The big game-changer (or alternately, the potential disaster) that came out of Twitter’s first big conference this week was Annotations. It’s a new way for developers to attach any kind of metadata to tweets. Twitter has long supported certain kinds of metadata like geolocation or what service the tweet was sent from, whether it was the web site or a Twitter client like Brizzly.
The new annotations open the possibility for making tweets and search much richer. One could search by tweet type, looking for all song tweets or tweets containing videos.
As one founder said to me, “It’s the most disruptive thing the company has done in two years.” It could also be a huge mess — if one developer adopts one type of metadata, other applications will have to recognize it to make it useful.
So what can happen now? Here are a few ideas I’ve heard people talking about through the week:
Media attachments: The basic stuff. Labeling tweets that have links to photos or videos in them. Or sticking the links to the photos and videos in the metadata itself. Or tagging the content of the photos in the tweets.
Hashtags: Migrate them to the allotted space for metadata.
Reviews: Add one to five-star ratings to tweets.
Finance: Make it easier to follow stocks for spikes or sudden fluctuations with a special category of tweets. In fact, make a special type of price alert tweet for all sorts of goods and services like plane flights and oil.
Coupons: Make a special category of coupon tweets so you know when hamburgers are half off in the neighborhood. Location data would make this especially powerful.
Music: Mark all tweets that refer to songs. Create Twitter song charts that span different services like Blip.fm and Hype Machine. Generate playlists from Twitter. Granted, these ideas are already out there. But people might be hesitant to tweet their favorite songs for fear of spamming their followers. Tagging the tweets would allow people to opt-in to ‘music-related’ content from friends.
Location: Yes, there is already location metadata. But there are different ways of thinking about places outside of lat-long coordinates or place names from local listings. What about all tweets about parks? Or tweets about schools? Traffic tweets? Finally tagging all the tweets that are check-ins?
News: A news story’s dateline (when a location is written in all caps at the beginning of story) has been a key piece of metadata paired along with copy for more than a century in newspapers. Now it can be put into the metadata of a tweeted headline. This solves a big problem: media organizations often tweet the headlines of stories about topics like French politics or airplanes being grounded in Europe because of the Icelandic volcano eruption. But these tweets may not actually contain the words ‘FRANCE’ or ‘ICELAND’ or ‘EUROPE,’ so they won’t get picked up by a cursory search. One could also tag tweets that are about the same ongoing news story like ‘michael-jackson’ in the case of the pop singer’s death.
Gaming: Now that Facebook’s notifications between game developers and users are fritzed, perhaps there’s finally a non-intrusive way of building them on Twitter. Part of the reason games never took off on the Twitter platform is that people were afraid of spamming their followers. But if Twitter developers came up with a common set of metadata for marking gameplay, Twitter clients like Seesmic, Tweetdeck and Brizzly could design around that and pull them out of the main feed. They could create special “gaming” columns where you’d learn when it’s your turn in a game. The gaming experience wouldn’t be on Twitter itself, but rather Twitter would be used to alert and lure players back to a game on an external site. This could help larger companies wean themselves off the Facebook platform and build their own properties. Granted, it still requires genuinely engaging games, because Twitter users would have to opt-in to follow the gameplay of their friends.
New commenting systems: Provided the size limit is large enough, you could put URLs in the metadata. You could call all the comments or tweets that reference that specific URL. There are services that do something like this already, but they generally attract retweets and not original responses. I’m not sure exactly how this might look or feel different from what already exists out there, but it could be powerful.