Twitter  Quartz found an unlikely inspiration for its relaunched homepage: The email newsletter.  
Nieman Journalism Lab
Pushing to the future of journalism — A project of the Nieman Foundation at Harvard

nytimes-logoDiscussions about the future of news often feature a dystopian, filter-bubbly argument: Let people personalize the news they consume and it’ll kill the common conversation essential to democracy.

I’ve always been suspicious of that argument. (That “common conversation” was always a bit of a charade, and old newspaper monopolies can’t be simply willed back into existence by trying to guilt-trip readers who’ve found options they prefer.)

But whatever your feelings about it, the reality is that very few news organizations have invested in technology that would allow for substantially different presentation of the news from person to person. If my neighbor goes to, he’ll see the same webpage I do; if my mother goes to, it’ll be the same page I see. (Some sites do minor shifts based on geography — local audiences vs. national audiences, for instance — but that’s still a long way from true personalization.)

The New York Times is one of the few major news organizations that’s invested in a recommendations engine that attempts to figure out what stories you, as an individual, might be most interested in. Today, they pushed a set of improvements to that engine, with more to come:

So what’s changed?

They’re not kidding about the real-time aspect; my recommendations have changed several times in just the past few minutes. (Although it keeps insisting I need to read this Miley Cyrus article.)

I’ve always found the Times’ recommendations engine to be quite good, and the early reviews of this edition are solid:

The Times is well positioned for an investment in personalization. First off, it produces an enormous amount of content — hundreds of stories, blog posts, videos, and slideshows each day. If you’re limiting your recommendations to recent content, you need to be producing a lot of it every day for a personalization filter to make any sense.

Second, the Times’ metered paywall approach — 10 free articles a month, payment required after that — means that getting another click can mean a lot more than earning another $0.004 in advertising revenue. The meter incentivizes the paper to do whatever it can to push a marginal reader’s story count higher. If you can figure out what a reader wants — and you can make those recommendations prominent, as the Times does by putting them in the sidebar of article pages — maybe you can turn a few 8-article-a-month (free) types into 15-article-a-month (paying) types. (In the beta version of article pages, recommendations aren’t nearly as prominent — but that could obviously change before the site redesign launches.)

My big question is when recommendations will break out of their shell and become more prominent on the front page of They’re there now, but several screenfuls down, beneath dozens of other links. When the layout and selection of stories at the top of the front page starts to be influenced by personal recommendations — when, say, 2 of the 15 articles on the top of my version of are different from your version — that’ll be a milestone for the algorithm. (Or the death knell of democracy, if you argue the other side.) There’s a lot of newspaper tradition arguing against that sort of personalization; I’m hoping the Times can allow itself to benefit a bit more from a pretty powerful tool.

Also, Jacob Harris has an interesting idea:

Update: This isn’t strictly related to the recommendations engine, but I also noticed that the Times is now available for programming on IFTTT, the web-automation tool popular among a certain subspecies of nerds. So you can, for example, now get an automatic email when a New York Times Magazine becomes very popular, or save all Well stories to Instapaper, or get an SMS when your company is mentioned in a Times story.

Will IFTTT generate earthshaking pageviews for the Times? Highly doubtful. But it’s another case, like the recommendations engine, of the Times (a) differentiating itself from its peers and (b) using technology to make the Times more useful to its readers. Those are wins.

— Joshua Benton
What to read next
Ken Doctor    Aug. 25, 2014
America’s largest newspaper company says it’s building for the future. But it’s hurting its own value proposition in the process.