I have tracked Ian Clarke for several years now and was very excited to have the opportunity to interview him. He founded FreeNet, a p2p network that enables people living in non-free countries (like China, Cuba, etc) to get secure access to the Internet. He has helped the Skype guys with their latest project Joost, co-founded Revver and is now running his own startup called Thoof. Thoof is a generic targeting engine designed to accurately provide predictions about users like Amazon’s “customers also bought” page. That in itself is nothing new, but Thoof is designed around providing accurate targeting for very small datasets. Ian’s technology can be applied in a many areas online including behavioural targeting and personalization.
Full Interview Audio and Transcript
Hobbies and Interests: Going for Long Walks, Working on Free Software, Political Activism: Free Speech Online, Anti-Software Patents, etc.
Favourite Sports Teams: Not a Huge Sports Fan but Does Follow the Irish Rugby Team.
- The Age of Turbulence by Alan Greenspan
- The God Delusion by Richard Dawkins
- The Four Steps to the Epiphany by Steven Gary Blank
- The Reagan Diaries by Ronald Reagan
- Palestine: Peace Not Apartheid by Jimmy Carter
Favourite Entrepreneurs: Niklas Zennstrom and Janus Friis.
Company Website: http://www.thoof.com
Adrian Bye: Today, I am here with Ian Clarke. Ian has done a number of interesting things in the past particularly around file sharing and founding a popular video site. Ian’s here to tell us a little bit about that and also his latest start up.
Ian Clarke: I’m an engineer by trade. I have a degree in Artificial Intelligence in Computer Science from the University of Edinburgh in Scotland. The initial project that really set my career trajectory was called Freenet, which grew out of some work I did at the university on decentralized systems.
Freenet is a 501(c) 3 non-profit, and the intent is to allow people in countries like China and Saudi Arabia to exchange information without fear of censorship by their government. Freenet has been an ongoing side project for me for the last seven years.
In parallel with it, I’ve started several commercial ventures. The first was called Uprizer, which developed peer-to-peer content distribution technology. I left Uprizer in 2002 and set up a consulting company called Cematics. Probably, the best known project that Cematics worked on was a peer-to-peer video distribution technology that formed part of a piece of software called Joost. We handed that off in mid-2006 to the team that then became Joost.
I had always been very interested in the concept of how to allow people to monetize their creative efforts online. Together with two co-founders, I started a company called Revver. Revver was the first online video company to attach advertising to the end of videos and share the revenue with content creators. Revver has raised, so far, $10 million in A and B-round financing.
I moved on from Revver in December of 2006 to start my current venture, which is called Thoof. The idea of Thoof is to figure out what people want, give it to them, and offer that capability as a service to third parties.
Adrian Bye: Tell us a bit about Revver. As I understand it, it’s basically a YouTube clone with people posting videos and getting a revenue share of the advertising money generated from their video.
Ian Clarke: It is similar to YouTube in that it’s part of the latest crop of video sharing web sites. We weren’t aware of YouTube when we started working on Revver in mid-2005. We found that Revver is the web site you go to if you actually want to get paid. We were able to attract semi-professional video creators. The problem that Revver always had was that this stuff would wind up on YouTube anyway because of YouTube’s lax enforcement of copyright. It wasn’t quite as big a differentiator as we originally expected it to be.
Adrian Bye: Why did you leave that company? It’s still a fairly successful company, is it not?
Ian Clarke: My skill is in solving interesting and hard technology problems. Revver was turning from the type of company where it has technology problems into the type of company where it has business challenges that needed to be addressed. I felt that wasn’t really my core competency. While at Revver, I’d also developed an idea I felt was extremely compelling in addressing a pretty fundamental problem and I wanted to pursue it.
Adrian Bye: I’d like to really understand what you did with Freenet and why it came about and what it’s doing today.
Ian Clarke: While studying Artificial Intelligence and Computer Science, I was very interested in a field called emergent systems where you have a bunch of individually simple components but when they interact with each other they exhibit sophisticated, often surprising behavior.
I was really keen to find some kind of practical way to apply this concept. As people were living in this dreamworld where the Internet would be the answer to the ultimate tool for freedom of speech, I really started to worry that actually it could be the opposite and a means for control. I started to think of a way to layer a technology on top of the Internet that would allow people to communicate freely and anonymously without fear of censorship.
You can surf Freenet just as you would surf the World Wide Web, and you can actually use an ordinary web browser to do it. It is slower than the World Wide Web because there’s a lot of cryptography and your computer is talking directly to other computers. You can use HTTP over Freenet, and people have experimented with SMTP or the e-mail protocol, and others as well. We don’t guarantee that you are 100 percent secure, but it’s safer than the alternatives.
Adrian Bye: Tell us about the start up you’re doing now and the problems you’re working on solving today.
Ian Clarke: The basic idea behind Thoof is to quickly figure out what people are interested in and present them with things that will appeal to them. I did some work on this type of thing at Revver where we were interested in figuring out what kind of videos interested people and then presented those videos to them. I actually developed two collaborative filters for Revver, which is kind of the classic way to solve this problem. One we licensed to a company called Reddit, which was later purchased by Condé Nast.
I realized while working on this problem at Revver that collaborative filters were not a good solution. That made me realize there was a real opportunity to create a type of recommendation engine that addressed what I perceive to be the key problems with collaborative filters.
Adrian Bye: If collaborative filtering isn’t the best way, what is?
Ian Clarke: Collaborative filters require a lot of data about users before they can effectively make recommendations to those users. Typically, collaborative filters will do one of two things. They’ll use a user-based collaborative filter, which works by identifying similar users and then looking at what those users like that you haven’t seen yet and recommending those things to you. That requires a lot of data before it can effectively find similar users. The other kind of collaborative filter will assign properties to the users.
The difference with Thoof’s approach is that it’s able to figure out what your interest is based on relatively little data, such as the type of data a web site will have about a user right from the first moment a user visits or very shortly afterward.
It’s like a meat-grinder algorithm where you can give it whatever information is available to you about users. The algorithm will spot patterns in user behavior and use those patterns to come up with an initial picture of the user’s interests. As that user continues to use your web site, you can feed in additional behavior you collect about this user that will refine the systems idea of that user’s interests.
Adrian Bye: Is this predictive analysis or something totally different?
Ian Clarke: It’s related to predictive analysis. Basically, it’s an inference engine of sorts. It looks for correlations in user behavior and combines that with an ontology of information about the things that are being recommended. It uses the knowledge it has about metadata and how the metadata is associated with other metadata in order to very quickly come up with theories about your interests.
When applied to advertising, our approach is similar to behavioral targeting. We assume that we may have very little information about a user, and our technology is optimized to take that small amount of information and turn it into a useful profile of what this user most likely will find interesting.
In the case of product recommendation, the metadata that you would have is about the product, such as its color, its price, its manufacturer, and other relevant characteristics. You would give this information to our algorithm, and it would actually be able to use that metadata instead of just looking at the product.
Adrian Bye: If I were to license your technology, how does that work?
Ian Clarke: We have a simple over-the-network API, with which we’ve worked hard to make it very easy to integrate. You sign our license agreement, and we give you access to this API. It is very easy to integrate with almost any platform, including Java, PHP, and Ruby on Rails.. Very simply, you send us whatever information you have about what you’re recommending to users. We will send you back a list of the things that user is most likely to find interesting.
Adrian Bye: Well, thanks for making time, this was an excellent interview.
Ian Clarke: Thank you.