Add to iTunesBret Taylor & Paul Buchheit - Co-Founders, FriendFeed
Introduction
I’ve increasingly become interested in the role feeds are playing in social networks. FriendFeed is a new Silicon Valley start up that is being run by a team of very successful former Google ‘intraprenuers.’ Bret Taylor and Paul Buchheit, two of the four co-founders, sit down to talk about FriendFeed on this week’s episode of Read/WriteTalk.
|
|
|
Interestingly, after our interview Paul agreed to join me on a panel at the upcoming Graphing Social Patterns conference on “Social Networks and The Need for Feeds.” I’d love to tap into the collective wisdom of this crowd, let me know questions you’d be interested in seeing us discuss.
Links
- FriendFeed: About Us
- Post Referenced on FriendFeed’s Usage
- Examining Feeds in Social Networks (RWW Post)
- Social Networks & the NEED for FEEDS (Graphing Social Patterns)
Transcript
| Sean Ammirati: This is Sean Ammirati with ReadWriteTalk. Today I have Bret Taylor and Paul Buchheit from FriendFeed. So just to start out guys, one of you could just give our listeners who aren’t familiar FriendFeed, a quick overview of the service.
Bret Taylor: Yes, thanks for talking to us today. FriendFeed is a tool that lets you see what your friends and family and co-workers are up to on the web. You could see an aggregated news feed all of the blog post that they share from Google Reader, the photos that they’ve published on Flickr, the Twitter messages that they’ve posted and a variety of other sharing activity aggregated from the most popular social websites on the web and through the integrated feed.And then it promotes Water Cooler discussed around those items. So everyone who’s subscribed to me on FriendFeed can comment on and favorite individual |
|
| 01:21 | Sean Ammirati: So the discussion is happening on the FriendFeed site itself, the Water Cooler discussion as you put it?
Bret Taylor: Yes, that’s true. And because the way FriendFeed currently works is that people can comment on my items if they’re subscribed to me more or less. So the discussion if I for example go on Google Reader and read a ReadWriteWeb article about FriendFeed and I share it, the discussion about that article on FriendFeed will be limited to people who are subscribed to me.And friends of friends and standard social networks privately controlled. So for a given shared item, say like a newspaper article, there actually be lots of different discussions in different parts of FriendFeed based on different social groups subscribed in the same article. |
| 02:15 | The discussion is on FriendFeed but what makes it distinct from say like the comments on ReadWriteWeb is that it’s isolated to different social groups. So it’s much more people who actually know each other, more of a water cooler discussion as opposed to public discourse.
Sean Ammirati: Ok, that’s a great overview. So let’s go back to your time, for people who aren’t familiar with you guys’ background, both of you are very successful I guess intrapreneurs at Google with Google Maps and Gmail. And then you are entrepreneurs in residence at Benchmark. And I understand the original business you’re working on at Benchmark was not what you ended up leaving Benchmark and starting. So talk about why you switched. What was so compelling in your opinion about FriendFeed. |
| 03:02 | Bret Taylor: Sure. Just to give you a little more, there such a background with a little more subtle than that. Paul actually left Google earlier. And I let him independently talk about when he left and why he chose to come to FriendFeed. I actually left Google with a long time friend of mine named Jim Norris. We actually lived together in college and have worked together in pretty much everything for the past 10 years.He came to Google and worked on Google Maps with me as well. And then we ended up joining forces independently with Paul and Sanjeev much later in the process. So our background was I worked at Google. And the primary things I worked on was Google Maps and Google developer programs. So the Google Map API started it but all of the other APIs and developer focused tools that Google released after that. I ended up like leaving Google in June of 2007 with Jim and we became entrepreneurs in residents at benchmark. |
| 04:06 | We’re originally working on some storage technology. And I ended up creating a very early prototype of FriendFeed on the side. We had been getting feedback and ideas from Paul and Sanjeev who are just at that time are doing a lot of angel investing and advising to start-ups. And I think the prototype of the side project was a lot more interesting to us than what we are building. And we ended up scrapping it and decided to join forces and work on it together full-time.
