Yahoo opens up

Yahoo's Open Strategy offers an open development vision

When Yahoo announced its Yahoo Open Strategy (Y OS), in late April, it offered a vision that, if realised, could give back the company the mojo it lost several years ago.

With Y OS, Yahoo pledges to open all its sites, online services and web applications to outside developers, and give users a "social profile" dashboard to unify and manage their Yahoo services.

The ambitious plan is designed to let Yahoo radically improve its position in key areas such as search and social networking, and thus make a run at competitors like Google, MySpace and Facebook. With Yahoo in the midst of fending off Microsoft's take-over attempt, the success of the strategy appears even more important.

This week, IDG News Service got more details about Y OS from Neal Sample, the company's chief platforms architect, who said the project began forming early last year and was launched internally in September. An edited transcript of the interview follows:

Did Y OS trigger a widespread revision of existing initiatives?

In formulating this strategy we had a lot of input from the properties [like Yahoo Mail, Messenger and Flickr]. We talked about what we wanted to do and how we wanted to forge a new Yahoo, and got everyone on board. Once we had that alignment, it helped shape the vision.

It's a staged deployment approach. The properties will reflect in coming months elements of the Y OS strategy. The first deployment is search [via the Search Monkey project]. We took one of our largest and most important canvases, and one which has one of the most stringent performance requirements, and we began by opening that up to developers.

Was the trigger for Y OS the realisation that Yahoo users have flocked to social networks like MySpace and Facebook, drawn by elements they haven't been able to find in Yahoo, even while they continue to use Yahoo services?

Absolutely. Yahoo was very homogeneous at first, but grew up very quickly over 14 years and branched out into many verticals and properties, organically and through acquisitions. We're number one in seven different verticals and compete in many more, so users come to Yahoo to get the utility [of our services and applications.] Now comes the next logical step with Y OS. We've learned from our competition that social [networking] is an important dimension, regardless of whether or not it's treated as a destination [site].

Ours is a very different proposition from some of our [social-networking] competitors who may be struggling to find utility in the application space — in email and instant messaging [services], finance, news or sports [properties] — and build that type of anchor property on top of their social network.

We have the foundation: very strong utilities and applications. The next evolution will be to add social as a feature on top of the utility instead of the other way around. When you realise the assets Yahoo has, it's funny to see our [social-networking] competitors, to some extent, try to discover and build those.

In being truly open, we have two opportunities that other networks aren't necessarily providing. The first is to bring Yahoo off-network by API [application programming interface] calls, so you can take your Yahoo connections and experiences to flavour these other social networks. Another possibility is bringing social networks into Yahoo. If the best applications written by developers leverage our competition and bring in their connections, data and experiences into Yahoo — because that's what our users want — then we're open to that as well.

Could you explain the social-profile concept in the Y OS context?

In terms of the mechanism underneath, that's fundamental to Y OS: the unification of the profile. It's taking the user data and making it centrally available, and therefore available to the user all across the network.

Yahoo has quite a few different [user] profiles in services like Kickstart, Yahoo 360, Mash and the like. [Network Division Executive Vice President] Jeff Weiner was quoted recently saying that Yahoo has 25 different profiles that he's trying to work through. If he's the EVP of products and sees that as a challenge, imagine what's it like for an end-user.

Part of our strategy is to normalise those profiles and collapse them into a single place and reduce user confusion. We'll make a single dashboard for them to update their information, and we'll go from looking like 25 different Yahoos with 25 different profiles to one Yahoo and one profile.

This dashboard "social profile" will be rolled out later as a Y OS component?

Yes. It'll be one of the larger user-facing components.

How will the APIs and development tools be made uniform?

We took the first steps in a relaunch of our YDN developer network site a few weeks ago. Beyond that, there's an initiative around Y OS to normalise web services and give developers a single access point. You'll use your developer ID or credential and use Yahoo APIs in a uniform way.

Again, Yahoo growing up quickly over 14 years means that some of our APIs we developed in-house and others we acquired, and in that growth we've seen a lot of variability in how the APIs look like. Much like with the profiles, there were lots of different views of what Yahoo was if you were a developer. Y OS is putting out a Web service standard for our APIs, so that developers can learn Yahoo in one place, one time and leverage that with any API, instead of having to learn each one separately.

It sounds like there is a big data portability component to Y OS.

Yes, that's our third phase, which is opening up Yahoo everywhere. It involves letting users, with full transparency and control, take their data with them. The data is essentially theirs and they can use it wherever they want to.

Could you provide an update on Search Monkey?

It's available in a limited developer beta, as announced last week, and in the coming weeks it will be an open developer beta.

Initially, the Search Monkey API is triggered off the URL. A user would install an application that would look for specific things in the URL and then it would augment the results from that URL using Search Monkey.

For example, instead of seeing a Wikipedia link and a snippet, you could call back to Wikipedia and show a photo, get more context about the article, maybe even present some of the external links or the edit history, whatever would be interesting.

So initially it'll be switched on URLs, but later we've talked quite a bit about the semantic web, the ability to leverage structured information in the search result itself. So in a Yahoo Local business listing, the business can put structured information about itself as part of the result.

Other kinds of matching for applications might be triggered by keywords. I might have installed Search Monkey applications that look for results with "travel" as part of the response, and it might pull up destination information from across the web and build up a travel portfolio.

That sounds useful for end-users, but how do you strike a balance between richer search results and Web publisher concerns about too much of their content being exposed and discouraging people from clicking over to their site?

Striking that balance is a big priority. It's been an even larger priority in considering impact on ad revenue. Sometimes if you give better search results and the answers can be found there in context, you might not have as many searches that are being issued because users are getting to the information right away. But we've decided to put search quality above any of the other considerations. If users get the information they're most interested in as quick as possible, then everybody wins.

Still, there's an ongoing debate industrywide over 'when does a search result cross the threshold of fair use of someone else's content.'

That's an interesting issue, and Search Monkey will let us evaluate the impact of these changes for the user. This is revolutionary, not evolutionary, for search: the ability for a user to install an application that remixes the results and takes [search] from a fairly static programmed experience and lets them pull all sorts of other possible information sources. As we learn more about it and users give us feedback, we'll have a discovery period; and based on the things we find, we'll refine to the greatest extent we can. The opportunity is tremendous.

What will happen to your existing social networks, Yahoo 360 and Mash?

We believe in our 360 community. It's vibrant and thriving. The users, even knowing that 360 is going away, have stuck with it, because they've invested their time and created rich profiles and connected to a lot of different people. We're going to preserve their [time and effort] investment. We'll carry their profiles over and take their connections over into the new product. Same thing with Mash.

Join the newsletter!

Error: Please check your email address.

Tags DevelopmentDevelopment IDYahooopen strategy

Show Comments
[]