Mike Arrington and I disagree on the future

[kyte.tv appKey=MarbachViewerEmbedded&uri=channels/6118/401143&tbid=k_3522&premium=true&height=500&width=425]

Last Saturday on the Gillmor Gang Mike Arrington wondered aloud whether Twitter had won a “winners take all” game and whether that meant that friendfeed was toast. He then wrote a blog post titled “FriendFeed is in danger of becoming the coolest app no one uses.”

If Arrington is right, then Friendster would have kept MySpace from happening. MySpace would have kept Facebook from happening. Facebook would have kept Twitter from happening.

The thing is that Arrington and I disagree about the future.

He is right about one thing. Twitter has won. It is now the favorite way for people to do microblogging. It LOOKS like a “winner take all” thing, right? After all, who will use a different service than all those celebrities now? To be serious we already know that Twitterers won’t switch, because last year Twitter was down all the time and no one switched, even when they were being abused by the technology.

So that game IS over and Arrington is right about that.

But, is the microblogging game where the cash is? No, I don’t think so.

Well, then, where is the cash?

Search.

“Huh?” I can hear you asking. “Twitter has real-time web search already.”

Yes, but it isn’t all that good.

Here, let’s demonstrate. Remember the plane that crashed in the Hudson?

I want you to find the original picture that someone sent in from their iPhone over Twitter.

Here, I’ll do the searches for you. Go to Twitter Search and type in Hudson Plane Crash. See all the noise? Yeah, the picture is there somewhere but it’s hard to find.

Now, here’s another search, this time using some of the filtering systems over on friendfeed. The picture stands right out.

See, what is going on here is that for search to work you MUST have metadata. Google built a multi-billion-dollar business on the metadata of linking. The next big business will build on top of the metadata of these three things:

1. Who shared or commented on an item. The search above I knew I had liked the picture, so I constrained the search to only things I’ve liked.
2. How many comments or likes are on an item. The search above I knew had tons of comments and likes, so I looked only for items that had more than five likes. That got rid of 95% of the noise.
3. What was said in the item. See on Twitter there’s only 140 characters and other people can’t add tags or info onto that item. On friendfeed I can leave a comment underneath a tweet and make it even more searchable. These comments get indexed in seconds now, which makes them very useful. I can add “cool tweet” to a Tweet and then search for that tweet later. Here’s an example.

NONE of these pieces of metadata are available in Twitter.

This is not just about friendfeed, either. Over on Facebook Mark Zuckerberg has a TON of extra metadata to study as well. He knows who has commented on each other’s walls or who has sent a message, so he knows who your “real” friends are.

Now, to be fair, Twitter does have some metadata to study as well. Retweets are metadata. Already Retweetfeed and TweetMeme are watching that metadata and there are other sites as well.

Twitter can also study hashtags, but I have been talking to people about why I think they are dead. Certainly hashtags are less relevant in the future.

But, notice what Twitter’s metadata is: it has to be included INSIDE a Tweet, where Facebook and friendfeed’s metadata is OUTSIDE the Tweet. Which gives you more characters to communicate with your friends.

So, why are the dollars in search and not just in microblogging?

Well, let’s look how people buy something.

Someone or something creates the need in their head to buy something. I have a baby on the way so I’m starting to look at strollers again since our old one isn’t good enough for a two-baby family.

That’s the “need.”

Then I start talking to my friends. That’s where Twitter and Facebook and Friendfeed come in. I also start researching. That’s where Google comes in and parenting sites and all that.

Then we make a decision. NOW that is the only place where monetization is possible.

Let’s say I decided on a Bugaboo stroller. I search Google for such. Notice all the ads along the right side!

Why are all those ads there? Because businesses know that’s where the money is. Not in blasting ads on the Superbowl. Not in bothering you with ads in my videos or on my blog, where there aren’t many people buying strollers.

No, they want to hit you AT THE MOMENT YOU ARE BUYING.

So, why would Bugaboo want to put ads inside your Tweets? Hint, they won’t. They won’t get many sales that way. But, what if you are searching for information on strollers? Absolutely! That means you’re looking to buy.

