Posts Tagged ‘Facebook’

Of the platform wars

Friday, January 16th, 2009

Platform wars are interesting, mainly because the winner will potentially be in the hotseat for quite some time. Take a look at Microsoft winning the desktop platform war – they’ve had a very well entrenched position for a considerable amount of time.

We’re more or less at that point with the web. We were at the beginning of the wars about a year ago when Google launched OpenSocial and FriendConnect, and then Facebook with their platform and Facebook Connect. Having a well adopted framework is very important – developers typically want to develop apps for a widespread platform. This allows them to build once, and get multiple distribution on different sites.

From the outset, it looks like most of these services are very similar, but on closer inspection they do have some subtle and important differences, which I’m grappling with at the moment.

To sketch a background, I’m doing research on a pretty exciting portal project. At the risk of giving too much away – we need a portal that users can customize with widgets. It’s not too far off from something like Netvibes or Pageflakes. The scripts to run these sites are a dime a dozen and pretty easy to get hold of, plus it’s not so difficult to build them from scratch. They’ve almost become as ubiquitous as white label social networks, blog platforms etc. To use the lingo, they’ve become a commodity.

We initially did some research into white label networks to add a social element to the site. But this poses new problems though – will users need to sign-up to the site again, and find all their friends on the site? Wouldn’t it be cool if they could get a current list of friends already on the site, and potentially on the site right now? We honestly don’t want to build a new network from scratch. Cue FacebookConnect. We decided on Facebook Connect over Google Friend Connect for the predominant reason that we find the social profile of Facebook much more expansive than Google. Even though lots of people have a Google account, we feel the data in Facebook is much richer and also much more organized. There is the benefit of action injection into the news feed, as well friend linking and profile integration.

So we’re going to go the Facebook Connect route – we’re pretty excited about the one-click login for users as well. Now however, we’re posed with a different problem. We want to outsource the development of new widgets to the portal. It’s a similar problem that Facebook had a while ago, and the reason for the platform development (or definitely a large part of it). Facebook has effectively allowed for the outsourcing of it’s ecosystem – things which make the site useful. They first started with Events, Albums and Videos, eventually they expanded in the Marketplace. Soon enough it becomes apparent that they won’t be able keep up with new apps – what if they want to add a TV guide? And if they build a TV guide, how would they build a localized one – for us here in South Africa?

An easier route is to allow outside developers to build those apps – effectively outsource the development of them. Enter the Facebook Platform. This allows for hyper-localized applications (e.g. a TV guide for DSTV) without Facebook needing to build them, and also better applications. Facebook is going “license” the building of the marketplace to Oodle.

So what we’re going to need for our portal site is a similar platform framework. We could build our own platform, but that won’t make much sense. For one: we also want to outsource some of our development. Say someone builds a DSTV app for Facebook, we want that app to be available on our portal as well without the developer having to build the app again. Secondly, we also don’t want to go through the process of building an entire new platform – we simply don’t have the people to build a new platform, plus we want to play nicely with the other platforms out there (why build something that is already built?).

And we’re lucky here as well – we have actually have a choice! We can look at Google’s Open Social, or we can go the Facebook Open Platform route. But once again, we’re more partial to Facebook, also because we’re not to keen to go down a route where we need to figure out a way to integrate Open Social and FBConnect. Yikes. Plus the OpenSocial implementations that I’ve seen are pretty mediocre.

So now some problems start cropping up – I’ve never seen an implementation of Facebook Open Platform, plus I have no idea what we will be able to do with it once we have it setup correctly.

  • Will the Facebook Open Platform be compatible with our portal?
  • Will we be able to filter apps that will appear on our site?
  • Will we have access to the applications?
  • Will people be able to build an application once, and it be used on Facebook, and our site?
  • What happens to users who aren’t Facebook users?
  • Where are other examples of these same implementations?

These questions are really bugging me at the moment, and if anyone has any tips, we’d really appreciate some feedback!

User Centric design and identity with Beacon

Friday, November 30th, 2007

Facebook Beacon has brought up a lot of issues – almost all bad of course.

What’s been happening with Identity 2.0 will hopefully solve this issue, and it also relates to personal information that other companies have. Jason from Bokardo started me on a thought process.

Personally, I have no problem with companies knowing my spending habits, but I want to know what spending habits they know. I want that to be transparent, and also changeable at any point. I want to have access to my user profile on Blockbuster (US example) and I want to know what data they are capturing. And this is for two reasons: I want to be able to see what dirt they have on me, and then I also want to be able to change that data so that Blockbuster can provide me with a better experience. In the case of movies, sometimes I rent only comedy’s, but I’m also actually interested in documentaries, specifically documentaries on street art and modern urban architecture and how public space facilitates discourse and society. But they don’t know that because they don’t stock it,  and I’ve never browsed/searched for it because I know they don’t stock it. If I can tell them this somehow (I can mail them, but who does that? For all my services?) then perhaps they would start stocking these movies.