Sean Ammirati: Cool! And Paul could you just give your background. I actually had missed that from some of the other coverage on you guys. Paul Buchheit: Yes. That’s right. I left Google in May 2006. So I just took sometime off and got involved in a number of different start-ups doing investing and advising. And in general, this time learning more about start-ups and running businesses as well. |
| 05:08 | And then earlier last year Sanjeev left Google. Sanjeev was the second person on Gmail and he has taken over the project when I left. And so we started talking ideas for a new company and we started working on a few ideas and also just talking to other people coming out of Google. Such as Bret and Jim.We discovered that they had a lot of the same values and ideas and hopes in terms of the kind of company we want to build. Similar ideas in terms of the product and everything else. And just decided that it will work best if we just all joined together to form one company basically which was FriendFeed which Bret basically started working on as side project over the summer. It’s a very compelling product at least to me. Immediately it just makes sense and it’s simple. So we were excited to do that. |
| 06:18 | Sean Ammirati: Cool! Going back to the initial thing you were working on over the summer. Just from using it a little bit I’m guessing the Facebook news feed was part of how you got inspired for this. Maybe talk a little bit about how did you come up with the idea?
Bret Taylor: Yes, absolutely. The feed from news feed was certainly an inspiration for a lot of what we did. I thought when it first came out in particular just the act of being able to see when people I knew publish photos was really interesting to me. It’s really integrated with the Facebook experience really well. |
| 07:03 | But my main issue that actually that I had a lot of family members use a lot of different photo sites. So the initial thing I did was just something to aggregate for my extended family and friends. So that’s actually a photo publication notifier. And then I ended up because a lot people I knew use sites like Google Reader and Digg and those sites supported some combination of Atom and RSS feeds and custom-made APIs.I started integrating other services like you just get notified when people I know did things around the whole internet. I immediately started relying on this daily just to see what was going on in the world. The main reason for that is like there’s just so much information published everyday and the news source that I rely upon are so diverse now. It’s almost impossible for me read like every post in every blog that I subscribed to or watch all the videos on the front page of YouTube. |
| 08:02 | So I just created a simple filter of what the people I know, what are they actually reading, what are they actually watching. It was one of the most effective ways of just sifting the amount of information published everyday. And I think that concept was really compelling to all of us.Just how do we generalized this? How can we use your social network as a collaborative filtering mechanism. It’s very personalized that you can find the most interesting blog post of the day or the most interesting videos of the day or the most interesting new music. And the act of finding social solutions to information and problems was really a compelling concept for us.
Sean Ammirati: Interesting. So you talked about some of these via APIs and some via RSS and Atom. I actually did want to ask you about that. I think you have about 30 services now that you can pull in and obviously I think RSS and Atom is one of them. But I’m assuming a lot of the other 30 services are really just RSS feeds as well. How does that breakdown? |
| 09:01 | Bret Taylor: I don’t like remember onthe top of my head.
Sean Ammirati: Just shape of the data. Paul Buchheit: Yes, I would say probably somewhere around 70% are feed-based. And probably 25% or so are customed XML-based APIs. And then probably another 5% like just a lot of work on our end. [Laughter] Sean Ammirati: Got it. Bret Taylor: So for example, we aggregate Google Talk status messages like Tweeter I guess. You change your Google Talk status message, it will show up as an entry in FriendFeed. And that uses the Jabber protocol for example which is obviously an open protocol but integrating it is a little bit more work than just crawling Atom and RSS feed. And we also have done a fair amount of work to make the user experience right. Pulling out the right thumbnails then cleaning up the titles of things just so you get that nice integrated experience. |
| 10:03 | That if you’ve ever subscribed to an RSS feed from say a photo site. If someone publishes his album of photos you’ll get 250 entries in your RSS reader which is not really a great user experience. We do a lot of stuff. I do smart clustering of related items. Or if someone Twitters 2,000 things in an hour, you’ll just get one cluster of entries that doesn’t take out too much of your feed to try increase the signal to noise ratio and then you want individual FriendFeed.