OK, here’s where we differ on the future.

We found our new doctor on Yelp. You’ll find all sorts of things this way in the future.

How about a restaurant? A plumber? A TV repair shop? A lawyer?

Consider that you’re walking down the street with a future version of Facebook or Twitter or friendfeed in your hand. You’re looking for a restaurant.

Which is going to be able to bring back the best restaurants that your friends care about?

That requires having metadata to study. That’s why Facebook copied friendfeed’s likes so that it can come back and say “there are four restaurants that have more than 20 likes from your friends within walking distance.”

Translation: the future hasn’t been built yet. That’s why Twitter has not won the entire game yet. That’s why this is a fun industry to watch.

Oh, and Mike, you only have 62 comments and 72 likes, which demonstrates you haven’t done enough with these systems to see the real value. I have 18,300 comments and 17,284 likes and now I have a database that ANYONE can search and find a LOT of value. Plus, I have now been to the future and you haven’t. Yet. You’ll get there. :-)

I gave my Twitter and friendfeed followers a look at the video I did, discussing our disagreement and here’s the comment area that erupted. Keep in mind that this is a live chat room and you can add comments and I’ll see and be able to answer them live on screen there. Oh, that’s something else that Twitter and Facebook can’t yet do. Like I said, I’ve been to the future. You should come and join me.

91 thoughts on “Mike Arrington and I disagree on the future

  1. this is so silly. a 15+ minute diatribe about Friendfeed and twitter, delivered with such drama. is this what you do for a living?

  2. Pingback: Punti di vista
  3. Hi Robert,

    Cool analysis. We’re clearly at the very beginning of the social media phenomenon. Way too early to call. It will replace many forms of communications and advertising over the next 5 years.

    Around minutes 13-14 you come close to describing YouGottaCall’s model.

    Except instead of advertising or ratings (which I categorize as static) we allow you to connect through your friends who trust them.

    Our target is the “long tail of advertising” – the $100B local ad market for SMBs that is currently underserved.

    Instead of metadata that supports the reliability of a local business, we use relationships. Real, local, trusted relationships. That’s why we call it a “Connection Engine” vs. a Search Engine.

    http://cli.gs/YGC-TSP
    http://cli.gs/YGC-slide
    http://cli.gs/vSZRbZ

    Our service is free – but our “Trusted Service Providers” (member businesses) pay a voluntary fee – only for results. The fees are paid out to the members’ favorite charities.

    (Imagine that – a monetization model! :-)

    Check it out. ‘Would love to hear your thoughts.

    – - Tim

  4. Hi Robert,

    Cool analysis. We’re clearly at the very beginning of the social media phenomenon. Way too early to call. It will replace many forms of communications and advertising over the next 5 years.

    Around minutes 13-14 you come close to describing YouGottaCall’s model.

    Except instead of advertising or ratings (which I categorize as static) we allow you to connect through your friends who trust them.

    Our target is the “long tail of advertising” – the $100B local ad market for SMBs that is currently underserved.

    Instead of metadata that supports the reliability of a local business, we use relationships. Real, local, trusted relationships. That’s why we call it a “Connection Engine” vs. a Search Engine.

    http://cli.gs/YGC-TSP
    http://cli.gs/YGC-slide
    http://cli.gs/vSZRbZ

    Our service is free – but our “Trusted Service Providers” (member businesses) pay a voluntary fee – only for results. The fees are paid out to the members’ favorite charities.

    (Imagine that – a monetization model! :-)

    Check it out. ‘Would love to hear your thoughts.

    – - Tim

  5. Good article Robert but I think I disagree with you on main point.

    Twitter owns its data… it’s in their databases and they can mine it and pick it apart as much as they want.

    Who’s to say that the next version of Twitter Search won’t sort results by “relevance” which would be determined by factors like Re-Tweets, number of followers (argh, I know), and other factors. And, if you’re logged in, you could opt to only show results from folks you follow… or people that your followers follow (2 degrees out)… or people X degrees away within your social graph.

