Blog Post

The NYT API: Newspaper as Platform

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

There’s been a lot of chatter about the newspaper industry in recent weeks — about whether newspaper companies should find something like iTunes (s aapl), or use micropayments as a way to charge people for the news, or sue Google (s goog), or all of the above — and how journalism is at risk because newspapers are dying. But there’s been very little discussion about something that has the potential to fundamentally change the way that newspapers function (or at least one newspaper in particular), and that is the release of the New York Times’ (s nyt) open API for news stories. The Times has talked about this project since last year sometime, and it has finally happened; as developer Derek Gottfrid describes on the Open blog, programmers and developers can now easily access 2.8 million news articles going back to 1981 (although they are only free back to 1987) and sort them based on 28 different tags, keywords and fields.

It’s possible that this kind of thing escapes the notice of traditional journalists because it involves programming, and terms like API (which stands for “application programming interface”), and is therefore not really journalism-related or even media-related, and can be understood only by nerds and geeks. But if there’s one thing that people like Adrian Holovaty (lead developer of Django and founder of Everyblock) have shown us, it is that broadly speaking, content — including the news — is just data, and if it is properly parsed and indexed it can become something quite incredible: a kind of proto-journalism, that can be formed and shaped in dozens or even hundreds of different ways.

Doing this with all of the various elements of the news — names, places, events, details — on a large enough basis can reveal hidden patterns or connections that might not only improve an existing story but lead to new and completely unexpected ones. At the moment, only the research departments of newspapers have the tools to do this, but opening up an API the way the New York Times has can put those tools into anyone’s hands, allowing them to pursue projects and avenues that newspaper reporters and researchers might never think of. And from the point of view of the Times as a media outlet and business, it turns the paper into a kind of platform for other services and features. That makes the paper and its content more valuable, and could lead to all kinds of commercial licensing possibilities and partnerships — not to mention being good marketing.

This kind of thinking is at the core of Jeff Jarvis’s book “What Would Google Do?” His main point is that virtually any business can benefit from thinking about making its data more open, allowing others to remix and manipulate it to see what comes out, and then taking advantage of what can be learned from those experiments. All the New York Times is doing is using its article database in the same way that Google uses its map database, or the Google Earth satellite-imagery database — as a foundation upon which other things can be built. The Times deserves kudos for pursuing such a open model rather than locking its articles up and trying to charge people for every view. I have no doubt that they will benefit far more from such an approach in the long run than would ever be possible with a pay-per-view strategy.

33 Responses to “The NYT API: Newspaper as Platform”

  1. Daylife ( has already done the “News as API” experience, and it is very impressive. NYTimes is an investor, and Jeff Jarvis worked with them in the early days (I don’t know his relationship with them now). Worth taking a look; the site itself is just a hint of what the have created for HuffPo, USA Today, and other information sources.

  2. Scott, sorry to correct you, but none of the APIs are “mashery made.” Mashery is indeed providing the key generation and request throttling service they’re known for, but the actual underlying API code is implemented by and supported by developers at the New York Times.

  3. I can’t resist the Dilbert take on this:

    In all seriousness, I think it’s great that the NYT is taking this initiative, and I’m looking forward to Times Open in a couple of weeks.

    As for Jeff Jarvis’s “What Would Google Do?” book, I encourage you to read my recent blog post:

  4. This is a great development; and the combination of open access to some of the best news and editorial coverage in the world along with the NYT brand is sure to spawn innovative new business applications (that NYT should benefit from).

    Interestingly, the NYT’s own Brad Stone wrote an article about major brands opening up APIs in September: indicating that the Times had been working on this since last May.

  5. Manju Mahishi

    NY Times certainly deserves a lot of praise for this effort.

    I fully agree that this open model has the potential to spawn many interesting and useful applications – while being a terrific marketing vehicle for NYT.

    Thanks for sharing this Matthew.

    -Manju Mahishi

  6. Alex Beamish

    This is pretty amazing — turning a newspaper into an application service. But I’m a little surprised that it’s taken this venerable newspaper so long.

    Someone in the NY Times seems to have been thoughtful enough that they permitted Perl developers to work on and donate work done on the Perl module Devel::NYTProf ( back to the community. Yet it sounds like all upper management could do about their decline of dead tree readership was to blame the Internet.

    It’s a business approach that seems weird, yet Google didn’t really have much of a solid business plan when they started — AdWords just tipped up as a Good Idea, and is one that now earns them piles of dough. I think the NY Times is going to see a very good return on this move, both financially and mood-wise.

    To quote a good friend of mine, “This is going to be huge.”