Sean Ammirati: That’s very interesting! Bret Taylor: In many ways, FriendFeed is really riding the trend of people supporting open syndication format and APIs and it’s something we’re really appreciative of. But there’s a time out of work to make that really genuinely useful where it doesn’t feel like you’re getting the output of a dumb machine. You actually do something useful so there is a fare amount of work to make that the right experience. |
| 11:03 | Sean Ammirati: Yes, I mean how are you doing that clustering of like similar photos? I guess I wasn’t aware of that as a user of the product but kind of interesting.
Bret Taylor: It’s not really rocket science right now. I mean for photos it’s mainly based on time and when you publish an album of individual photos. It tends to happen within a relatively short period of time. So it’s fairly easy to detect when things are part of a conceptual group.And then for other things, if someone uses any particular service really, really frequently we’ll try to cluster things together just so that your front page of FriendFeed have a good diversity of people and different services. They are not overwhelmed just because one of your friends happen to use one service really, really often. |
| 12:00 | Sean Ammirati: Yes and I think Twitter is a good example of that.
Bret Taylor: When we launched, we want a set-up on our blog and a histogram of the services people use. But we’re actually to update that. Obviously our user base had run since then. What’s funny is that Flickr I think at that time is the most popular service among FriendFeed users.A very large percentage of FriendFeed users use Flickr. But by far Twitter was the most common individual entry. Twitter users use Twitter incredibly often. It’s really amusing. The statistics of how people use these services is really interesting to me. Sean Ammirati: Yes and I actually saw that post. That was fascinating. But I think if you use Twitter and have any type of network in it, it’s obvious how voluminous some people can be. Bret Taylor: Yes. Absolutely, absolutely. Sean Ammirati: So either one of you can take this. But I’m curious, who would you deal as your competitors? Paul Buchheit: Well there’s a lot of product out there that are kind of superficially similar. It seems almost like there’s another one every week. And so at some level it seems like we have a lot of competitors. But at the same time there aren’t really any products that have quite the same user experience. |
| 13:17 | And I think that’s something that we all learn at Google and that a lot of times it’s the subtle details can make a big difference. And so there’s a lot of products that are focused more on the just the aggregation aspect of it. But we’re really trying to go beyond simply aggregating to actually creating a pleasant social experience around with the content.So far there haven’t really been any other companies doing that. For example, Bret was talking about the comments and the Water Cooler discussion aspect of that. And the surprising thing about that to me once we realized what was going on is just that, that experience almost doesn’t exist on the web today. |
| 14:02 | Typically where will you have comments. The comments are from these all kinds of random people. And they tend to have a certain flavor to them as a result. The worst case is like the YouTube comments or something like that. But it’s a very different experience when it’s just your friends or maybe your friends of friends commenting on something. It’s actually to me it is a lot more interesting and enjoyable because it’s a different kind of dynamic. And it’s a lot friendlier.
Bret Taylor: Just to give you an example of this. We recently launched this feature so that if you’re subscribed to someone like Paul and Paul comments on one someone’s entry that you’re not subscribed to, that entry will show up on your feed with Paul’s comment highlighted.So basically you’ll see entries from friends of friends that are interesting enough to warrant in a comment from someone you know. And that feature is actually that incredibly popular among our user-based because it broadens the discussion by one level in the social network. |
| 15:08 | But it also just means that your FriendFeed is filled with like the most interesting discussions and the most interesting news items from your friends and your friends of friends. And so for example when we launched that feed I posted a blog, posted on the FriendFeed blog and there was like a really big discussion from users. My people subscribed to me and their friends about ways we can approved the features writing context. And it’s really interesting to have not just the aggregation of the content but that semi-private discussion about that comment, at that content. And I think that’s the dynamic. There’s a bunch of blog post by FriendFeed users talking about it as the main place they hang out online and I think that’s the experience that we’re trying to create. Sean Ammirati: Yes, that’s great! The people I that I was assuming you would say your competitors I don’t think have that. |
| 16:04 | Sean Ammirati: The Twitter, Pounce, Tumblr where you can put on feeds. One I guess it’s interesting is how do you view Google Open Social? We had Kevin Marks on ReadWriteTalk not long ago and talked about activity streams. I mean in some ways they’re trying to aggregate social feeds and let networks do some things on top of it as well.