    Twitter has all that data and a whole lot more and they’re just not using it right now. This doesn’t mean they *won’t* use it, it just means right now Twitter Search is “good enough” for most people so they’re lax a bit in updating it.

    The FriendFeed search you linked to regarding the Hudson crash was very impressive, but not a lot of information being shown on that page is totally unique to FriendFeed. Twitter could show much of that page with similar sorting/weighting algorithms using the data it has. When will that happen? Perhaps as soon as they deem it important enough.

    FF’s search has the leg up on Twitter right now but *not* because they have superior data metrics to parse through… it’s only because Twitter hasn’t put their full weight into Search yet, and when they do the new Twitter search will be good enough to replace FF search. As I’m sure you know, when you’re the king of an industry and somebody is behind you nipping at your heels, good enough is usually just that… good enough to knock back the #2 for good.

  6. Good article Robert but I think I disagree with you on main point.

    Twitter owns its data… it’s in their databases and they can mine it and pick it apart as much as they want.

    Who’s to say that the next version of Twitter Search won’t sort results by “relevance” which would be determined by factors like Re-Tweets, number of followers (argh, I know), and other factors. And, if you’re logged in, you could opt to only show results from folks you follow… or people that your followers follow (2 degrees out)… or people X degrees away within your social graph.

    Twitter has all that data and a whole lot more and they’re just not using it right now. This doesn’t mean they *won’t* use it, it just means right now Twitter Search is “good enough” for most people so they’re lax a bit in updating it.

    The FriendFeed search you linked to regarding the Hudson crash was very impressive, but not a lot of information being shown on that page is totally unique to FriendFeed. Twitter could show much of that page with similar sorting/weighting algorithms using the data it has. When will that happen? Perhaps as soon as they deem it important enough.

    FF’s search has the leg up on Twitter right now but *not* because they have superior data metrics to parse through… it’s only because Twitter hasn’t put their full weight into Search yet, and when they do the new Twitter search will be good enough to replace FF search. As I’m sure you know, when you’re the king of an industry and somebody is behind you nipping at your heels, good enough is usually just that… good enough to knock back the #2 for good.

  7. I agree with you, Robert. I think that FriendFeed and Twitter should merge, and together they can kick both Google’s and Facebook’s asses. :-) But the problem I see with FriendFeed is the same I saw with FeedBurner – that is, it’s very specialized in what it does well. FriendFeed needs 3rd party apps and other novel ways for people to use the service. Otherwise, it’s no more than just a “Forums 2.0″ app.

  8. I agree with you, Robert. I think that FriendFeed and Twitter should merge, and together they can kick both Google’s and Facebook’s asses. :-) But the problem I see with FriendFeed is the same I saw with FeedBurner – that is, it’s very specialized in what it does well. FriendFeed needs 3rd party apps and other novel ways for people to use the service. Otherwise, it’s no more than just a “Forums 2.0″ app.

  9. It is silly to quote the doctor example. If you look at the decision tree, perhaps a primary physician would remotely fall under the “I have a friend who uses this doctor, therefore I should”.
    There are more than 20 variables to result in that decision. With false-negatives being very high.

    All this blog-wanking about business models by just scratching the surface is a little tiring.

    Also any comment by Mr Obasanje — who should understand the power of networking, given that his close relation in Nigeria was implicated in massive corruption (which included Microsoft providing the country with computers and software) and his position in Microsoft — is morally weak in my opinion…

  10. It is silly to quote the doctor example. If you look at the decision tree, perhaps a primary physician would remotely fall under the “I have a friend who uses this doctor, therefore I should”.
    There are more than 20 variables to result in that decision. With false-negatives being very high.

    All this blog-wanking about business models by just scratching the surface is a little tiring.

