Descending Into the Dark Clickstream of the Soul: An Interview With the Founder of Voyurl

There’s a pretty good chance you don’t think about your clickstream too often, that trail of visited URLs that in total paint an entirely unique picture of you as a citizen of the Internet. Early last spring, a somewhat strange web app called Voyurl launched, presenting its users with one mega-clickstream of all of its users, a constantly-updating list of websites visited and by whom (most use screen names). There was also, necessarily, an anonymous option.

It wasn’t scary so much as it was very real and, well, audacious in a time of increasing awareness — and fear — of being watched online; the dawning of the realization that you are being watched all of the time and by many different parties. The OG Voyurl was that laid totally bare.

The site relaunched last week as a bold, potentially revolutionary take on recommendation engines. I talked to Voyurl’s founder, Adam Leibsohn, as he took a break from four nearly sleepless days and nights in front of a computer guiding his creation back into the public.

So is Voyurl finally public now?

I can't get into it out in the open, but there's some reasons we have to stay in beta for now. But it's not too terribly hard to get in. [_Motherboard users can access the private beta here: www.voyurl.com/motherboardtv.

It looks like a way different web app than I remember from last winter.

The whole design aspect of things was a pretty considered process. We did a lot of user feedback actually; we did a bunch of user surveys and got a really good response. From those surveys, we learned three pretty big things. One was that people were looking for and expecting to get out of Voyurl so really decent content recommendations, along the lines of what they hadn't seen before. /Show me more of what I like, that I haven't seen before/. The next was insights and analytics into the [users] themselves. And the last one was some mitigations of privacy concerns.

There was some trepidation about exposing people's clickstream into the open. Understandably so. We were trying that as an experiment to see how people would go with it. We didn't intend for that iteration to be more of a proof-of-concept, but I think that's what it ended up being. Could we build the tech to be stable and scalable; could we get it to act in real time? And could we distribute it and see how people behaved with it?

And so we took that early service and we just sort of bolted on the options we knew that we would like it to do, and we got done and we realized that we'd Frankensteined the whole thing, and we'd sort of hooked everyone up to the firehose. When you logged in, you were faced with that everyone-stream. And what we were finding is that's a really hard issue for folks to deal with. We were basically asking you guys to make sense of all of it yourselves, and that's a lot of work to do. And, two, it negates the context and the temporal nature of things. If you just throw everybody in a stream, it's really hard to derive meaning from it. It just sort of washes over you. For five minutes, it's pretty interesting, but for 40, it could hurt you.

How did evolve into this?

We wanted to figure out how to make better sense of that for everyone and faster. And do it in a way that didn't require any feedback from users or any of that kind of stuff. At the end of the day, we set out with those three insights, privacy, analytics, and recommendations, and we got really simple about it. With a philosophy of that we believe that these days technology platforms and services and apps are asking people to do a lot of heavy lifting for them: you have to tell it what you bought or check in or scan this thing of plus-one or like. Humans have become sensors for their software, and we don't want to do that anymore.

We want something that is feedback free that actually works. Our theory is that technology should be good enough now that it does more for you while asking you to do less. It's the ethos we started out with. You install the plugin and you're good to go for the rest of the time. You don't have to worry about anything after that. And from a privacy standpoint because there's no public stream anymore, the only thing you really publish is your profile.

So how does this work now?

The cool stuff about that is that it's still a really minor exposure. You have control of what goes in there. And we realized that what we'd built without trying is this sort of antisocial social network. Because what we're really interested in is how you behave online and not who you are. We separate your behavior from your identity. You can be totally private and still get the recommendations and analytics. All we want to socialize is behavior; we want to match behavior to like behavior. You don't need to friend or follow anyone because we'll take care of that in the background.

At the same time, we don't require a Twitter or Facebook Connect. We don't need those things. One, we think they can be a little inaccurate. If you have 1,000 friends, it's hard to say that you're tightly connected to all 1,000 of them. And I think it's an interesting problem to solve. We're trying to make sense of the more accurate data, which we think is browsing data, match like to like and recommend as such, within those clusters. You browse the web and we analyze how you behave with things that you like and don't like and we match you up and start recommending things that you haven't seen before and people that behave like you. That's s how that works out.

Some of us thought the mega-clickstream was pretty interesting.

It's tricky, right? I personally wouldn't disagree with you. I think it's really interesting—to people like you and I that want to see that all the time. I do want to poke my head in, take a look, and poke out. At the same time, allowing that ability would I think then would bring out some of those sharing concerns and oversharing concerns. We don't want to scare anybody away. We want this to be a useful service that allows you to get what you want out of it with pretty low impact.

What we'll have to think about moving forward is how to reach that middle ground without alienating anybody. That's really important to us. It's about information; it's about data. It's about your data. And we want to be as honest as possible. If you find that people aren't trusting of the web, they're going to stop using it. And if they stop using it, the apps and products you want to build can't thrive because there's not enough to support them. If we want that ecosystem to work well, people need to start building trust into the products they build. So we would just flip that switch on tomorrow and say "Oh we brought it back." We would risk offending a lot of people that trusted us.

Do you worry that there's just too many recommendation engines out there?

It's considered. And we're definitely thoughtful of it and aware of the other services out there and how they're operating. We believe that we're different. We don't ask you to do anything to get the benefit of that information. We don't ask for interests, we don't ask you to check back in later. You don't have to maintain anything. In that sense, it's sort of the evolution of a recommendation engine. But I think it's also combined with the evolution of the RSS feed. Maybe that's where this is heading, the merging of those things. Our goal isn't to serve you 4,000 links; I don't think anyone can process that. Our goal is to get you a handful of choice morsels that you've never seen.

Other services don't have access to your clickstream so they can recommend you things that you've already seen quite often. A lot of recommendation engines have gotten really good at proving that they know who people are but they don't know what they want, what they haven't seen.

What was some of the worst feedback you got about the mega-clickstream?

There wasn't like a defining issue. You know what I think it was? We really believed that there was a broader awareness to how people are tracked online without their knowledge. I think perhaps we found that it wasn't as broad as we assume. I think people were surprised that this is a behavior that people would want to participate in. And the funny thing was that what we were trying to do was hold a mirror up to that tracking data that we don't get access to and expose it. Not to freak people out.

I don't mean to say freaked people out because it wasn't everyone that was freaked out. It was a few people, but that's alright. We were hoping to open people's eyes and make them aware of these actions taken towards them on the sites that they visit. There was nothing ever scathing. The most it got to was "I just don't think I'm ready to share that much." There was no recoiling in horror or anything like that.

Thank god.

Reach this writer at michaelb@motherboard.tv.

Follow Motherboard on Twitter and Facebook and Tumblr, and get our newsletter here.