NextWeb – Digg wants to predict better

NextWeb live blog:
Keynote Discussion: Scott Rafer & Kevin Rose
“Inside look at Serial Entrepreneurship”

[Health warning: Live notes, excuse typos, much of this is paraphrased quotes]


The nice thing about starting in the early days was not having too much traffic. Not worrying about “features”. Now it takes a long time to get stuff done. We have a huge dept to check out code, issues, etc. A lot harder than getting new things live on the site.

Where we’re taking the product is acting as a collaborative filter for all content. Behind the scenes we have a new hire in Anton Kast, our first Math PHD, and his main goal is to improve the algorithm and work in ways we can share stories with people by comparing diggs and looking at what happens in real time. We’re working on omething called Digg Stories. It’s about what else are people digging that you might find interesting. We want to make these connections and introduce them to people before they get big, like on the front page.

One of the issues we have is that the upcoming stories section is broken. The way people browse to get stories on the home page does not scale. So behind the scenes you have to get smart about how to get stories on the home page. We have to determine which stories are likely to get popular. The more we can do that the better we get at recommending a story before it gets popular, because it has 6 diggs from people you are connected to / are like etc. [paraphrasing here – ed]

We’re also working on a new comments system.

Question: Pownce versusTwitter?

Kevin: Pownce is an amazing platform for sharing media with your friends and having a conversation around that media. That’s what it excels at. Our application uses standard ports, not like sending files over IM – that’s the niche it’s filling.

How many shows are you planning to do with Revision3?
Kevin: It’s all about high quality content for the Web, originally created by people on a forgotten cable channel. So we took that to the web and took the time constraints off the content. We don’t care where the content goes, iPhone, desktop etc. As long as we can count the views, thats all the advertisers care about.

We definitely have a look at the best upcoming content, like cool new shows. If you look at video in the Bay Area, it’s not like we are competing – we cross promote and plug each-other’s shows.

Q&A section

Would you sell Revision3?
Kevin: With all these companies, it didn;t start with an idea about a sale. Like Digg was started before people started calling things “Web 2.0”. I was working in TV at the time. I wanted to create a company that could pay my rent, that was all. So now I still have fun and have the best job in the world. The only way acquisition makes sense is if two companies can create something bigger and better. We talk to lots of companies all the time but we don’t have any hard plans or dates [for a sale].

Lots of firms want to hook into your underlying technology?
Kevin: We’re going to be everywhere, we’re going to be in as many platforms and places as we can. I want it so that you’re on a news article and you click and you’re done, that’s it. We joined the data portability group and now support hcard – so we’re absolutely behind moving your data and social graph around. Always been creative commons , public domain.

How do you plan to monetise data?
Kevni: We’re not a very bandwidth intensive application. Digg is a tool for the collective wisdom of the masses and what they are producing. I would rather lose and ad and make the experience easier than lose people. We don’t need to take more money, we’re on track to be profitable. We just want to be the collective place where the masses unearth stuff they like. We don’t have to be an amazingly profitable company, just want to to be the place where all that happens.

Scott: The way the world’s going online, remaining dominant is more important that revenue times pages.

Kevin: There’s 26m people coming to the site per month, but only 2.5 million are active – so we want more people to participate.

What about Digg clone startups – do you look at their features, or like Propeller, or Yahoo Buzz. What competitors worry you the most?
Kevin: The fact that Yahoo Buzz is in the space makes me really happy. It validates us. Look at clones, they might be great but we have a huge head start. What would worry me would be something we haven’t seen yet.

Who are most authoritative diggers? Making that visible would create an amazing reputation system, no?
Kevin: I haven’t thought about that a lot but have thought about recommendations and recommending users. Such as finding the the best user who finds the best LOLcats stories. It’s going to get really interesting.

How much gaming of Digg is there?
Kevin: Lots – some people have hired hundreds of people to do this. We have a team figuring out who is gaining the site.

Scott: Yahoo Buzz will get worse – there is lots of money to be made from hitting the front pages of these kinds of sites.

Kevin: What we’ll push in the future is category areas on the site, such as the best political stories. Creating a universal home page that scales to everyone’s interests is very hard to do. We won’t always appeal to everyone on the front page. The best meat in the site is in the categories.

Advice for startups?
Kevin: Make sure it’s something you’re passionate about. Not who you’re going to sell it to.

Scott: It’s very easy to put together a logical plan for a startup and a bunch of VCs. But unless your day job gets you out of bed then you’re not going to perform.

Could you have done Digg earlier?
Kevin: I wouldn’t have been able to do this in the 90s – we relied on open source and rented servers etc. Now it’s hard for different reason because there is so much noise. I’m looking forward to the next year when there’s a shakeout and the interesting companies will come through.

DVD key issue:
[Didn’t get all this] Rose backed his decision to post the DVD key to Digg