    Also any comment by Mr Obasanje — who should understand the power of networking, given that his close relation in Nigeria was implicated in massive corruption (which included Microsoft providing the country with computers and software) and his position in Microsoft — is morally weak in my opinion…

  11. 1st off I think grouping services as “micro blogging” is the first mistake. Twitter does twittering, Friend feed does friend feeds etc. They aren’t all takes on the same activity in the same way as “Web 2.0″ was a big murky grouping.

    If you must put them head to head a critical factor is the inventory the service produces, twitter is generating way more inventory in the form of tweets. With huge data sets pattern mapping gets better and better and then there isn’t reliance on crowd sourcing to produce this supposed useful output. Friend feed needs lots of people to shut up and just vote rather than say something and I don’t think there is enough social currency in being a reviewer in such a short form medium.

    The signal to noise issue really doesn’t come into it for normal people who are and will use the services. Most people follow people they know or have an affinity with. With the naturally narrower data set basic searches can filter efficiently enough to manage the torrent. AT&T or NIKE May have a similar noise problem from volume that you do Robert, but don’t ask me to use my time to solve AT&T or NIKE’s problem by creating that meta data. Sure, if that meta is created inadvertently as I do what I want to do, say by mentioning them several times in a tweet exchange. That’s cool because it costs me nothing.

    I get your reasoning though, got me thinking too so thanks ;-)

  12. 1st off I think grouping services as “micro blogging” is the first mistake. Twitter does twittering, Friend feed does friend feeds etc. They aren’t all takes on the same activity in the same way as “Web 2.0″ was a big murky grouping.

    If you must put them head to head a critical factor is the inventory the service produces, twitter is generating way more inventory in the form of tweets. With huge data sets pattern mapping gets better and better and then there isn’t reliance on crowd sourcing to produce this supposed useful output. Friend feed needs lots of people to shut up and just vote rather than say something and I don’t think there is enough social currency in being a reviewer in such a short form medium.

    The signal to noise issue really doesn’t come into it for normal people who are and will use the services. Most people follow people they know or have an affinity with. With the naturally narrower data set basic searches can filter efficiently enough to manage the torrent. AT&T or NIKE May have a similar noise problem from volume that you do Robert, but don’t ask me to use my time to solve AT&T or NIKE’s problem by creating that meta data. Sure, if that meta is created inadvertently as I do what I want to do, say by mentioning them several times in a tweet exchange. That’s cool because it costs me nothing.

    I get your reasoning though, got me thinking too so thanks ;-)

  13. Scoobie, what I dont get is why doesn’t FF simply postion itself as the leading Twitter client? And then extend the service with comments, threading, voting, blah, blah…

  14. Scoobie, what I dont get is why doesn’t FF simply postion itself as the leading Twitter client? And then extend the service with comments, threading, voting, blah, blah…

  15. Fantastic post and video!

    I think I’m in the minority, but I’ve always approached FriendFeed as a data flow. Maybe it’s my science fiction fandom, but it strikes me that FriendFeed is akin to the Gibson (Neuromancer) or Stephenson (Snow Crash) version of data.

    FriendFeed provides UI and tools that allow me to extract information from the river of data. Some of that will be social because part of the data is social in nature and the interaction (the meta data) is also social.

    But I wouldn’t call FriendFeed a social network – it simply contains social data.

    The filters rolled out in the new FriendFeed beta increase my ability to extract information. As you mention, the meta data (likes, comments) helps turn the data into information. The data is annotated. It is transformed.

    Extracting information (e.g. – search) from Twitter is painful. The idea that Twitter can compete here is just flawed.

    http://www.blindfiveyearold.com/twitter-is-not-a-google-competitor

    Twitter is very good at what it was designed for – an update service. As you note though, how that actually translates into revenue is anyone’s guess. Though the recent sponsored microsites are an impressive first step!

    I believe FriendFeed has a leg up on information extraction. That could be flat out search or it could be something altogether new and different. I’m also of the mind that ‘linear search’ is declining and ‘non-linear search’ will become the way in which we learn.

    Part of the ‘chaos’ that FriendFeed delivers is a diverse subject set. I’m not just reading about one topic. I’ll skip from topic to topic, back and forth. To some this sounds unproductive. But how many times are you able to apply lessons from one topic to another? Plenty!

