The New York Times aims to predict which stories will do well on social media with a bot that runs within popular messaging app Slack.
This new tool came about as a result of the publication’s need to sift through over 300 new stories each day to decide which ones to publish on social media.
Of the 300 or so stories NYT publishes each day, only about 50 end up being shared on social media. The new bot, called Blossom, is designed to analyze a selection of articles and pick out which ones will do best on social media.
Using Slack, sending a direct message to Blossom with commands such as “Blossom Facebook?” will return the bot’s prediction about which stories to post.
Blossom can even send alerts if it detects a recently published post has the potential to go viral.
What functions like a simple chat bot is said to be made up of advanced machine learning, involving programming involving Java, Python, and MapReduce.
According to Blossom’s developer, stories recommended by the bot have been getting 120% more clicks on Facebook than stories the bot didn’t recommend.
The bot’s impressive ability to predict successful social media posts is cutting edge for a company like New York Times — company that has a well-documented history of being slow to embrace digital technology.
Facing declining revenues as a result of being late to adopt a digital strategy, this may be a sign that NYT is now headed in the right direction when it comes to digital.
Does Blossom Have Potential?
For the time being, Blossom is being developed to be used internally by the New York Time’s editorial staff. There’s currently no talk about a wider release of this intelligent bot, but you can’t rule out the possibility either.
Could such a thing take off and lead to more social content being aggregated by bots?
I knew SEJ’s Social Media Strategist, Brent Csutoras, also a known Futurist, would have the right insight into this kind of topic.
“The idea of using algorithms and machine learning is nothing really new, as sites like the current Digg rely on similar concepts to populate their site. It is something new to see publishers focusing on this technology internally, especially a company like NYT.
We [Brent’s company Kairay Media] developed a tool internally to identify old content based on current social trends a couple years ago, which helped a lot for promoting content you might have forgotten about at a time it could perform its best.
The implications of what NYT is doing could change the publishing world completely.
I would love to see this catch on and 3rd party tools start jumping up to allow more publishers and sites to take advantage of this strategy.”
Content publishers of all types could stand to benefit from this bot if it were to become more widely released. It could also be a way for New York Times to start building a new reputation as a leader in digital publishing.
Sounds like something to keep on the radar in the event that it does get a public release.