into the present.
Wondir is, at its core, a question answering service. Wondir itself is more than two years old, but Koll only recently took the “beta” off the service and turned it into a for-profit enterprise. While there are loads of question answering services on the web, this one is different in some important ways. First off, it feels like a search engine. That’s intentional, Koll told me, as he feels the process of finding answers through chat rooms and usenet like forums is cumbersome and unintuitive. Secondly, Wondir aggregates questions and answers through the architecture of participation, essentially getting its questioners to become answerers, and vice versa. This is non trivial – getting people to answer questions is not as easy as it might seem. But Koll has thought through all of this, and I like where this service is going.
You don’t have to register to ask a question, but it pays if you do, because then your answer can be sent to you (and you can also tell Wondir areas of your own expertise, and it will notify you of questions that come in that you can answer if you wish to). When you do ask a question (in plain english), Wondir does a number of clever things. First, it parses your question’s text and categorizes it in any number of potential topic clusters. It then alerts registered users who have raised their hands as willing to answer questions in those topics, either through email, RSS (soon), or IM. It also posts the question right there on the service, in a scrolling ticker below the search box. Wondir has any number of categories in a pull down menu, and when you select a category, the questions scrolling across the bottom change as well (the questions in the “mature content” area are a hoot).
Now, that alone is not enough to get this service to scale, and Koll knows it. So he’s done a few more neat tricks. First, he’s cut vertical content site deals, distributing Wondir out into the web in areas where the expertise and the community lives, complete with the question ticker. For example, there’s a Wondir question/answer service at ichef.com, ratemyteachers.com, and icerocket (that new engine backed by Mark Cuban). Those more tightly integrated affiliates create scale and databases of questions and answers, databases that are then folded back into Wondir’s overall engine, meaning that the more questions that are asked and answered, the better the overall engine gets. Neat, huh?
It gets better, at least theoretically. Koll has also cut a deal with Six Apart for a Typepad implementation, which will allow bloggers to share ecosystems of question answering. So, for example, Danny, Gary, Andy, and I might have a Search-related Wondir implementation. Eventually, we’d be able to share revenue in that model as well.
Revenue? So what is the business model? Well, it’s paid search, of course. That’s the beauty of it. A site like Wondir, or its affiliates, is very intent driven, and very specific, making AdSense a natural fit. That might answer the major question I have about the service – why, beyond good kharma and self promotion, would anyone want to get in the habit of answering questions for free? (Although, I’m not so sure that being helpful isn’t in our human nature to begin with, and it’s cool to have a service like Wondir to test that theory. )
Of course, Google Answers has been around for a long time, but as Koll points out, you have to pay for those answers, and the business has a limited architecture of participation. Koll claims it’s doing only hundreds of questions a day, and Wondir, while still pretty much in stealth mode, is doing thousands.
Koll told me he wants to get the word out on Wondir, and hopes we’ll all bang on it and help him make it better. I for one plan on using it for a while, and I certainly hope the service hits a tipping point. The implications are pretty darn cool.
Bonus link: Chris Sherman did a nice write up of Wondir back in 2002.