Twitter Annotations are going to change your life

While the news of Twitter Annotations has been around for quite a while, it was announced at Chirp in April, outside of a geeky developer audience it hasn’t captured the imagination of the Twitter global echo-chamber. For me the potential of Annotations could mean, for Twitter, this changes everything. Again. At the same time it could lead to a lot of divergence and confusion in the client marketplace.

Twitter Annotations allow metadata to be delivered along with a tweet, that is additional structured data outside of the 140 character limit. Now some Tweet metadata already exists, such as geolocation and a specific tweet you may be replying to and there is also some unstructured user created metadata such as #hashtags and @replies. Twitter Annotations is going to allow developers through the API to deliver any form of structured metadata

How it looks inside the machine

For the technical inclined the Annotation format is going to look something like

[{type1 => {attribute1 => val1, attribute2 => val2, ..... attributen => valn}},

{type2 => {attributen+1 => valn+1 ... }},

......

]

For the not technically inclined, this means that a tweet can contain a number of keys and values, grouped into a type of attribute. For example if you were watching the awesome Argentina v Germany match. You might send a tweet like this, that Tweet may contain an attribute like:

[{"world cup match" => {"round" => "Quarter Final", "home_team" => "Germany", "away_team" => "Argentina", "home_score" => "4", "away_score" => "0", minutes => "89"}]

This information is very unlikely to be entered by the user Tweeting but instead added by the client dynamically depending on the context of how it is being used, much like how a mobile Twiiter application will use GPS to insert your current location into a Tweet at the moment (random statistic as of late last year only 0.23% of tweets were geotagged.) You could easily see how a BBC iPlayer Live Twitter application, or Google TV Twitter application could add that data while watching a match.

How this data is then used depends if the receiving client (or website) knows about that annotation format and what it means. In the Tweet above I added some metadata using the unstructured format that is the #hashtag, by adding #ger Twitter have displayed a little German flag, a nice touch but you could see how this data can be expanded to provide awesome insight and commerce potential. Twitter have said annotation tracking will be supported in the Streaming API and likely in the Search API which means in the example above, a site or app will be able to collect and display all the tweets of people watching that match or TV programme (combine that with some demographics analysis and you could get an awesome contextual advertising)

Changing everything

Now the example above is quite complex but it could be as simple as including metadata for a link to a website article you are Tweeting about, saving those valuable characters and making link shorteners less relevant. It could also include an image or video URL to be shown inline in a Tweet. The possibilities are quite literally limitless.

And it is those limitless possibilities that also cause the problems. For Twitter Annotations to be relevant, they need to be produced by and application and consumed by a Twitter client in the same way. This could lead to a free-for-all of competing standards and markup. Twitter have announced they will do no annotation “validation” on their side, but they will lead by example and have already published some recommendations for recommended types. The annotation structures that “win” will likely be a darwinistic process driven by the crowd depending on usefulness and adoption by popular clients. Twitter will need to get this right for annotations to be successful and plan to publish an Annotations “explorer” with statistics of most used, most adopted and trending annotations to identify “winners”.

What we are likely to see first

Twitter’s recommended types give us an indication of what the first use cases will be:

webpage

The title, image (screenshot?) and url of a webpage a tweet is about will allow for a better link sharing experience and more analytics about popular links. Expect to see this pop-up in your Tweetie soon.

place

We have already seen the introduction of this annotation in Twitter Places which allows you can to tag a Tweet about a place or from a place. The integration of this with Foursquare and Gowalla shows how Annotations are going to allow all kinds of new API remixing.

review

The content, rating and url of webpage you are reviewing in a Tweet could create a new kind of Google PageRank based on crowd user-reviews of the internet. Expect to see this integrate as Firefox and Chrome plugins soon.

song

The title, url, artist, year, genre and album metadata is going to add whole new dimensions to what you are listening to now, new insight to what the world is listening to not to mention driving usage of iTunes and Spotify. Expect to see a Justin Bieber trending annotation soon.

movie

This has title, url, director and year and will undoubtably be linked to IMDB. Expect to see mixed up with review metadata. This might lead to new indications for box office takings.

tv show

With title, url and network. Second-screen Tweeting is about to explode and this annotation could be big. I expect to see lots of variations and additional data here, should be hotly contented.

book

Has ISBN, author and year. Integration with your iPad and Kindle maybe, I don’t expect this to be a hot annotation really, more than likely will be replaced with the webpage annotation pointing to amazon.

product

This has brand, model, price  and category. Not sure about the format of the metadata but you can expect some serious commerce potential here, especially with integration into price comparison engines. A whole new revenue stream for clients and for Twitter.

stock

Has the symbol and price. A replacement for the $ tag used to indicate stocks and prices.

offer

With an end_data, price and discount. A new way for brands and e-commerce to use Twitter. Expect to see “deal” aggregators using this annotation but will it lead to increased loyalty?

topic

Just has a title. A replacement for the # tag.

A new layer of the internet

This is just the early stuff. The open nature of annotations can turn Twitter into more than just a content distribution network but into a new infrastructure layer for the internet, one driven ontop of our social graph. For those techies familiar with OSI 7 layer network models I propose that Layer 8 is the social layer.

Twitter Annotations will be in the API soon, probably within the next few months and then It’s going to be a whole new world of Twitter. Hopefully the move to structure won’t replace one of the things that makes Twitter great, the user-driven need to create order, which can then be subjugated for the pursuit of humour ala #ashtag.

I’ve embedded Raffi Krikorian, one of the tech leads on the Twitter Platform with a sneak peak presentation.

Raffi Krikorian’s Extremely Preliminary Look at Twitter’s Annotations from Farhan Rehman on Vimeo.