Nieman Foundation at Harvard
Class is still a taboo topic in the U.S. The Guardian’s ambitious new rural reporting projects are tackling it
ABOUT                    SUBSCRIBE
April 18, 2011, 11 a.m.

Wisdom of the (developer) crowd: Key lessons from news organizations using open APIs to ramp up R&D

Editor’s Note: At the International Symposium on Online Journalism earlier this month, one of the most interesting papers was by our own Seth Lewis, who’s now teaching at the University of Minnesota, and Tanja Aitamurto, a Finnish Ph.D student working as a visiting researcher at Stanford.

Their subject was the ways in which top news organizations — like The New York Times, NPR, and the Guardian — have created and used their own application programming interfaces (APIs) to work with outside developers. What models have they followed, and how have they worked out? Tanja and Seth have written up a summary of their findings for the Lab; you can read the entire paper here. I’ve also embedded their slide deck below.

Among its other troubles these days, the news industry has an R&D problem. The work of digital journalism used to be pretty straightforward: serve up content for websites that were viewed on personal computers (so quaint!). But now there are multiple platforms, multiple devices on those platforms, and multiple operating systems for those devices — altogether making the work of online journalism far more complex. This is especially true given the rapidly shrinking life cycles for products, and the speed with which new things emerge and become standardized (think: the iPad, barely a year old but seemingly ancient in Internet time).

How do news organizations keep up? Amid declining resources, how do they accelerate R&D, better serving their customers via product development — while at the same time not increasing costs but, in fact, finding new revenue opportunities to offset declines in legacy operations?

One possible solution has begun to emerge: application programming interfaces, or APIs. (Not sure what APIs are? As Will Sullivan nicely put it: “My brief ‘how to explain it to my mom’ summary is: APIs are like RSS feeds on steroids.”) At least for a few of the largest news organizations — the ones with the human/financial capital to invest in this kind of thing — APIs have begun to bear fruit as a revenue-generating R&D engine: catalyzing innovation both inside and outside the organization, giving new life to old content for revenue purposes, and creating opportunities for leveraging brands and building networks of developers across the web.

We’re talking specifically about “open” (or public) APIs, those that are intended to be picked up and used by outside developers. This is common practice in the tech industry, where Facebook, Twitter, Google, and more offer APIs that let developers build applications around their data (though, it should be noted, there can also be a shift away from openness, as in the recent clampdown on developers using the Twitter and Android platforms).

APIs are not exactly new to the news industry — NPR launched what is believed to be the first in this space in 2008 — but only in the past year or so have APIs become hot-hot in online journalism. Sullivan described them as the “big winner” at ONA 2010, potentially marking a coming “Age of APIs.” In the news sphere, the four biggest players with open APIs are NPR, The New York Times, the Guardian, and USA Today. (See previous Lab coverage of these and other APIs for news.)

As part of our academic research on innovation and journalism, we examined these four cases via interviews with key developers at each organization, seeking to identify the benefits, challenges, and opportunities of open API initiatives for news organizations. (Our research is ongoing, so we’d like to hear from other cases of APIs at news organizations; you can reach us at and

At the four news organizations we studied, their open API initiatives differ in offerings and conditions. For example, via its open API the Guardian gives wide access to its content, providing full-text access to more than a million articles from the past decade and beyond. By contrast, The New York Times eschews full text — not surprising given its pay-model approach — and instead delivers links back to

Despite such differences, though, some definite commonalities emerged among the organizations and their experiences with open APIs. Here are five key lessons:

1. Open APIs can serve as external R&D labs

News organizations are acknowledging and tapping into collective intelligence in their R&D by implementing open APIs as a part of their product development strategy. This becomes something of an outside R&D test lab: When external developers experiment with and build on news organizations’ content (and not just news content, but also information such as bestselling books, as in this API offered by USA Today), this accelerates the R&D process, as the news organizations can draw on existing experiments (outside) to help frame their own exploration (inside).