    I’ll use a jigsaw puzzle metaphor. Traditional information gathering (linear) means you do the edges first – all of it – before moving on to another section, which you must then complete. Non-linear means you could do part of edge, then some of the castle, serendipitously put together some difficult sky pieces and then back to edges.

    I think the latter is far more effective now that we can receive information at such high velocity

    Finally, the idea of experiential data (doctor, mechanic, restaurant) is interesting because it again contains a social element. While I believe Facebook or FriendFeed can help aggregate these experiences, there are still issues of trust and relevancy to be resolved.

    If I like my scrambled eggs loose and you like them firm – that could have a profound impact on how you rate a diner. Or if you like eggs at all and would rather have pancakes. Do you value a bedside manner in a doctor or not? How much?

    Certainly meta data can help resolve this – but I sense there should be better ways to extract the right matches.

  16. Fantastic post and video!

    I think I’m in the minority, but I’ve always approached FriendFeed as a data flow. Maybe it’s my science fiction fandom, but it strikes me that FriendFeed is akin to the Gibson (Neuromancer) or Stephenson (Snow Crash) version of data.

    FriendFeed provides UI and tools that allow me to extract information from the river of data. Some of that will be social because part of the data is social in nature and the interaction (the meta data) is also social.

    But I wouldn’t call FriendFeed a social network – it simply contains social data.

    The filters rolled out in the new FriendFeed beta increase my ability to extract information. As you mention, the meta data (likes, comments) helps turn the data into information. The data is annotated. It is transformed.

    Extracting information (e.g. – search) from Twitter is painful. The idea that Twitter can compete here is just flawed.

    http://www.blindfiveyearold.com/twitter-is-not-a-google-competitor

    Twitter is very good at what it was designed for – an update service. As you note though, how that actually translates into revenue is anyone’s guess. Though the recent sponsored microsites are an impressive first step!

    I believe FriendFeed has a leg up on information extraction. That could be flat out search or it could be something altogether new and different. I’m also of the mind that ‘linear search’ is declining and ‘non-linear search’ will become the way in which we learn.

    Part of the ‘chaos’ that FriendFeed delivers is a diverse subject set. I’m not just reading about one topic. I’ll skip from topic to topic, back and forth. To some this sounds unproductive. But how many times are you able to apply lessons from one topic to another? Plenty!

    I’ll use a jigsaw puzzle metaphor. Traditional information gathering (linear) means you do the edges first – all of it – before moving on to another section, which you must then complete. Non-linear means you could do part of edge, then some of the castle, serendipitously put together some difficult sky pieces and then back to edges.

    I think the latter is far more effective now that we can receive information at such high velocity

    Finally, the idea of experiential data (doctor, mechanic, restaurant) is interesting because it again contains a social element. While I believe Facebook or FriendFeed can help aggregate these experiences, there are still issues of trust and relevancy to be resolved.

    If I like my scrambled eggs loose and you like them firm – that could have a profound impact on how you rate a diner. Or if you like eggs at all and would rather have pancakes. Do you value a bedside manner in a doctor or not? How much?

    Certainly meta data can help resolve this – but I sense there should be better ways to extract the right matches.

  17. As a general statement there’s a huge assumption being made here that there is even a concept of a pepetual “win”, I see no reason to reach this conclusion.

    Twitter is certainly currently winning (present tense), but that’s a long way from having “won”. It takes an incredible amount of build a genuinely sustainable “monopoly” which generates super-normal profits.

    Twitter, Facebook and FriendFeed are simply nowhere near mature enough in their technologies yet. They also struggle considerable because they are not associated with a typical user’s purchase decision. This is, currently, where Google sits (in many markets).

    Facebook’s ad programme is a case in point here. The segmentation on offer is incredible, but still it only carries ads for credit reports and dating. And that’s because people are not in a transactional frame of mind when they are using Facebook, the same applies to Twitter.

