Stowe Boyd: Building Social Applications [en]

Warning: these are my notes of Stowe‘s workshop at LIFT, meaning my understanding and interpretation of what he said. They might not reflect accurately what Stowe told us, and might even be outright wrong in some places. Let me know if you think I really messed up somewhere.

Update 05.2007: enjoy the slideshow and the video of his presentation (not the workshop!).

Questions to play ball with:

1. What makes social applications social (or not)
2. How can we make applications more social?
3. What are the common factors in successful social applications?
4. What is worth building?

5. iTunes vs. Last.fm; also non-social applications which implement, at some point, some social component.

“Software intended to shape culture.” Stowe Boyd, in Message, August 1999

steph-note: a step further than “groupware”

Applications which are qualitatively different. But they haven’t replaced the rest: people are still building applications which allow people to buy stuff online. But we’re looking for ways to stick the humans back in there (“what do the top 10 authorities on cellphones recommend?”)

Read: The Great Good Place by Ray Oldenburg (Third Place, not home and not work)

Decreasing affiliation in the USA (Putnam — sp?). People spend less time “hanging out” with people. steph-note: cf. danah/MySpace More TV. Commuting isn’t that significant, but hours in front of the TV is. The light at the end of the tunnel, the only hope we’ve got left, is the internet. Social hours spent on the internet are hours not spent watching TV (steph-note: yep!)

TV is not involvement in people, but in this “entertainment culture”. TV reached lowest numbers in the USA since ’50s.

One way we can measure the success of a social application is how much it moves us in that direction.

Social: me first. Put the individual in the centre. Look at the difference between traditional journalism (disembodied third voice) and blogging (first person, you know who’s writing and who’s reading). Need to start with needs and desires of the people using it (?).

Adoption happens in stages. First, the application needs to satisfy the needs of an individual, in such a way that he/she comes back. And then, there needs to be stuff to share that encourages the individual to invite his friends in.

my passions — my people — my markets

Start with the people. Put the people in the foreground (but how?) Easy to fail if you don’t do that right. How are people going to find each other? Second, support their networks/networking.

Only third: realisation of money — markets — shipping etc.

Give up control to the users: “the edge dissolves the centre”.

To review a social app, you need to use it “for real” over an extended period of time.

Xing: the edge doesn’t dissolve the centre. E.g. can’t create a group. Need to ask them by e-mail, and they try to control group creation and management.

Build an environment in which people are “free”. Allow them to find each other.

Success factors for a social application: me first and bottom up. Otherwise, it won’t spread.

Blogging: primary goal is social interaction and networking (steph-note: half agree, there is the “writing and being read and getting some recognition” goal too — and that is not necessarily social interaction and does not necessarily lead to network contacts)

What suicide girls get right: low price, real people, real lives, social stuff like chat, pictures, etc. They have the connections between the people as the primary way to go around.

Semi/a-social

• iTunes
• Bestbuy.com
• Pandora (until recently)
• After the fact (eBay: reputation, Netflix: friends in a tab, Amazon: recommendations from other users, Basecamp: not that social, fails some of the critical tests)

The Buddylist is the Centre of the Universe…

A case against IM being disruptive: the user chooses how disruptive the client is (blings, pop-up messages, etc… same with e-mail)

Totally acceptable to not answer on IM. But also, maybe at times your personal productivity is less important than your relationship with the person IMing you.

“I am made greater by the sum of my connections, and so are my connections.”

(Give to others, and they’ll give to you. Help your buddies out, be there for them, and others will be there for you when you need them.)

List of hand-picked people who are on your list.

Groups help huge communities scale, in the way they bring them down to manageable sizes for human beings again. (Dunbar constant, roughly 150 people.)

Six degrees of connection doesn’t work. People are strangers. Even second degree is really weak.

Difference between people you really talk to, and “contacts” (often people will have two accounts => should build this kind of thing into the service — cf. Twitter with “friends” and “people you follow”).

Me, Mine, and Market.

