Tribal search


tribescape is a newish entry into the social search arena that allows searchers to share search results with their peers through Twitter.  Rather than e-mailing URLs, you simply click on button next to each search result, pick the followers to whom you want to tweet this result, and you’re done. Convenient, yes. Collaborative? Maybe.

I can certainly see using tribescape to support a rudimentary form of UI-level mediation, but it suffers from several shortcomings. You get to see the URL that’s being shared, but not the query that retrieved it or the other documents that accompanied it. SearchTogether, on the other hand, does a much better job of preserving the context of the query while also sharing the results with a group of searchers.

Another limitation is that Twitter messages, unless saved in some manner, tend to disappear after some time. This makes tools such as tribescape useful for sharing  results in the moment, but may break down on on-going collaboration. The system might benefit from a TwapperKeeper-style archival mechanism that would help searchers preserve their results. Finally, the system would benefit from showing more than eight hits per query, since despite calls for the demise of recall as a measure of interactive search performance, exploratory search benefits from being able to explore the result set more deeply.

tribscape is one of a class of tools that have emerged recently that try to leverage social networks in support of information seeking. It would interesting to evaluate these tools empirically (can you say, Masters Thesis?) to assess the relative contribution of the social aspect to information seeking, and to compare the effects of different levels of  of mediation (result exchange only, search context, and algorithmic mediation) on the outcomes. Come to think of it, perhaps this more ambitious goal may be PhD-worthy. The trick will be to see how many of these search engines survive long enough to be studied.


  1. Thanks for the write-up Gene.

    Collaboration – how about asynchronous collaboration for the time being:)?

    Archiving – when you share a result with your colleagues/follower, a copy of this result is available with Tribescape. You/your recipients could browse it from within the Tribescape system along with few “local” hit statistics.

    Features – few trivial collaboration techniques along with few nifty ones are planned for Tribescape, just waiting for some user traction to see where it goes.


  2. Thanks for the comments, Dhruba!

    I missed the saved results feature when I played with the system; that’s certainly useful!

    What do you mean by asynchronous collaboration? Do you mean that the collaborators don’t have a shared view (WYSIWIS)? If so, I think you’re right and that such a view is typically not necessary. On the other hand, sharing data as we wrote about in the synchronization post, is key to collaboration, as it allows collaborators to influence each other, thereby improving the quality of the results through both the sharing of knowledge and the division of labor.

    Glad to see that you’re innovating in this space. I was wondering if you had considered Yahoo! BOSS as a search platform. It would give you deeper result sets that could be used to surface less common (but perhaps more interesting) results. We’ve found that to be particularly useful for exploratory search, which also benefits from collaboration.

  3. I used the term asynchronous from a time “chat” perspective. Two or more collaborators need not be present on the system at same time. Here Twitter is very handy as it will send your recipients some form of alerts – email or SMS.

    Task splitting, keyword association and peer reviews charts are other things, which I would like to add. I have added a feedback form (powered by uservoice) to let users cast their views/votes.

  4. You might think about forming groups up front: join a group, and then automatically distribute data among that group. Data could be documents that people found useful, queries that have been run, documents that are retrieved but are not useful or redundant, etc.

  5. I second that feature. Ref. May I add that your definition of Data is on dot.

Comments are closed.