Search the conversation

Now that Microsoft and Google are going to search Twitter, how to make that useful? Social search is clearly part of the answer – filtering results based on social proximity, based on friend/follow lists. There’s another piece that is missing – the context of the conversation. In Twitter, conversations are represented implicitly by a series of replies between users. Twitter itself does not show that explicitly, though there are clients that do so.

The thing is, in Twitter, each message is very short, and often depends for context on a poster’s previous tweet, and on her replies to other correspondents. So in order to deliver meaningful results, it would be useful to algorithmically reconstitute the conversation.

The border of a conversation is fuzzy. In the recent conversation between Howard Rheingold and his Twitter followers on multi-tasking, there were a series of back and forth exchanges, that interspersed a bit with other topics. An algorithm would approximate the cutoff points where the topic changes, and the conversation ends.

Then, the search result could be shown in the context of the conversation, and make more sense.

The spark for this post is a conversation between me, Thomas Vander Wal and Alan Lepofsky on Twitter.

14 thoughts on “Search the conversation”

  1. Not sure about the role of wave architecturally in the scenario, but the advertising scenario and business model seems right.

    Where Google typically does well is aggregating information across a wide variety of sources, being algorithmically good at showing what’s relevant, and selling ads on their site & others. I see their back-door strategy in a way that is similar but different than the way Jeremiah does. Where I see the leverage is less in individual apps, such as sidewiki and wave. Instead, I see leverage in driving standards that accelerate the progress toward decentralized social networking, which increase Google’s potential value as an aggregator and syndicator of ads.

    The reason I’m skeptical about wave is its social model which seems so… inchoate… in its early form. If anything (in its current form) wave is interesting for in-depth small-group collaboration with ad hoc groups – which doesn’t seem to me to draw as interesting a social graph, as, say, well-adopted Twitter lists. It’s interesting secondarily as an embedding tool only if there are primary uses that are well-adopted.

    This comment was originally posted on A Social Interaction Design (SxD) blog on Web 2.0 & Social Media

  2. Not sure about the role of wave architecturally in the scenario, but the advertising scenario and business model seems right.

    Where Google typically does well is aggregating information across a wide variety of sources, being algorithmically good at showing what’s relevant, and selling ads on their site & others. I see their back-door strategy in a way that is similar but different than the way Jeremiah does. Where I see the leverage is less in individual apps, such as sidewiki and wave. Instead, I see leverage in driving standards that accelerate the progress toward decentralized social networking, which increase the Google’s potential value as an aggregator and syndicator of ads.

    The reason I’m skeptical about wave is its social model which seems so… inchoate… in its early form. If anything (in its current form) wave is interesting for in-depth small-group collaboration with ad hoc groups – which doesn’t seem to me to draw as interesting a social graph, as, say, well-adopted Twitter lists. It’s interesting secondarily as an embedding tool only if there are primary uses that are well-adopted.

    This comment was originally posted on A Social Interaction Design (SxD) blog on Web 2.0 & Social Media

  3. I’m thinkin’ that Google just learned what they needed to from the "Beacon Fail" and will implement a deeper and more personally invasive experiment of their own. Will it be Wave? I dunno, I need to see what people do with it to really get what it’s capable of vs the hype. I definitely believe there will be more and more narrowing of privacy, no matter what the official line is.

    The next question is; as our networks expand beyond our really close friends how much value is there to mining that extended network? We may not know each other or even influence each other at all.

    This comment was originally posted on A Social Interaction Design (SxD) blog on Web 2.0 & Social Media

  4. It has to be one of their end games but I’m not sure it ends with social search ads. Recommendation ads, ads based on your social graph and ads based on what your connections do/like are still not as relevant to the user as ads based on their own likes/habits/activities online. Finding a way to include the social data without actually diluting the quality of accuracy of targeted ads is going to be tricky for Google (I think).

    This comment was originally posted on A Social Interaction Design (SxD) blog on Web 2.0 & Social Media

  5. The good news is that it appears Google and Bing will take different approaches.. Bing – "what’s trendy now", Google – "What’s the latest news". Google’s social search MAY help: I find the most important information is comments from people I know and respect, THAT provides the most relevant context!! But Twitter with it’s free form structure and limited length will be a bitch to post process into relevant data.

    This comment was originally posted on FriendFeed

  6. “The thing is, in Twitter, each message is very short, and often depends for context on a poster’s previous tweet, and on her replies to other correspondents. So in order to deliver meaningful results, it would be useful to algorithmically reconstitute the conversation.”

    I agree that understanding context will be critical for social search. But I think it will be much more complicated than looking back over past actions within a service e.g. previous tweets. Many conversations drift over multiple media. For example, people talk face to face, send each other links on the back of the conversation, continue the conversation over the phone, etc. So, in order to accurately reconstitute the conversation, we’ll need to mine phone calls, text messages, email, I.M. and social networking. And we’re still missing face to face conversations.

    Not to mention the perceived invasions of privacy…

  7. Sure, conversations extend over multiple media, and following those links could be creepy from a privacy perspective. But with regard to Twitter conversation, it may be socially worse to find a word out of context from a single tweet, than a word in the middle of a conversation with 3 people and 6 posts.

  8. Another great post, Chan. Really Makes me think again about just how much of the Web happens via Google. One phrase that really stuck out to me, that I think about often, is just how invisible Google is to many end-user experiences. While MySpace has faded from attention, and Facebook tries to gain as much real estate across the Web as they can, Google is already way ahead. Consider how much Web real estate is owned by Google. Across heavy opensocial properties like MySpace, Orkut. Across blogs with either blogger, Google analytics, Reader, banners. Now aggressively indexing (a phrase they’ve made feel innocuous as they store every bit of data) Twitter.

    Chrome and Wave are merely alpha platforms experimenting with all the data and intelligence they’ve collected, and how little they can give back to people in exchange for that much more direct input.

    One to toss back to you to think about:
    What if Google levied a $15/month Internet tax on every user, based on your five most used IP addresses / ISPs, and if you didn’t pay it, they would not allow you access to any page Google code sits on. What would the Web look like?

    This comment was originally posted on A Social Interaction Design (SxD) blog on Web 2.0 & Social Media

Leave a Reply

Your email address will not be published.