Notes on Charlene Li's Future of Social Networks SF AMA talk

Last night I went to an interesting talk by Charlene Li at the SF American Marketing Association -here are my twittered notes. See also, Charlene's slides.

says @charleneli: Theme is "social networks will be like air" - her better phrasing of my "Social Cloud" idea
says @charleneli: in future we'll say "wasn't it quaint that we had to go someplace to be with our friends"
says @charleneli: "I want Amazon to have a 'friend's reviews button on there - or anywhere else they could be"
says @charleneli: we'll have a feed of the presedential debates with our friends tweets on - like I did in 2004: http://bit.ly/IRCdebate
says @charleneli: universal login with OpenID lets you tie your IDs together, and sites can import friends from your networks
says @charleneli: I had to friend my co-author Josh 35 different times on different sites - Portable Contacts should save us from this pain
says @charleneli: Profiles where they are useful - eg LinkedIn profiles showing up in Lotus Notes via email
says @charleneli: your friends activities in context with GetGlue.com's plugin - Iron Man wikipedia page and IMDB page shows friends reviews
says @charleneli: 2 sets of standards exist Facebook's own protocols and the OpenStack backed by Google, MySpace, Plaxo, Yahoo and more
says @charleneli: advertising has evolved - content targetting for demographics; Search marketing for intent; behavioural targetting
says @charleneli: how many of you have gone to a social network site and remember seeing an Ad? or clicked on one?
says @charleneli: Who wants to be a fan of FiberOne on Facebook?
says @charleneli: people want to tell each other about things they care about - need new ads for this
says @charleneli: examples of new Ad types - branded virtual gifts, shown to you as your friends gave or received them
says @charleneli: SocialVibe has profile sponsorships that donate to your favourite charity eg colgate ad to leukemia
says @charleneli: the Tipping Point argued that there are influencers that can make a product go viral [I disagree see http://bit.ly/watts ]
says @charleneli: social graphs and interests, culture of sharing and online and email behaviour can create context for ads
says @charleneli: vendors who identify influencers include 33across, lotame, media6 degrees, unbound technologies
says @charleneli: network neighbourhood modelling in interesting - homophily is a good predictor for clusters - you are like your friends
says @charleneli: Google tracks who I email most - very useful to me: "In Google I Trust" http://bit.ly/BtvV
says @charleneli: Media6 identifies you by profiles you view on SNSs - shows ads to your friends based on your purchases
says @charleneli: Media6 gets 3-7x increase in response rates on banner ads through this homophilic targetting - no PII involved
says @charleneli: Influencer strategies are a misnomer, btu clustering works
says @charleneli: People will demand greater contol over when, where, how profiles + friends are used. Detailed permissions - a UX nightmare
says @charleneli: remember when people didn't trust callerID? Now if you turn it off, people won't take your call
says @charleneli: setting up lists of who can see your pictures is a pain - have to categorize people - reclassifying is hard
says @charleneli: there's a need to better articulate and detect sub-groups of friends so this is less of a chore
I pointed out the power of asymmetric friending eg http://bit.ly/publics and @charleneli and audience agreed that it reduces awkwardness
says @charleneli: people will pay real money for virtual gifts
[ChrisSaad @kevinmarks asymmetic is good, the term friending is not great. I prefer follow or subscribe ]
@ChrisSaad agreed "following" is a better term for this
Audience: when will people profit from us using their profiles? @charleneli says we all have our own CPMs
[clynetic @kevinmarks What is CPM?]
@clynetic CPM is marketingspeak for 'cost per thousand' - I suppose CPA ( cost per action) is better
says @charleneli: don't give up your social capital for short term gain me: don't be the Amway guy at the party
says @charleneli: behavioural targetting is often faulty, as behaviours change
says @charleneli: social media advertising experiments are waiting for turnaround
says @charleneli: GYM (Hotmail for M) will test social media integration with webmail
says @charleneli: Facebook Connect and Open Stack gaining traction with media co's
says @charleneli: Social shopping experiments start - we want our friends recommendations
says @charleneli: identify where social network data and content shoudl be integrated in your sites
says @charleneli: leverage existing identity and social graphs where your audience is
says @charleneli: get your privacy and permission policies aligned with an open strategy
says @charleneli: find your trust agents - in google I trust? do you trust facebook?
says @charleneli: the media buyers are still trying to buy demographics or content, not better targetting

My twittered notes on the Leweb Social panel

Platform Love: Getting Along - Panel

Panelists:


• David Glazer - Director of Engineering, Google
  
• Jeff Hansen - General Manager, Services Strategy/Live Mesh, Microsoft Corporation
  
