The New York Times is using machine learning to serve readers a personalized newsletter containing stories they “might have missed.”
"The
Times publishes countless stories across news, opinion and features every day, and we're focused on finding the best ways to introduce readers to the breadth of our journalism,"
stated Alex Rainert, head of product and design at
The New York Times.
Delivered every Friday, the newsletter — “Your Weekly Edition” — will be
curated by editors with the help of machine-learning algorithms designed by
The New York Times.
The
Times is calling the newsletter “experimental,”
adding the technology and design behind “Your Weekly Edition” will continue to improve.
Newsletter editors will sort editorial into three categories. “News you might have
missed,” “Best of The Times,” which will contain the week’s most important news, analysis and features, including video, graphics and other media, and “Discover,”
with highlights from the features well, with stories from travel, food, health and other lifestyle sections.
The Times will algorithmically select the articles
it hopes will be most interesting to each individual reader, based on that user’s reading history on its site and apps, and what other readers have found interesting.
“We track
article popularity by using another algorithm that does not rely on personalization. We do this by recording how often stories appear on our site, and then how many readers click to read them. But
those data are aggregated and therefore anonymized,” the company said.
For example, a user may regularly read movie stories and reviews written by the Times' Culture
reporters. That user's "Your Weekly Edition" would identify and display articles the user has not already read but may be interested in reading.
The Times says “Your
Weekly Edition” is “the first of what we hope are many experiments using personalization technology at The Times.”
Last month,
The New York Times introduced a pop-up newsletter series with ideas of what to eat, drink, see and do in New York City each weekend, called Summer in the City.
Is this really a step ahead with " machine learning," or another use of standard analytics data to improve the likelihood readers will be drawn and stay with the stories?
Will be interesting in monetizing this percieved level of higher emailed personalization with ads...and the usual engagement metrics...if that works. They they will have a "best of the TIMES" content that spews a nice revenue gusher of highly targeted activity via its readers Now, if they can externally track from NYT site, meaning I view energy news on FT.com..but not NYT..its another way to divert competitor's readers for specfiic coverage..and to try the NYT, e.g. I'll read the times for domestic economics, but not international, that's with The Economist..so how are these new NYT emails going to gradually validate "hey, we at the NYT also have compelling coverage on this energy topic" and with that show "peronalized" energyb content the subscriber hasn't seen YET on the NYT site...but now that subscriber may be motivated.