Your TwitterBase
I recently read a post by Mark and Sean Evans over on their blog talking about Twitter. They touched on something that I’ve been hearing and thinking a bit about lately, but defined it in a particularly succinct manner:
“Twitter has accomplished want nobody, not even Google, has yet to figure out - crowdsourcing search. How often have you seen a request for advice, recommendations or suggestions come across your Twitter feed? “
They have really hit on what could be considered the “killer app” for Twitter: the ability to get instant advice on any number of things from a group of people you have essentially hand picked (I like to call it my “TwitterBase”). Traditionally, forums have been my go-to resource when I need to know something that can’t be answered by a Google search. More recently, Metafilter and a few other services have sprung up as well but it is hard to beat the near instant access to handpicked advisers that Twitter can give you.
Of course, there are a few advantages to user group or forum based methods of accessing advice:
-First, when you post or access advice on a forum you are directing your search towards a very specific subset of people who are likely to have the answer you may need. On Twitter, there may be a few people who follow me who can offer some advice about where to find a good email client for my Treo, but to a lot of people who follow me that question is just going to be noise in their Twitter Feed.
-Second, forums generally have some sort of metric in place that allows you to make some assumptions about the particular people offering replies. It could be as granular as showing how many posts a user has made next to his or her nameplate, but it is still marginally useful in intuiting to some extent how trustworthy the answer may be. Also, because of the persistent nature of each thread conversations tend to go on longer and allow for bad ideas to be refuted. “Junk” posts become obvious.
Twitters’ “Other” Scaling Problem
It’s no secret that Twitter has had trouble scaling it’s back end to meet the increasing demand. That isn’t the real challenge though. Hardware can be upgraded, software can be tweaked. Like anything in this ever-accelerating information age, the real bottleneck in throughput is the parsing software that sits directly behind the eyes of the user.
Twitter, forces you to maintain the “trust’ database” yourself in your own brain. Ostensibly, this works great if you only have a few select people that you follow on Twitter, but the problem is to be a truly useful friend sourcing tool, you want a large pool of people who have different areas of knowledge. Also, the larger and more diverse your personal ‘think tank’ of friends is on Twitter, the harder it is to maintain an optimal signal to noise ratio both incoming and outgoing. You have to depend on your internal memory in order to determine who to listen to, and how much to trust them on any given topic.
Using Twitter to get answers on a topic right now is akin to yelling out a question in a room full of people. Now, if there are ten people in the room then sure, it’s going to be easy to recognizes the guy who answers you. After all, you probably just talked to him over by the appetizer tray. But what if there are 400 people in the room? Is that guy in the leather jacket that just gave you advice on your car insurance the dude who had that really insightful chat with you last week, or is he the guy that you met at some party 4 months ago who turned out to be a real dummy? Twitter doesn’t give you the external systems you need to keep track of all this unless you can do it in your head. And that, sadly enough won’t scale to terribly well.
I’d love to see a couple features either integrated into Twitter itself or as a third party service to address this. Each of these features focuses on the ability to scale your personal friend sourcing think tank efficiently so that it operates as well when you have 2000 people you are following as it does when you have 200.
What Twitter needs to do to become a truly effective friend sourcing tool is to incrementally offload 2 tasks from the users’ brain: Content parsing (signal-noise ratio) and Context (Who is this person?).
-Custom user notes It would be great if you could type out a short line or two about a particular person that you could access when mousing over a twitter by them. For example, if I mouse over John Smith, maybe the note I made for him says “This guy helped me with my Treo Question” It’s a nice, easy way to offload that “trust database” so you don’t have to maintain it in your own brain.
-Groups I’d like to be able to assign Twitter users to groups that I create so I can filter groups on or off, or Tweet only people in a particular group. Lets say I have a “Web developers group” If I have a question about PHP, I’d like to not have to clog everyone’s Tweet feed with it if I can efficiently ask people I’ve already decided will likely know the answer. Conversly, this way if I’m perusing my Tweet feed and don’t care to see a lot of stuff on web development, I could turn this group off and hide Tweets coming from anyone in it temporarily.
-Custom styles for groups/users. There are certain people who I like to make sure I don’t miss when I peruse the Twitter feed. I’d like to be able to Style their Tweets so I don’t miss em. This could be organic, using simple up-down Reddit-style arrows next to tweets that promote/demote users to more or less visible styles. If you wanted to get really interesting you could parse mouseover and Twitter reply data to adjust the styling for the Twitterer’s I interact with most frequently atomatically.
-A ’save this conversation’ option. This button would archive that thread and allow you to acces it later from a link on that persons’ twitter homepage or when you mouseover a Twitter by them. I propose accessing it only in these contexts because the primary purpose of saving these clipping would be to easily let you provide yourself context for who this Twitterer is when you encounter them later on. When you save a Twitter, you are basically saying “Hmm, if I see this later, it will remind me who this person is so I’ll know to listen to them.” Favorites don’t contextualize things enough to make it useful for this.
One could argue that self-assigning is a headache, (You are in essence creating your own subsets and networks as opposed to simply seeking out pre-existing ones) but I the the value here is that you maintain the one thing that really does make Twitter valuable: everybody on your follow list is pre-vetted by you to some extent.
Are there tools out there that provide this functionality yet? I’m not sure but I think people would find it highly useful. What ways have you found to manage your own personal ‘Twitterbase’ ?