    The search/browse/recommendation and selection processes of online purchases are very specific and, I don’t believe, well served by general conversation media. That said, you know my business is online retailing, so my views could be considered to be skewed.

  18. As a general statement there’s a huge assumption being made here that there is even a concept of a pepetual “win”, I see no reason to reach this conclusion.

    Twitter is certainly currently winning (present tense), but that’s a long way from having “won”. It takes an incredible amount of build a genuinely sustainable “monopoly” which generates super-normal profits.

    Twitter, Facebook and FriendFeed are simply nowhere near mature enough in their technologies yet. They also struggle considerable because they are not associated with a typical user’s purchase decision. This is, currently, where Google sits (in many markets).

    Facebook’s ad programme is a case in point here. The segmentation on offer is incredible, but still it only carries ads for credit reports and dating. And that’s because people are not in a transactional frame of mind when they are using Facebook, the same applies to Twitter.

    The search/browse/recommendation and selection processes of online purchases are very specific and, I don’t believe, well served by general conversation media. That said, you know my business is online retailing, so my views could be considered to be skewed.

  19. This is the single most important video/point I’ve heard you make in several years Robert. As I said over on FF (http://beta.friendfeed.com/scobleizer/927f84a3/mike-arrington-and-i-disagree-on-future), there’s magic in this real time stuff and major $$$ waiting to be extracted by the one who can attract the attention of the crowd and show them the path to the awesomeness that is riding the real time river. Whoever builds the best visual metaphor for slicing, dicing, and mashing up the river will win these precious $$$. Exciting times indeed. As to whether Mike gets it or not, I can’t really comment on that because I haven’t met him (*yet*) and wasn’t privy to your conversation. Great post.

  20. This is the single most important video/point I’ve heard you make in several years Robert. As I said over on FF (http://beta.friendfeed.com/scobleizer/927f84a3/mike-arrington-and-i-disagree-on-future), there’s magic in this real time stuff and major $$$ waiting to be extracted by the one who can attract the attention of the crowd and show them the path to the awesomeness that is riding the real time river. Whoever builds the best visual metaphor for slicing, dicing, and mashing up the river will win these precious $$$. Exciting times indeed. As to whether Mike gets it or not, I can’t really comment on that because I haven’t met him (*yet*) and wasn’t privy to your conversation. Great post.

  21. Man, seriously fascinating video Robert. I thought it was going to be all about your disagreement with Arrington, but no – you’ve got me thinking.

  22. Man, seriously fascinating video Robert. I thought it was going to be all about your disagreement with Arrington, but no – you’ve got me thinking.

  23. I agree on the value of metadata, but feel that non social meta data is being overlooked. We need to use available content to provide context, hence building a “deeper” experience. We need maps (google maps), concepts (wikipedia), reviews (yelp), etc added to the info our friend feed is putting in front of us.
    Social meta data in the form of “likes” or “comments” can be use to filter or, with enough volume, can end up adding to the noise.

  24. I agree on the value of metadata, but feel that non social meta data is being overlooked. We need to use available content to provide context, hence building a “deeper” experience. We need maps (google maps), concepts (wikipedia), reviews (yelp), etc added to the info our friend feed is putting in front of us.
    Social meta data in the form of “likes” or “comments” can be use to filter or, with enough volume, can end up adding to the noise.

  25. Great read Robert, but consider product deliverability and accessibility. Simply put, Friendfeed just isn’t as accessible to the mainstream yet. They need to do a better job of communicating its purpose and simplifying the application before the mainstream will take notice. Sometimes, it takes more than simply having the best product.

    On the other hand, Twitter is establishing a trust with users as the destination for real time information. That gives them a big head start, even if Friendfeed’s search is significantly better at the moment.

    In the end though, it’ll probably be Google that cracks this nut. I predict they’ll strike deals with Twitter and others to gain access to their streaming data. Then the game is truly over.

  26. Great read Robert, but consider product deliverability and accessibility. Simply put, Friendfeed just isn’t as accessible to the mainstream yet. They need to do a better job of communicating its purpose and simplifying the application before the mainstream will take notice. Sometimes, it takes more than simply having the best product.

