APIs: The new distribution

The Guardian just announced that it is releasing all its content through an API as well as making available many different data sets through a data store, all of which can be mashed up into others’ sites and applications. They join other organizations – the BBC, National Public Radio, and The New York Times – in releasing APIs; notes that it’s the creme of news that sees the wisdom in APIs. The Guardian’s offers more than headlines: articles, video, galleries, everything. It also adds one more important element to its offering: a business model, creating an ad network for users of the API.

Upendra Shardanand, my partner at and the founder of Daylife, has been saying for a few years that APIs are the future of distribution. The Guardian says its API will put its content “into the fabric of the internet.”

The moment that led to the title of my book – when I told publishers to ask what Google would do – came when I was trying to convince a roomful of them to think distributed, to stop believing that their brands were magnets sufficient to attract their entire audience, to go to where the people are, like Google.

The reflex of publishers – the few who agreed – was to think of distribution in terms of widgets (the trend that never was). They also produced RSS feeds, though limited. Note that a few weeks ago, the Guardian also shifted to full-text feeds (I asked about the business impact of this and they told me that there seemed to be none as traffic continues to rise).

Now APIs take distribution to its logical – if unknown and sometimes frightening – limit. Now I could build an application around the news of at least these four outlets – and, with Daylife, headlines from and analysis from thousands more. In the case of a Guardian story, you may read it via my application and not go to the Guardian’s site. Isn’t that insanity? Isn’t that what publishers are complaining about with aggregators? Actually, no, aggregators display only headlines and give links; this is worse if you’re trying to protect your content and traffic to your site. But that’s the old, centralized mediamind way to think. In the new, distributed world, you want to be where the people are. The frightening part has been that once you release your content as data, you lose control of the display, branding, data collection, and revenue. That’s why the Guardian is trying to add its advertising business model – because it wants to release everything; it wants its content to be used all around the web. This is the new distribution.

News organizations already lost control of packaging, whether they all knew it or not, when most of us most days come to content not through carefully designed home pages but through search and links and now Facebook. The media brand is less a destination and a magnet to draw people there than a label once you’ve found the content, wherever and however you found it. So the more places you can find it, the better.

: Disclosures: I should have added that I write and consult for the Guardian and, again, I’m a a partner at Daylife.

  • Cranston P. Hughs

    Another stellar example of primary journalism, Jeff. Nice interview quotes–truly original content.

    • What’s your problem? I interviewed the creator of the API last week (the news was embargoed); the piece to which I link does a better job than I would have done; no sense repeating when one can link, eh? In any case, I’m not saying this is original reporting. I’m commenting on the news. Blogs do that sometimes, you know.

      • J. Blare

        Jervis is right. Why duplicate effort? Doing old school reporting is a waste of time and resources. It’s much easier just to recycle what’s already out there–it’s convenient and you needn’t ever leave the comfort of your couch.

      • Mike G

        What’s an API? I followed the link but still didn’t see. You kids and your new technology!

        • Sorry, Mike. It’s an applied programming interface: the key to the kingdom that lets someone get at your data and program an application around it.

        • I think you invented a new term there. Applied.. I actually like it..

          API = Application Programming Interface


          It’s basically the mechanism by which you access data and services through software.

          With the Guardian, some of the API is based on commonly used data structures and constructs to access the services, but the Guardian has it’s own specific API based on;

          a) how it wants you to find the data/article you are looking for, and
          b) the format of the data/article and associated images/interactive content.

  • Pingback: NewsCred Blog » Blog Archive » The Guardian Launches Open Platform: Why This is Really Big()

  • Pingback: The Guardian Newspaper launches its own API (And Why this is Exciting) | native()

  • “Now I could build an application around the news of at least these four outlets” – The New York Times doesn’t have the full text of the articles in their API. It’s still a big step for them, but not the type of paradigm shift we have with the Guardian opening up their API. They provide the full text AND encourage commercial use, which is forbidden in the NYT Terms.

    Despite those details, these are exciting times as news organizations are finally starting to “get” the new, open web. Here’s to a thousand business models blooming…

  • Lyle

    Great to see this Jeff. Opens up all sorts of opportunities for peer distribution of news. Enjoyed watching and commenting on The Future of News yesterday. Lecture theatre is not the best places to pull that off. Botanical gardens with participants in a circle better. But then the product manager from google head office would have looked as if he was from the NSA wouldn’t he? Haha. Great job anyway.