• Dave Morin -Senior Platform Manager, Facebook

  
• David Recordon - Open Platforms Tech Lead , SixApart
  
• Max Engel, Head of Data Availability Initiative, MySpace

Moderator: Marc Canter - CEO, Broadband Mechanics

Watching the 3 Davids, Max, Marc and Jeff talk social at LeWeb
says Marc Canter 'open is the new black' - and asks about the Open Stack
says @daveman692 google, yahoo, microsoft all building on the open stack - won't FaceBook become the underdog when openness wins?
Canter suggets OpenID will be the brand that ties the Open stack together
max of MySpace "what we're doing with these standards is moving the web forward - when the web hits a roadblock it routes round it"
max of MySpace:"90% of our users think of themselves as URLs so OpenID is a natural fit for us"
Dave Glazer: the goal is to let users do anything they want to, with others, anywhere on the web. OpenID lets you log in anywhere
Dave Glazer: openSocial solves a different bit of the puzzle - JS APIs to run the same app in different social contexts REST APIs web to web
says @daveman692 the web is designed to be distributed, and the Open Stack fits this model
Jeff of Microsoft: live mesh is built on symmetric sync - supports Open Stack, OpenID shipping, OAuth looks good, support PortableContacts
Jeff of Microsfot: we're evaluating the OpenSocial gadget container
Marc canter "we're putting all our balls into ev williams vice"
Jeff: we offer lots of languages. Marc: lots of ways to put our balls in your vice
Max: we support OpenID, Oauth, OpenSocial but you can too
Marc: anything good for the Open Web is good for Google
Marc Canter wants a URL for each Gmail? DG: each one does have that, but only you can see it
Dave Glazer: there are 3 classes of information: Public, Private and Complicated - users should never be surprised by who can see what
says @davemorin facebook wants people to have a social context wherever they go
says @davemorin FaceBook had to create a Dynamic Privacy model for FB Connect @daveman692 calls shenanigans - LJ had those in 1999
asks @daveman692 of @davemorin why are you giving microsoft access to all our email addresses wihtout asking permission?
Max of MySpace - we've shown that security and openness work together by using OAuth, and can revoke them from in MySpace
Dave Glazer: need to separate the technical levers from the social customs. technology can't stop people putting your bizcard on the web
says @techcrunch "call bullshit on facebook" - broke integration with google. FB don't want an open stack, they may be forced into it
says @tommorris how can MS be on the panel after the debacle of Office OOXML which wasn't open or XML?
says @dave500hats could we get contacts with certain features eg tennis fans?
Dave Glazer: there's an open spec process to define new attributes in the spec - if you want to add one go and propose it

Cycling to new layers of freedom

Dave Winer used the public beta of Google Friend Connect to reflect on tech industry cycles:

A new generation of young techies comes along, takes a look at the current stack, finds it too daunting (rightly so) and decides to start over from scratch. They find that they can make things happen that the previous generation couldn't cause they were so mired in the complexity of the systems they had built. The new systems become popular with "power users" -- people who yearn to overcome the limits of the previous generation. It's exhilirating! [...]
The trick in each cycle is to fight complexity, so the growth can keep going. But you can't keep it out, engineers like complexity, not just because it provides them job security, also because they really just like it. But once the stack gets too arcane, the next generation throws their hands up and says "We're not going to deal with that mess."

Now, I may be a few years behind Dave, but I think he is throwing the baby out with the bathwater, or the stack out with the cycle here. Back when I started out, to get my computer to generate sound, I had to make my own D to A converter to attach to the parallel port, and for non-character graphics, my hardware hacker friends swapped the character generator ROM for RAM, and I had to code in assembler to swap the display data in time.

Now my son thinks nothing of mixing 10 polyphonic Midi tracks in an afternoon or editing hi-def video (and yes, it's on an OS I helped to make capable of that).

Dave's revolutionary impulsiveness has a germ of truth, but what really happens is that successful technologies become invisible infrastructure for the next things that build on them.

I no longer need to write assembler, heck I no longer need to write C code. Dave's very URL - scripting.com - shows how we have built up layers of utility to work upon.

HTTP, HTML, JSON, Atom and Javascript are infrastructure now. Our deepest role as developers is to build the invisible infrastructure for the next generation to take for granted, so they imagine new abstractions atop that. Dave did it with feeds.

What we're doing with the Open Stack — OpenID, OAuth, PortableContacts and OpenSocial— is part of this evolutionary cycle too. We're combining building blocks into a simplified whole that makes sense to people who want their websites to become social.

It comes down to what you can take for granted as the baseline to build the next exciting cycle on.

OpenSocial’s birthday today

OpenSocial Reach chart
Originally uploaded by Kevin Marks

Just over a year ago, we launched OpenSocial to the web, with a few example applications and a lot of potential. Now, a year on, over 600 million social network users can use OpenSocial applications in their preferred social network sites.
Then, applications had to be embedded in sites as gadgets, which makes the social context clear for users, but means developers have to write some Javascript, and can only run code when the user is looking at the site.
With OpenSocial 0.8 rolling out, the REST APIs mean that developers can integrate with social sites using server-side code directly, potentially delegating user registration, profiles and friend relationships to an already-trusted social site, and feeding activity updates back into them.
To do this, we are building an Open Stack, based on OpenID, XRDS-Simple, OAuth, PortableContacts and OpenSocial. By composing open standards in this way, we can make each one more valuable. The advantages of OpenID over email login in itself are not that obvious to users, but if the OpenID can be used to bring in your profile and contacts data - with your permission via OAuth - suddenly the added value is clear to users and developers alike. This connection was one of the exciting discussions at the Internet Identity Workshop this week - here's a video of myself, Steve Gillmor, David Recordon and Cliff Gerrish talking about it.