    On the other hand, Twitter is establishing a trust with users as the destination for real time information. That gives them a big head start, even if Friendfeed’s search is significantly better at the moment.

    In the end though, it’ll probably be Google that cracks this nut. I predict they’ll strike deals with Twitter and others to gain access to their streaming data. Then the game is truly over.

  27. The value I see in Twitter is not in the post itself, but the link in the tweet. Of course, there are time when the tweet says it all, like the ones @pogue does on “Tonight’s Meditation”, but for me atleast, the “value” is in the links and the micro-description of the link in the tweet. Helps me do less of surfing on RSS feeds to get my news.

    So all major bloggers, celebrities have their “broadcast” channels in the Twitterworld, I can go assemble my favorites list of channels choosing who I follow …

  28. The value I see in Twitter is not in the post itself, but the link in the tweet. Of course, there are time when the tweet says it all, like the ones @pogue does on “Tonight’s Meditation”, but for me atleast, the “value” is in the links and the micro-description of the link in the tweet. Helps me do less of surfing on RSS feeds to get my news.

    So all major bloggers, celebrities have their “broadcast” channels in the Twitterworld, I can go assemble my favorites list of channels choosing who I follow …

  29. You got the point. Twitterverse may generate huge amounts of data, but lots of this data is automatically being pushed to FirendFeed and covered with metadata in the form of likes and comments. Eventually, this leads to FriendFeed’s higher signal to noise ratio.
    Generally speaking, FriendFeed is the next generation of communication services. It builds atop of previous generations services like Twitter and Facebook, filtering relevant information for its user (providing higher signal to noise ratio), and this is where the money are.

  30. You got the point. Twitterverse may generate huge amounts of data, but lots of this data is automatically being pushed to FirendFeed and covered with metadata in the form of likes and comments. Eventually, this leads to FriendFeed’s higher signal to noise ratio.
    Generally speaking, FriendFeed is the next generation of communication services. It builds atop of previous generations services like Twitter and Facebook, filtering relevant information for its user (providing higher signal to noise ratio), and this is where the money are.

  31. Renee, totally agree with you. When people need “trusted” recommendations, they generally rely on friends or friends-of-friends (e.g. “my friend knows a great gardener”).

    A colleague and I have developed a service (http://kelpp.com) that does exactly this – currently at the “alpha” stage of development, but fully functional and we’d love to have folks try it out and send feedback.

    In a nutshell, the service integrates with Facebook and lets you ask your friends for recommendations on anything, and tracks their responses for you. You can also search for existing recommendations (not a lot in the system as we’re new) and we automatically prioritize reviews from your friends and friends-of-friends.

    We plan to integrate more deeply with Twitter in the future – you can see our current integration by following http://twitter.com/kelppme

  32. Just replied to another comment on this thread, but couldn’t resist replying to this one as well. I think AJ’s point about relevancy (whether I like eggs vs. pancakes will impact my review for a restaurant) is right on target.

    Put another way, the “social graph” is just one element needed for a rich recommendation service. As we developed Kelpp, we’ve focused on incorporating the following in addition to the social graph dimension:

    “Goods or services” – these are tags that describe what you are reviewing or looking for. (e.g. “scrambled eggs”, “inexpensive vegetarian lunch”, “great city views”, etc.)

    Business / locations – the “place” you are reviewing.

    “Requests for recommendation” – these indicate what “goods or services” a user is looking for recommendations on. These requests are communicated to the user’s friends (via Facebook).

    Reviews – e.g. “user X strongly recommends business Y for tags A, B, and C”

    You can see this in action by taking a look at http://kelpp.com/places_of_business/view?pob_id=42

    This business has multiple reviews for different “tags”, which allows for very precise recommendations. You can also explore the “tag space” to find related businesses – for example if you click on the “vegetarian” tag you will see several other restaurants that have been reviewed for that tag.

    Would love to hear if folks think this concept makes sense and is valuable.

Comments are closed.