This is especially useful at a time when media organizations (of all kinds, not just news) are trying to speed up R&D in order to reach customers on a growing variety of devices and operating systems. As one example: using the NPR open API, a Google engineer voluntarily developed an NPR application for the Android platform, and another external collaborator built an audio player for Unix, a rather marginal platform in terms of user volume. Also, the first iPhone application for NPR content was launched by an external developer using the API. When that app (NPR Addict) quickly became popular, that got the attention of NPR’s in-house product development team and spurred them develop the official NPR app.

When external developers experiment with the content, the process serves as a kind of a market research tool for news organizations, helping them identify opportunities to “give new life to old content” — i.e., leverage existing information across new (and potentially revenue-generating) channels, and perhaps in ways that the organizations themselves might not have imagined. As Tim Carlson, director of mobile development for USA Today, described it:

Is there a demand out there for our bestselling books beyond sort of the website where it’s always lived? I mean, we’re basically giving new life to this content, and we want to see if people out there bite and are interested in it.

When news organizations embrace the flow of ideas inbound to the organization as well as outbound from the company, they are actually implementing one of the main principles of open innovation in their product development strategy. Open innovation, a concept coined by Henry Chesbrough, has become popular in the academic literature on management and technology; in our paper, we argue that open APIs can be understood as one of the first manifestations of open innovation strategy in the news industry.

2. Open APIs present a revenue opportunity

News organizations are not introducing APIs simply for goodwill: There is a deliberate business strategy behind this deployment, and open APIs are providing new ways to commercialize news as a product.

There are two approaches to gaining direct revenue via open APIs: First, by letting collaborators use the content for free but expecting them to take the news organization’s advertising key with them. In the case of the Guardian, this means that when an external developer builds a mobile app using its API, the application must present advertising from the Guardian ad network — though the Guardian also allows republishers to present their own ads. The second method for direct revenue via open API-built products is to charge a licensing fee for the content. The Guardian and USA Today both use this method.

Apart from direct revenue streams in the form of advertising and licensing fees, news organizations can find other, more indirect ways to strengthen the value of their products through an open business strategy. One of them is to drive traffic to their website, as each of these organization tries to do via its API. The Times is a good example: the content delivered via its API has links back to that must be displayed.

The Guardian uses a similar approach in its three-tiered business strategy: the news organization allows collaborators to access Guardian headlines, but not the body copy of the article. At this level, the developer can publish Guardian headlines and metadata for free, and keep the advertising revenue.

To be sure, open API initiatives at news organizations are at an early stage, and aren’t generating huge revenue (and may never). However, they’re a good example of monetizing existing content via a self-serve system that doesn’t require too much extra work for the news organization.

3. Open APIs help weave news organizations into the fabric of the web

When collaborators can interact with and build on content, that content spreads more widely on the web, reaching more people and niche groups of users. This is a crucial part of a larger transformation from news site to news-and-information platform. Matt McAlister, who leads the developer network at the Guardian, describes the thinking behind the Guardian open API:

We came up with this kind of big, broad statement about weaving the Guardian into the fabric of the Internet. It was a realization that we needed to be a part of the Internet and not just on the Internet.

This is a fundamental realization for news organizations, and one that marks a big shift from traditional, print-era thinking in which benefits come from controlling the content rather than “letting the content go.” But these news organizations are smart: They are ceding control in a structured way, and with a business logic attached.

And that is where this becomes interesting: However contradictory it may sound, by opening up their content for public use, news organizations (in their view) are actually gaining more control over that same content. As Daniel Jacobson, who designed the architecture of the NPR API and now directs API engineering at Netflix, says:

When you let people to use the content and attribute it, they treat it respectably. The API actually gives you an opportunity to control that in some ways that the website doesn’t, and in fact the people who use the API have more respect for it anyways.

