Praxis

Om Gives LinkedIN API a "D"

2009-10-28T22:19:21-07:00

In a spot-on post today, Om criticized the LinkedIN API and gave it a "D", for "Disappointing".

I couldn't agree more. In the commentary, Darren nails it on the biggest issue (which I'll expand upon below) and Steve Newman correctly points out that a properly managed API with OAuth can address the privacy concerns.

At Mashery we power dozens of successful APIs. Before we set out to build our first release almost four years ago, we did research that looked at hundreds of APIs and compared those that were "successful" (had a strong ecosystem of apps around it) with those that were not (had few apps or developers). Across the board, the biggest factor that correlated with a lack of success was the lack of the ability to self-provisoin a key at any time, day or night, and get to work developing. Of course that key can be limited in quantity of data, available; we have also seen "provisional" keys limited to accessing sample data or limited to accessing only the developer's own personal info during a test phase. Either way, though, the signup-and-wait barrier is the single biggest factor in the lack of adoption. Click on the "Developers" link in the footer and you are given three widgets as options, plus a tab or link that goes to a form requiring all kinds of info about your intentions before you get a key. I talk to a lot of developers. I wish I had a dollar for every time I have been asked if I can help someone (large or small) get access to the LinkedIn API. There is a massive ecosystem of people out there who want to make LinkedIn more valuable. I look forward to these upcoming announcements, and hope they will include a truly open API

Production and Distribution - The Future of News

2009-08-16T09:23:35-07:00

Over the past few months I've thought a lot about the future of the news business. Mashery has several great customers who provide and distribute news and who have developed great APIs to increase their reach and, ultimately, increase their revenue.

Much has been written about this topic, with some news providers warning us that unless people pay for news content no one will be able to afford to staff a Baghdad bureau, investigative journalism will end, and society will suffer. I'm a huge fan of great journalism. I understand its important role in our world.

But like other media and entertainment, it is clear that the business model has to evolve. I believe it will evolve in a way that in the vast majority of cases explicitly separates news gathering and content creation from distribution.

This trend began a long time ago. The traditional big-city daily paper had bureaus around the world, covered all the local news, and created an inexpensive and efficient distribution channel for local delivery. Over time, it became more efficient to share resources for some of the broader national and international news, and to allow the local experts to share local stories of national interest with papers published elsewhere. Hence the wire services.

Thus the big-city dailies became a mix of internally generated content and syndication of externally generated content. The financial success resulted from the economic benefit of a strong local distribution capabilities coupled with cost sharing on gathering national and international news.

Blogs and citizen journalism notwithstanding, I would argue that the economics of newsgathering have not changed a lot. Good journalism requires talent and experience, and the money to pay for professionals to research, write and edit.

But the economics of distribution have changed dramatically.

Much of the debate on the news business seems to take as axiomatic that production of the news and its distribution must remain vertically integrated for the business to survive. I believe the opposite is the case.

Fred Wilson made an excellent point in a post this morning:

If the front page of NYTimes.com linked to everything interesting on the web instead of just their own stories, they could play the same game. I understand the organization reluctance to do that, but I wonder if they have any other choice.

I totally agree. I think we'll see things evolve in a similar way to the motion picture business:

Originally, there was the "studio system". Studios hired actors, made them stars, hired writers, produced movies, marketed them and distributed them. The studios had very strong brands.
Over time, the production and distribution divisions of the major studios began to separate. Independent producers began to package and produce content which would be distributed by either major studios' distribution arms, or by independent distributors.
"Talent" became free agents. Some would join together on a project-by-project basis, set up their own production companies, or commit to work with a particular distributor or production company in exchange for the financial support needed to conceive and develop projects.
Meanwhile, the studios' distribution divisions use their economies of scale to market and distribute a mix of films produced by their production divisions and films produced by third parties. Some of those third-party films receive production financing from the studios, while others are "picked up" once they're completed
Today, we have a complex ecosystem of big-budget studio-financed films, independent art films, and everything in between. But the era of the vertically integrated studio system ended decades ago.

If the news business is to survive, production and distribution must be decoupled. The strong national brands will likely be able to play on both sides, just as the major studios do, but even there the production and distribution divisions will need to be autonomous.

Meanwhile, we're going to see new business models in which newsgathering and production can be done efficiently and profitably, and distributors will figure out how to make money while providing the necessary economic incentive for the producers.

News producers who are opening APIs and experimenting with a variety of syndication business models are likely to lead the way. Those that aren't are sitting on the sidelines and hoping that when the innovators "crack the code" they'll still be able to jump in and join the party. By then, it will likely be too late.