Bret Taylor: Yes it’s true. I’m not an expert in like all of the stuff that changed in Open Social since launched. I haven’tbeen following it extremely closely so please correct me if I’m wrong.But I think the main Open Source activity stream API was meant as a way for applications within an Open Social container to publish events.So that essentially within a social network you could find out that so and so did something on Scabulous or did something else with widget apps. So you get the viral distribution of widgets within a social network. |
| 7:08 | As far as I read anyway, there hasn’t been a lot about aggregating stuff across social sites that much. And much say happen to implement a widget in Open Social and what not. And so I think that there’s a lot of people thinking about similar problems.I think there are some not so subtle product distinctions between what we’re doing and what Open Social is doing which is I think more or less a standard for creating widgets for social networks. And then ability for those widgets to publish activities so that the widgets can spread virally within the social networks which is all being arguably a fairly fundamentally different problem that we’re trying to solve.
Paul Buchheit: I think they are fairly complimentary offerings really. We already have a Facebook application for FriendFeed that brings your FriendFeed into your Facebook profile and also within the Facebook UI shows you your friends’ FriendFeed and your friends on Facebook. |
| 18:14 | And that’s actually one of the very popular aspects of FriendFeed. And so we’re not just trying to pull everything into FriendFeed. We’re pushing it back out to all of these other sites as well with the Facebook app and there’s an iGoogle wigdet and a blog widget. It’s just a matter of time that as Open Social becomes a little more established we’ll definitely have an Open Social widget as well so that we can appear on the various Open Social pages and hopefully publish it to their event streams as well.
Bret Taylor: You can think about a lot our users are using it as a way of connecting their Facebook profile to all of the other products they use around the web. So when you add the FriendFeed’s Facebook application and you publish some photos on Flickr, it will show up in your Facebook news feed and show up on your profiles. |
| 19:11 | And if you bookmark things on Del.icio.us or share something on Google Reader it will show up on your profile. So it’s a really one-step way of connecting your profile to the products that you use that don’t happen to be a part of Facebook.
Sean Ammirati: Yes and actually I use it on my Facebook. It’s actually great! The one thing you got to be conscious of is I had both the Twitter widget directly and then the FriendFeed and my Twitter username in there. Bret Taylor: Yes, we’ve contact Jim and we definitely are very aware of that problem. Because of how popular aggregation services are nowadays and Facebook applications. You end up with this weird interceptions and duplication issues. And I think we’re aware of that. We haven’t got a good solution |
| 20:00 | Sean Ammirati: My friends just called rainman because I said everything twice for a little while.
[Laughter] Sean Ammirati: It has worked out fine. I just took the Twitter updates out of the feed because they were coming through FriendFeed anyways. But it is actually a great service for people. Actually we’ve touched on some of the UI challenges a couple of times through. But Paul you had an interesting post on the FriendFeed blog where you announced Kevin Fox had joined your team.For you who aren’t familiar, I don’t know his exact role at Google. I think within the blogosphere, he was considered at least the leader of the Google user experience team or at least a lot of the teams within there. I thought that was amazing that he was joining your team. What do you guys look at as the biggest interaction challenges for FriendFeed moving forward? Bret Taylor: Well there’s a number of fairly significant challenges. One of the first things we’re really wanting to focus on is making this into a product that really appeals to a fairly broad audience. We don’t want to become something that’s only for people who use 10 different social services or something like that. Because I think that’s a fairly small group. |
| 21:21 | And so we really want to turn this into something that all kinds of people are interested in and enjoy using. And a lot of that just comes down to how we present ourselves in it. A lot of our most active users today actually don’t even use any of the other services. They don’t use Twitter or Flickr, any of those things.They just come to view what their friends are sharing and to comment on that. And so we already have our products that works for other people but we don’t communicate that very well at this point. And the interface doesn’t communicate that because it sends a different message. And so one of our biggest goals is just to make it so that it is appealing to more normal people. |
| 22:07 | And then beyond that is obviously information overload issue as well as we pull more and more information in. It’s a challenge just to summarize all of that in a way that you don’t feel like you’re just being buried in data. And so into that kind of another big focus for us.