4. Open APIs create an ecosystem effect

Having an open API has proven to be a “good conversation starter with developers” — like a coffee-table item that has spurred further conversations and collaborations with the likes of Google as well as a host of independent developers. Having an API is “good street cred,” as one put it.

This has been a welcome side effect of the API initiatives: Opening up their content has allowed news organizations to attract crowds of developers and startups interested in using high-quality content in their applications (amid the surfeit of low-quality content found elsewhere). The Guardian alone has 3,000 developers around its Open Platform, which includes the open API initiatives and its Data Store. This cluster of collaborators creates an ecosystem effect of reciprocal change, nurtured by meetups and hack days with with developers in the community. Such community building has resulted in business opportunities and partnerships, and built an extended network of brainpower around the news organizations.

5. Open APIs can be a challenge, especially within the news organization

Lest it seem all rosy and risk-free, deploying open APIs at news organizations comes with its own set of challenges. We found that the toughest problems for these news organizations weren’t technological, but cultural; existing technologies could be retooled relatively easily compared to the hard work of getting buy-in from top management, which generally is nervous about “giving away” content in any fashion. In one case, a developer found that he could get his superiors on board by “de-emphasizing the open side of things” and pitching the API initiative as “Business Development 2.0” — speaking the language that news executives like to hear.

There are other internal challenges as well: the fuzzy legal implications associated with rights management in database-driven applications, as well as the difficulty of synchronizing the structure of APIs (variable-driven databases) with the basic unit of journalism (narrative accounts). Whereas the latter are intended for linear (human) consumption, the former are designed for nonlinear (machine-driven) engagement, and that presents a problem when, say, you have to pull apart a roundup story to fit the parameters of a movie reviews API.

Looking externally, the main challenge is preparing for the inevitable uses of the API that will not be to the liking of the news organization — and yet must be negotiated carefully, so as to maintain a consistent and close connection with the developer community. As one interviewee described it:

There will be someone who’s going to be very noisy, who’s going to do something with our content that makes us very, very uncomfortable, and we’re going to have to decide, ‘Do we turn off their [API] key, or do we let it go and let it be?’ Our default is to let it go at the moment. We haven’t seen anyone do anything that we felt was damaging to our brand.

Conclusion: Open APIs are not transformational, but still significant

Open APIs alone aren’t the solution to the news industry’s R&D problem, nor will they somehow “fix” a broken business model. However, the open API phenomenon is an important step toward a form of online publishing that is more in sync with the ethic of the web (i.e., of the web, not just on the web); and, along the way, it takes advantage of the wisdom of the (developer) crowd to accelerate the work of R&D, far beyond what a news organization could accomplish working by itself. In this sense, public API initiatives offer a first glimpse of what the open innovation paradigm could bring to the news business, with a range of new benefits and opportunities — for revenue, partnerships, and more.

As part of our ongoing research on open APIs, we’re looking to gather all possible information on APIs, future opportunities, and latest developments (in journalism and elsewhere). Please reach out to us at and Thanks!

Image is an excerpt of a Jer Thorp visualization of usage of the words “Iran” and “Iraq” in The New York Times since 1981, built using the Times’ article search API. Used under a Creative Commons license.

POSTED     April 18, 2011, 11 a.m.
Show comments  
Show tags
Join the 35,000 who get the freshest future-of-journalism news in our daily email.
Class is still a taboo topic in the U.S. The Guardian’s ambitious new rural reporting projects are tackling it
“We want to support people who actually live in these places reporting on their own states, about inequality, and then we want to bring them to traditional elite audiences.”
Do you trust the news, or do you trust your news? In the U.S., there’s a huge gap between the two
Plus: A bill to outlaw fake news in the Philippines, and the question of whether real news outlets should cover fake news.
Vox’s healthcare newsletter (with ads sold out) is filling a role beyond “articles on the Internet”
“I’m keeping in mind that there are actually people reading these stories who are relying on us for information.”