Back to user centric design
This is perfectly solvable with APML. What does APML do?
It’s great. Here’s what it does:

What APML does

APML allows users to share their own personal Attention Profile in much the same way that OPML allows the exchange of reading lists between News Readers. The idea is to compress all forms of Attention Data into a portable file format containing a description of ranked user interests.

That is exciting. APML is nothing else than your own piece of market research, but the great part is that it sits with YOU. Not Truworths, Stuttafords, NIKE or Facebook. These services inevitably only have a limited view on your attention – they only track your interactions with their own products. NIKE has an APML file for you, but it’s only for you using their products. Facebook Beacon was an attempt to aggregate all of this, and they are in a very good position to do this – they already have a very good profile of you, and now they’re starting to gather market intel on you.

Marshall Kirkpatrick has a pretty good argument behind how Google botched this up with their new Feed Recommendations. Why do I think? Because I didn’t have access to that Attention Profile that they have on me. I’m sure they have VERY good algorithms etc, but at the end of the day, me reading icanhascheezburger.com daily, every single update, doesn’t really mean that lolcats are high on my attention list – it’s only that one blog. And I think only I can tell you that.

It’s got 1565 of my RSS subscriptions, thousands of Gmail messages (32k unread ones, in fact), several Google Custom Search Engines, my GCal life history, search history and more I’m sure – all tied to my Google Account and all it can give me is 20 new sources?

Basically, I want to be able to control that data and construct my OWN profile, and know which sites know what from me. At the moment these attention silos are not being centered around the user, they are being centered around services that a user uses. Woolworths has a different profile of me than NIKE does, which is sometimes good, but they aren’t cross selling properly.

Back to OpenID and OAuth
Principles behind OpenID and OAuth aim to solve this. OpenID allows me to login to all these sites using a single sign on (in a strict sense of the word). OAuth allows me to tell these sites what I want them to know about me. So, I sign in to Flickr and it asks a bunch of details from me. I have already filled in all these details on my OpenID server, http://unodewaal.identitu.de. I then point Flickr to my OpenID server, telling it, get the details there. I then get asked on my OpenID server: Hey, Flickr is asking for a whole bunch of data of yours, what do you want to tell it? Name, Surname, Email, Country, Time Zone (cos that’s always a hack), etc etc.

All this data is transferred with a click of a button. But not only that, I can then manage what info Flickr has of me, from my OpenID server (not sure if Identitu.de can do this yet).

We’re trying to do this with Utterbuzz. You will have multiple profiles, on multiple sites, all managed from a single interface. So, you join a network for online dating, that has a specific profile that you play up (you’re taller, healthier, leaner, increased your salary by 10x and your breast by 3x), but you don’t want any of that data to cross over to the school network that you’ve joined. But it’s still managed in a central profile. Ning has done a similar thing with their profile management – it’s basic (you can set different profile pics for different networks), but it’s an idea.

OAuth is the technology behind all of this for standardizing the authentication process. Listen to this podcast on Oauth with Larry Halff, Eran Hammer-Lahav and Chris Messina for more:

Three of the minds behind the Oauth initiative join us to tell us about this emerging “open protocol to allow secure API authentication in a simple and standard method from desktop and web applications.”

Back to Facebook Beacon
I believe the Beacon experiment shows that users want this kind of transparency. They feel pissed off to know that Blockbuster and Starbucks might be sharing data via Facebook. But, if you knew what data they had, and you could control that data, would it be different? I believe it would. Every time I access Amazon, it asks me: “Hey Uno, I have no idea what books you like. Do you want to give me access to your APML data?” I say yes, and I flag my APML manager to “Always allow Amazon Access” so that I don’t have to do this all the time. I also say, “Only give Amazon access to this data, not the other. They don’t need to know that.”

When I browse Amazon they create a profile set for me, kept on their servers. I want that kept on my server, or at least have access to that file, because when I go over to Blockbuster, I can say: “Hey Blockbuster, I’ve been looking at these books, do you have any similar movies?” And Blockbuster spits them out. I then flag Blockbuster as: “No, I don’t want to give Blockbuster permanent access or allow them to keep my data” because I’ve heard that they resell data to spammers. Also, I don’t want to let all my friends now about the products I’ve bought, that option should sit with me.

At the end of the day Dave Winer says it best:

Long-term, however they both have problems because advertising is on its way to being obsolete. Facebook is just another step along the path. Advertising will get more and more targeted until it disappears, because perfectly targeted advertising is just information. And that’s good!

If Facebook Beacon went the way that OpenID is going then we would have better Attention Data and we would all be better off.