No one knows the future - that's the point of APIs (or, Teri Takai has it backwards)

2009-05-14T18:34:45-07:00

Thanks to Jen Pahlka for twittering an important (if troubling) comment made by California CIO Teri Takai at an event today:

Asked my own ? to Teri Takai: what are plans for CA to open up data to citizens?

A: we want to know what will provide value b4 we do it.

Teri, you have it backwards. The whole point of opening data through APIs is that you have no idea what will provide value. You have no idea how the data will be used.

APIs and open data allow innovation to occur outside the organization, which is what makes it so important that all levels of our government open as much data as possible and let citizen-developers experiment with it to find opportunities. That is the whole point of transparency in government.

Open data is a necessary ingredient for innovation, transparency, and the creation of all kinds of value that none of us can predict today. If you only open what you think has value now, you will miss all the opportunities that you can't anticipate.

Sir Tim Berners-Lee, who invented the worldwide web (so he knows a thing or two about open data) implored all of us at TED - now is the time for data to be open. Teri, our state should be the leader in this. It is our data. Please open it up to us and let all of us help promote change, and build value.

Facebook, APIs and Monetization

2009-05-01T16:01:28-07:00

Image via CrunchBase

One of the topics we frequently talk about at Mashery is "how does a company monetize its API?". More often than not, the answer is that the API is a new distribution channel for the company's core service, and as such is monetized however the core service is.

So if you are a content company, you monetize your API through driving more traffic to your site (as the New York Times does) or through including advertisements with the content served over the API (as the Guardian does).

If you sell monthly subscriptions to movie rentals like Netflix, you monetize your API by improving the subscriber experience, so your subscribers get more value from your service. Churn goes down and revenue goes up.

Or if you sell access to specialized data like Urban Mapping, you monetize your API directly by selling access to a premium API.

So where does this leave facebook?

So many pundits are scratching their heads about facebook's new activity stream API.

"They are going to lose so much traffic!"

"How will they sell ads?"

"Isn't facebook giving away a potentially valuable source of revenue?"

Take a deep breath and think, everyone.

I think we can all agree that facebook's core competency is not around creating a universally acclaimed user interface. What they have done is created a sharing and communications engine accessible in a variety of ways that, for a variety of reasons, two hundred million people find to be compelling and useful. Along the way, they have gathered a ton of really useful information.

Google claims that their mission is to "organize the world's information and make it universally accessible and useful". Lovely, laudable and largely true. But when it comes to information about individuals - what and who they are passionate about; self-reported demographic info; social graph; employment and job title; you name it - facebook has organized a ton of information that Google can't offer.

It is this information that is facebook's core asset, and their sharing and communications engine that will ensure that the value of this asset continues to grow. It's true that at this point, facebook monetizes that asset by running ads on its site, but the API unlocks many more ways to generate revenue - and to grow a lot faster.

Advertisers buy ads on facebook because they're not just buying keywords - they can buy people who work at specific companies, or like specific movies, or have specific demographics...or all of the above. That won't stop; plenty of people will continue to buy ads on facebook.com because many people will continue to use facebook.com as their primary means of accessing facebook.

But just as Google brought us Adsense, so will facebook be able to monetize the users who access facebook through third party sites and apps. Except rather than being the mess that Adsense has evolved into - many advertisers have found that spending money on the "content network", as Google refers to Adsense publishers, is not nearly as effective as buying the Adwords that are displayed in google.com search results, and smart publishers are able to manipulate their sites and pull in spam links to attract the appropriate ads and clicks - the facebook network will be every bit as effective for advertisers as facebook.com.

So how can facebook monetize the activity stream API? Here are a couple ideas...

Adsense 2.0 - people place ads with facebook with the same ultra-tight-focused targeting that they can on facebook, and facebook pushes those ads to the third party sites or apps and shares revenue
Facebook sells the targeting data itself through an API, and the third-party apps/sites can query the API to figure out the appropriate ads to serve. Facebook either shares revenue or gets a CPM fee for access to that API that is lower for basic demographics and much higher for more specialized information (employer, job title, university, favorite movies, etc, etc). One opportunity with this model is that not only can the developer serve standard ads, but they can also do a whole range of enhanced sponsored experiences that could generate significant revenue. This is especially true with third-party apps that would be freed from the constraints of the browser.

And I think we'll see even more creative ideas as people begin to understand the power of the API.

But whatever happens, it will be based on facebook's core competence, not on what can happen within the constraints of the www.facebook.com website.