Sean Ammirati: Yes. I think the filtering is an interesting challenge for sure. Just to take it the other direction there, too. A couple of times in an interview, you’ve talked about it as, I think at one point you called it “social collaborate filtering”. Is there any plans to do discovery as well. Because if this becomes people’s filter for information there’s a little bit of a risk for self reinforcing…you keep looking at more and more the same things. I think in the blogosphere, it’s called the “echo chamber” sometimes, that kind of thing. |
| 23:01 | Paul Buchheit: Right. Yes. I think we definitely want to do some things to bring in a little bit of related contact. I think somewhat where you’re making things a little too magical so that users don’t really know what’s going on. I think people want to have control of what they’re seeing in their FriendFeed. And so we want to make sure we don’t haven’t harmed that aspect of it. But bringing in more content is definitely, can make it more valuable if done properly.
Bret Taylor: I think that will also help some of the information overload problems as well. Our system have a greater understanding of the content in your feed so that you can present it in the most concise and least intimidating way. But also as you mentioned bringing in content that’s not nested. It’s more about discovery than what your present directly shows. Sean Ammirati: I’m sure that you saw that Om Malik or whoever from Google and wrote the piece. They ended with a little bit of a starkly comment which is and at some point the little project might have to make some money. So I was curious if you want to respond to that. I’m talking about how your planning to make some money. |
| 24:18 | Paul Buchheit: Fortunately, we are having some time to work on that problem. So we can really focus on the user experience to make it useful right now. But I think obviously in the longer term we do need to make money. These obvious things are more towards advertising. And given that word we have the information about people’s activities and so forth right there in the feed. I think it’s actually potentially a pretty useful channel for advertising. And then there’s some other possibilities. |
| 25:04 | Bret Taylor: Because we’re a relatively new service, we’re really focusing on creating our user experience and create our user-base first. But as Paul mentioned I think we obviously will try advertising approaches initially. It’s a relatively new type of service so I don’t think there’s too many examples right now to figure out how well advertising will work. So we have a number of other ideas and try after that.
Sean Ammirati: I realize you do have some time. Have you announced any funding or is that still under wraps? Bret Taylor: Unfortunately. we’re in the process of finishing that process so we can’t really talk about it. Sean Ammirati: Sure, that’s fair. Ok. I think it’s a great service. And I think it illustrates some interesting things about feeds and sort of the activity you can build around them. I really appreciate both of you guys taking some time to talk to us today. And we’ll have a post up on ReadWriteWeb and ReadWriteTalk. But thanks for joining me today. Bret Taylor: Thanks for taking the time to talk to us. Sean Ammirati: All right. Take care guys! Paul Buchheit: All right. Bye. |
March 15th, 2008 at 3:50 pm
[…] Ammirati interviewed the founders of FriendFeed at ReadWriteTalk in February (podcast+transcript). Tech blogger Louis Gray has taken the lead in covering FriendFeed and related services. See also […]
March 15th, 2008 at 5:23 pm
[…] we discussed include: *ReadWriteTalk interview with the founders of FriendFeed *The RWW Primer on Lifestreaming *Our list of 35 lifestreaming […]
March 22nd, 2008 at 6:41 pm
[…] to an extent as I’d like comments to my blog post on my blog, but according to the podcast on Read/Write Talk there is a subtle difference. This is a way for you, your friends, and their friends to comment on […]
March 30th, 2008 at 4:19 pm
[…] post Scoble’s getting an interview on Monday … I must point out we interviewed them a few months ago on ReadWriteTalk ahh well, I guess he is MSM now - jk […]
April 22nd, 2008 at 7:09 pm
[…] Bret Taylor & Paul Buchheit - Co-Founders, FriendFeed (Podcast) by Sean Ammirati (ReadWriteTalk) […]
May 2nd, 2008 at 8:03 am
[…] As discussed last month, ReadWriteTalk will be syndicating the Semantic Web Gang. On this episode, the members discuss Bret Taylor’s idea of a Wikipedia for Data and look at the role that semantic technologies should play in connecting diverse pieces of data together within and between organisations.  (Note: Bret is also the founder of FriendFeed and was interviewed with his co-founder on ReadWriteTalk.) […]