  • Pingback: Webbmedier väljer: öppet eller stängt « David Hylander()

  • You didn’t mention the Reuters API.

    But I totally disagree with this idea of the same content being used all around the web. This is diametrically opposed, Jeff, to your notion of do your best and link to the rest. Is it not?

    When full content is being used in multiple places around the web, how do aggregators (like Google News, etc.) get rid of duplicates? Where will all the Googlejuice go? What chance is there for the semantic web to settle on a “canonical URL” for the content?

    I think the Guardian’s full-content API plan is a horrifically bad idea. A step backwards into old printie AP-style distribution model.

    • Supporting the semantic web would be easy for the Guardian, simply make the API RDF-based. I think they missed a trick here, it would help with the issue you raised of needing canonical URLs and resolving duplicates.

      Looking at the Guardian API, it can contain linkbacks to guardian.co.uk and according to the T&Cs you must provide the link. The googlejuice is still there.

      I don’t think this is a step back, why did the AP become the largest news organisation? because of it’s syndication model. The AP model works even better when there are even more places for your content to go.

      If AP opened up a usable/searchable API, and added an ad model instead of paid model, it would dominate internationally I am sure. The content could be hyper-distributed, free of existing barriers, which are primarily related to a) finding the content you want, b) contracting with the AP and c) paying for content upfront, regardless of performance.

      • The AP had a great model for print, sure. For the web, it’s completely wrong.

        I completely agree, though, that if AP opened up it could kick serious ass.

  • Pingback: DYSPEPSIA GENERATION » Blog Archive » APIs: The new distribution()

  • Pingback: Dysart & Jones » Content: Into the Fabric of the Net()

  • Pingback: Marc’s Voice » Blog Archive » 2nd week of March ‘09 blogging()

  • Pingback: Another Newspaper offers an API | Reinventing Yourself...()

  • Pingback: APIs: The new distribution « BuzzMachine « Spill The Coffee()

  • Pingback: The Guardian Launches API: News as Platform()

  • Pingback: bwl zwei null · Der Guardian surft ganz oben: Mit offenen Schnittstellen in die Zukunft()

  • Steve – UK

    Content APIs where the originator maintains at least some control of the advertising, or a share of the customer spend, sound like nirvana for dying newspapers. But I doubt a single newspaper alone will achieve sufficient scale to make this commercially viable. The obvious way to go would be a federated content platform managing distribution for member contributors ranging from bloggers to the Guardian. It would also give content creators a much stronger negotiating position with the search engines.

  • Pingback: Guardian’s Open Platform - some thoughts from the blogs | Journalism.co.uk Editors' Blog()

  • K. Trot

    @ Steve-UK

    The only business model that can succeed is an impossibility at this point because you can’t compete with Free. News Organizations should never have given their content away in the first place. I realize this is an unpopular opinion among readers of this blog but the truth is that it takes money to gather news. If we lived in a cooperative society, Free might work. Unfortunately, we don’t. Our economy is driven by competition. These are not my rules, and I’m far from a fan, but it’s what we have. Giving stuff away for free doesn’t work for anyone. The model that Jarvis proposes where news gathering organizations can support themselves through advertising revenue has proved false time and again as can be seen each day with reports of news organizations closing. Every time he is asked why things haven’t worked out the way he predicted, he back pedals and blames news orgs for moving too slowly, or just not getting it.

    These days, more than anything, Jarvis has found success in sales, by convincing newspapers to part with their intellectual property so it can be repurposed through companies like Daylife. Ask him if Daylife–the company with whom he is a partner–will part with their code and algorithms and you will get a resounding NO. Shouldn’t Daylife’s platform be an open source platform? This is not in the spirit of a truly Open internet now is it?

    The irony of all this, is that the API model can only succeed if news organizations succeed. As news orgs fail, there will be less content to be purloined, and repurposed. Until then, the coders and algorithm heads, will make a small fortune selling APIs and then move on to the next industry.

    • Tom in KC


      You paint a bleak picture of change as doom. Though the changes are painful, I think we are seeing a process of creative destruction as the news business evolves.

      Here’s a less drastic way to look at the API model: the Guardian is inviting programmer/entrepreneurs to become the carrier/delivery agents for their nonphysical product.

      In an API model, publications would be smaller and more focused on their core functions as news orgs. To me, that does not sound like a formula for doom. (Unless you work in the printing plant.)

