FriendFeed ‘Likes’ Index: Case Study in Value of Distributed Conversations

By keeping comments distributed, or decentralized, more than one discussion is able to take place. New ideas are likely to be heard since readers often start with a blank slate and are more likely to participate.

Shey Smith, introspective snapshots, The Case For Distributed Conversations

Today, a great example of the value of distributed conversations took place. What started as a blog post here ended up with three different developers coming up with innovative new scripts that FriendFeeders were digging. And it all happened because of distributed conversations, not despite them.

The very smart and in-tune Fred Wilson wrote a piece yesterday decrying the distribution of conversations all over the Web, including on FriendFeed. Mathew Ingram followed up with a concurring blog post. I understand where they’re coming from, but I think they overlook the value of distributed conversations.

What I’d like to do is briefly describe the action today, and then point out how distributed conversations made innovation possible today.

FriendFeed ‘Likes’ Index Calculators

Wednesday night, I posted a piece titled FriendFeed ‘Likes’ Compatibility Index. The post reported some number crunching I did to figure out who most often Liked the same things that I do. The idea was to see what other FriendFeeders shared the same interests. At the end of the post, I made a request for someone to automate the analysis.

From this post, two separate conversations emerged. The RSS feed for the blog post hit FriendFeed (Original Post). And Louis Gray shared it on Google Reader (Shared Post), which started a second conversation.

What happened? There were three different places where conversations were happening: on this blog and on two different items in FriendFeed. And it resulted in three separate developers coming up with solutions.


Yuvi, a 17-year old wunderkind who does amazing stats analysis, was interested in automating this analysis. He posted the same comment on all three locations: “I could automate this…if friendfeed fixed this bug.” Yuvi was concerned about a bug in FriendFeed that won’t allow you to go more than 11 pages back in your history.

His comment generated responses in FriendFeed on both the Original Post and on the Shared Post.

Original Post:

  • Phil Glockner: “Yuvi, does that bug exist when doing queries against the API?”
  • Yuvi: “Yes, it exists in the API too.”

Shared Post:

  • Shey: Yuvi, could you automate it up to page 11?
  • Hutch: Does the limit of going back beyond page 11 risk the script failing? Or does it limit the data collected?
  • Yuvi: @Hutch: Limits data collected.
  • Yuvi:@Shey: Well, I could… But, it’ll be of limited use, no?
  • Bwana: I say do it now so when they do fix it, it’ll be ready, plus there seems to be an interest
  • Shey: @Yuvi Limited yes, but I think 11 pages of data is of some use for analysis of recent data
  • Cyndy: Yuvi, I’m not sure it’s a bug. I think it’s a limit that they have set. Since the variable is passed in the URL, if you try to go past that number of posts manually, it still won’t go. Could be that they are only pulling from cache?
  • Yuvi: @Cyndy: Well, they’ve been mum on this – so I don’t really know. But, if even *I* can’t access my old stuff, isn’t that wrong on at least “some” level?
  • Benjamin Golub: I don’t think it’s a bug either. I feel that there DB sharding might be setup such that it is very very quick to pull recent data.
  • Bwana: Well if there is a limit imposed, pages after 11 shouldn’t even be shown. It’s a bug of some kind either way.
  • Yuvi: @Benjamin: Yep, agree on that, but there should be ‘some’ way to get the older data out, no?
  • Yuvi: Just repeating – the API has the same limit in place. Script ready anyway – First Target – LouisGray 😉

So in that sequence, you see that fragmented conversation, away from the blog post itself, resulted in Yuvi creating a script to determine who shares your Likes.

And Yuvi blogged about it, linking to my blog post and even mentioning me by name. Everything a blogger could want.

Do you see what I mean Fred and Mathew?

Ole Begemann

On Louis’s Shared Post, a second developer Ole Begemann weighed in:

  • Ole Begemann: I’ve written a Python script that does this, too (for practice). Interestingly, Phil is no. 12 on my list of Louis Gray’s most compatible likers. If there’s interest, I’ll try to wrap it up on a web page (it’s command line at the moment) and publish it.
  • Hutch: @Ole – Yeah, I’d like to have a page where you could see these results.
  • Ole Begemann: I’ll get around to it Hutch. It might take me a few days. It’s my first try as a Python programmer. 😉

A second developer came up with a script for this. Again, via conversations that happened entirely away from the originating blog.


Finally, back on my Original Post in FriendFeed, a developer named felix added this comment:

“I just created a little javascript to go and grab the last 30 likes of anyone and do a basic calculation. Have a couple more features I want to add, but no more time today – what do y’all think?

That link goes to a blog post, where Ole links back to my original blog post. Again, as a blogger who wrote something I thought might be interesting, this is all really good stuff.

None of it occurred on my blog. And it doesn’t bother me in the least! in fact, check out felix’s blog post. You’ll see that he, Yuvi and Ole are having a conversation about FriendFeed API limits.

Why the Distribution of the Conversation Made a Difference

Three points to make here.

1. Go where the conversations are

If I’d been hung up on forcing everyone back to my blog for comments, this likely would not have been as successful as it turned out. FriendFeed offers a dead simple commenting function that makes it incredibly easy to comment. People find it easy to interact around content, rather than everyone having to travel from blog to blog to hold conversations.

Some blogger removed his RSS from FriendFeed recently, because he didn’t like all the FriendFeed comments along with it. Really? I remember the story, but can’t find the link to his blog. Seriously.

2. Connect to people outside your blog subscriber base

Digg, StumbleUpon, FriendFeed…all of these give exposure to your blog outside of those who subscribe to it or bookmark it. And when conversations about your blog occur on these venues, you’re getting vital exposure.

Make no mistake about this. A Like or a Digg or a Stumble is great. But if you really want to attract people to your blog post, comments are king. They tell people that the post is interesting, and that they better go read to get in on the discussion.

Louis Gray has a bigger, and different, community than I do. So his share of the post on Google Reader, and the subsequent conversation, attracted people who might never have bothered with my post.

felix, who developed the really cool app where you can see who shares your Likes, does not subscribe to me in FriendFeed, nor does he subscribe to my blog. I looked at his subscriptions, and we do have a number of FriendFeeders in common including Louis. I presume that’s how he found his way to the conversation about the blog post. Would he have been attracted to the blog without the conversation going on inside FriendFeed? Unlikely.

Embrace distributed conversations. They are free advertising for your blog.

3. Use the everyone search feature

Have people figured out this one yet? On FriendFeed, you can run a search for your blog post title in the ‘everyone’ tab. It can be a little hectic, but also fascinating. Click here for the everyone search for the FriendFeed ‘Likes’ Compatibility Index post.

Note that not only will you see all the different instances of my original blog post. You’ll see Yuvi’s post as well as Thomas Hawk’s post on the subject. I like seeing comments on those related posts as well.

As a blogger, I get a lot of value out of seeing who liked the blog post, and all the conversations among the different tribes. They help me improve.

Final Thoughts

Would that blog post have resulted in three separate scripts being developed if conversations only happened on the blog? No. At least not for me. If you’ve got a huge subscriber base like Fred Wilson or Mathew Ingram, it might.

But if you’re small fry, the distribution of conversations provides enormous value. Now let me go see who shares my Likes on FriendFeed…


See this item on FriendFeed: