When I talk to my friends and family, their unanimous complaint about Twitter is, "I follow too many people. My timeline is just too noisy."- which leads to using Twitter less and less.
The problem is Twitter issupposedto be an interest graph, but really it is just a people graph because you can't follow interests on Twitter, you follow people... this leads to a home timeline that is muddled with lots of tweets you don't care about with a few interesting ones mixed in. [Formatting altered]
We've been saying this for years now. Not about Collections specifically, but about the concept at large. I didn't even know Collections was a thing that existed until I read this post, linked by @mathewi.
Twitter is an invaluable network. I used to live on Twitter, but I rarely check it all the time anymore. It's exactly because of the problem identified above - my timeline is full of noise, and signal is very, very rare. That's even after I pared down my following list.
There are people I want to follow. Generally speaking, I don't care about 95% of their tweets. Sorry. I think you're a cool person and I want to stay in touch, but our interests have very little intersection.
There are other people who have very interesting tweets and content - but I don't know or care about them as individuals.
There are users who retweet a large volume of content into my feed. I usually don't care about users' retweets unless they also happen to intersect with the reason I followed them to begin with - who they are as a person, or their content.
Then there is the low percentage of my following list that consists of news, businesses, organizations, celebrities, and public personas. While many cite these as the reason that users end up on Twitter, they're not very important to me.
So how can all this content be organized? How can the flow of the firehose be controlled intelligently so that I have a better chance of seeing what I want to see and not seeing what I don't want to see?
Jazzychad brings up a few comparisons to other social networks. Pinterest allows users to follow a user or pick and choose from their boards to follow - user content curation. A comparison to Facebook was also drawn in a negative light with regard to algorithmic curation.
I think Pinterest is an interesting conceptual analogy, but it doesn't hold up quite as well. For one, Pinterest isn't used for communication the way that Twitter is; it's an image-focused network, and in that way, the nature of curation differs.
Say what you will about Facebook and its algorithmic curation, but I think it's working for them. The network is similar to Twitter in that you follow people, not topics or ideas, so you get all the noise of their posts regardless of topic. But the algorithm sorts through this noise and presents it in the manner it thinks best suits the user, so you (hopefully) get lots of signal and little noise.
The upshot of this is that any time I've visited Facebook in the past few months, drawn there by notification of comments on my automatically-crossposted tweets, I actually end up reading a good chunk of my News Feed. It's stuff that's relevant and interesting to me. My biggest complaint when I stopped using Facebook was that the content was just irrelevant and, honestly, infuriating. I don't really have that complaint anymore. Facebook keeps tempting me back with relevance.
Twitter's gone the other way. Whereas I used to keep on top of Twitter regularly, checking it dozens if not hundreds of times a day, I just can't seem to be bothered anymore. The same has happened to my RSS aggregator. I get most of my news and content from reddit - another user-curated network, where what I see is what thousands of users have deemed important or relevant.
Google+ tried to walk the middle path by implementing Circles, but that wasn't a perfect solution either, because it also wasn't interest-based. You still followed people and all their posts - so even if I followed an industrial designer in a Design circle, I would see all of their non-design-related posts as well.
So what does make sense? Tags. Now, I don't know how Pinterest Boards work since I'm not a Pinterest user, but I imagine that they are like folders, as I imagine Twitter Collections are. I don't like folder systems as organization tools - I much prefer tags. The difference is that you can only put a piece of data into a single folder, whereas it can have as many tags as you put on it.
Take the example tweet, "Twitter acquires TweetDeck". What folder would I put this into? Twitter? Social Media? Tech Acquisitions? News? Or, I can tag it with every topic that's relevant - twitter, social media, acquisition, tech news, and so on. This is the system that Tumblr uses.
Of course, Tumblr doesn't let you follow a blog's tags, only blogs themselves. And you can search tags the same way hashtags on Twitter work. But I see no reason why Collections can't be based around tags, or you couldn't follow a user's tags directly.
So, for example, let's say for my username @gojeffcho, I have +collections and .tags. +Technology could be .apple .technews .ios .software .hardware .android .linux .google and so on; +Personal could be .work .home .life .kids and so on. You could follow @gojefcho, @gojeffcho+Technology, @gojeffcho+Technology+Personal, @gojeffcho+Personal.android.google.linux, etc.
It's messy, unrefined, and inelegant - but I'm just throwing ideas out as they come to me. Having three classes of followability - user, collection, and tag - seems like a bad idea, but it allows massive flexibility. For one, you wouldn't be limited to collections that the user creates - you could create your own collection and curate multiple @user.tags.
You could also set tags of interest on your profile as part of initialization, like how you select your subreddits when you create your account and curate on an ongoing basis. If you list your interests as .diy .woodworking .powertools .crafting .homerepair and so on, Twitter could automatically prioritize those tags (and related topics, algorithmically) in your curated timeline.
Do I think that Twitter is going to do this? Absolutely not. It's clunky and it would confuse users way more than it would help them. But it's one way of looking at a problem that exists within the construct of the network itself, and how to architect around interest graphs instead of people.
People graphs make more sense in networks like Facebook and Google+, where messaging and communication (Messenger, Hangouts, Groups) are core concepts. Interest graphs make a lot more sense for a network like Twitter, and they really should have implemented more to amplify that a long time ago.