      The problems and opportunities of “free” are faced by software developers, too. Many of them create free applications. Many others compete against those free ones with commercial applications. And many offer BOTH a free and paid version of the same product as a marketing technique. The API model allows these developers to apply their intellects to the problems and opportunities of “free”. Seems to me news orgs need that kind of help if they are to adapt to the network economy.

      — Tom B.

      • K. Trot

        Hi Tom,

        Never said it was a formula for doom. I was actually responding to Steve-Uk’s comments as to whether APIs improve the commercial viability of these news orgs.

        What is needed is a business model that allows everyone to prosper. If you expect to be a commercially viable enterprise, you can’t give things away for free and expect to survive, at least in a competitive free-market.

        I just thought it was interesting that Jarvis–a major proponent of the Freetard movement–consulted with these news orgs and convinced them they should make their content available for free. Then, he turns around and partners with a company that repackages and distributes that content. Problem is, these companies are all failing, precisely because they made their content available for free. You’ve got to give Jarvis credit, though. He’s a brilliant businessman.

        -K. Trot

  • Tree Frog

    This is making its rounds, but David Simon wrote a great piece about investigative reporting in Baltimore:


  • Ian

    Jeff, I was interested in your comment about the Guardian’s move to full text feeds. I used to subscribe to a number of Guardian feeds – via Bloglines. Only Media Guardian and Sport are left now and I leave these until I’ve checked all my other feeds since they are so painfully slow to browse through and quite frequently crash my browser. I’ve tried contacting the Guardian via their website and have even posted a comment on the Media Guardian facebook wall to see if anyone can advise how to improve the performance. (The only other feed among the 42 I subscribe to which performs anywhere near as badly is MoCoNews – which also offers full text feeds and was acquired by GMG last year!)

    I’m really disappointed that nobody at the Guardian has felt it worthwhile responding to my pleas for help. Not very web2.0, more like DellHell 2.0!

  • Pingback: Blogonomics: APIs | Refinancing()

  • Pingback: Die Revolution der Netzinhalte - Thesen, Links und Zitate | lab()

  • Pingback: More coverage of the Guardian’s Open Platform API - paulcarvill.com()

  • Pingback: The Guardian of the Future » Awake in This Life()

  • API definately means API = Application Programming Interface.

    May be ur defining it differently.

  • Pingback:   Guardian goes data crazy by andydickinson.net()

  • Simon Bucks

    Meanwhile the Manchester Evening News , part of the Guardian group, has sacked 150 journalists and centralised its news gathering into the Manchester office. In the future will there be less and less original on- the-ground reporting and more recycled content?

  • Pingback: Printed Matters » New Guardian API - a quibble()

  • Pingback: jardenberg kommenterar - 2009-03-13 — jardenberg unedited()

  • Pingback: Concept of the Week: Pareto Improvements | brianfrank.ca()

  • Pingback: LSDI : Guardian: un nuovo sistema di distribuzione dei contenuti()

  • Pingback: Why Fishwraps are Hooking Less Readers | The Minority Report()

  • Pingback: Il ReteGiornale - la Tua Voce in Rete» Libertà d'informazione » Guardian: un nuovo sistema di distribuzione dei contenuti()

  • Pingback: laOtra TPA » El periodista como red social()

  • Pingback: Footprints (17.03.09) | Chris Deary()

  • Pingback: Matt McAlister » Response to the Open Platform launch()

  • Pingback: New York Times API Recap()

  • Pingback: 1. Solve journalism’s data problem. 2. Kill the AP. 3. Invest in the next market. « BuzzMachine()

  • Pingback: Today Google News Timelines, Tomorrow a Google News API? « Difference Engineering()

  • Pingback: F.A.Z.-Community()

  • Pingback: F.A.Z.-Community()

  • Pingback: Unbundling the Magazine « Ergo McHenceforth()

  • Pingback: Unbundling the Magazine « Ergo McHenceforth()

  • Pingback: Generative Media Networks: Fueling growth through action: What it means for journalism | Matt McAlister()

  • Pingback: Le journalisme des données : Avéyo()

  • Pingback: Tim Berners-Lee: Content wil vrij zijn | Hugo Louter()

  • Pingback: Veritabanı Gazeteciliğinin geçmişi ve gelişimi @arzugokdai @pinardag @dagmedyaveri #verigazeteciliği | Dağ Medya()