Market: it’s the marketplace where the application builders are going to be able to make money by supporting my interaction/networking with “mine”.

You can’t “make an app social”, you need to start over most of the time.

Think about the social dimension first, and then what the market is. E.g. social invoicing app, what could the market be? Finding people to do work for you. And then you can invoice them using the system.

E.g. Individual: “I need a perfect black dress for that dinner party.” => who knows where to shop for the most fashionable stuff? => market = buying the perfect black dress, with commission to the recommender. (New social business model!)

Facebook profile: all about flow, it’s not static. It’s a collection of stuff going on in my world. Information about my blog (posts), friends… I don’t have to do anything, and it changes.

It represents my links to the world. People want to belong. Be in a context where what they do and say matters. Make it easy for users to find other people who will care about them.

Orkut failed because it was just social networking for the sake of social networking. Not targeted at a specific group of people. Nobody who cares! Disease-like replication and then died down. Nothing to do there.

Swarm intelligence. People align around authority and influence. Some people are more connected then others. Inevitable. Swarmth = Stowe-speak for measure of reputation. As soon as reputation brings something to those who have it, charlatans step in and try to figure out how to game the system. Need to be aware of that, to discover those cheating mechanisms and counter them.

General principle: things are flowing, and we want to support the rapid flow of information (ie, stuff that goes in my profile). “traffic”: do you make it possible for people to get information from a variety of sources delivered quickly to them? (e.g. Facebook bookmarklet) (traffic=possible metric).

The media hold the pieces, but not the sense of the conversation. You need to immerse yourself into the flow to get it. How transformative is it to get a constant flow of information from people you care about? Can’t evaluate that from the outside.

Tags

cf. David Weinberger: tags matter for social reasons. The power of classification is handed out to the users. They use it to find information and to find each other. They define implicit social groupings.

If people don’t “get” tags, the interface isn’t good. Because the concept is really simple. (e.g. Flickr, del.icio.us get it right)

Discovery

Primary abiding motivator of anybody on the internet: discovery (things, places, people, self)

One of Stowe’s pet peeves: Groups and Groupings

Networks are asymmetric, accept it. Everybody is not equal in a group. The groups are always to some extent asymmetric.

Groupings are ad hoc assemblages of peope with similar interests (from my point of view). (My buddy list categorisation.)

Groups try to be symmetric.

Community of tags. They happen automatically.

Power Laws

There will always be people with more power than others, get over it. The recommendation of somebody with more swarmth should count more than that of one with no swarmth.

Accept and work with the imbalance of power.

But careful! The people decide who has more swarmth. And you need to constantly counter the games. Natural social systems are self-policient (sp?).

Reputation

Measure and reward swarmth (steph-note: !== popularity, quantity)

Reputation is not transportable from one network to another.

Deep Design

• last.fm (neighbours!)
• upcoming.org (events are nothing without people!!)
• Facebook
• ThisNext (about design and fashion)

First, just build the social app. Once the social stuff is in place, build the market (see Last.fm).

Journal where you can integrate music references. With backlinks from artists.

Mistake? tags aren’t source of groupings.

steph-thought: Flickr groups are not just about people, they are about editing content (creating collective photo albums).

If you have an existing social app, and an entrenched body of users, to make people switch to your new product you need to be an order of magnitude better.

Tag beacons: a recommended tag (e.g. lift07)

If you make people tag an item, the tags used stabilize over time. After a while, the same 10-15 tags. Little chance a new user two years latter will suddenly introduce another tag.

ThisNext is pretty. A piece of social interaction stuff missing however — can’t communicate with other people. Profile just leads to recommendations.

Cautionary Tales

Basecamp and the Federation of Work: multiple logins, domains — fragmentation. Wanted to be able to pull everything in a single place. Not simple to keep track of everything one has in the system. Pervasive static models with hardly any flow. It’s an online groupware app, not a social app. It doesn’t put me in the foreground.

Outside.in is about finding people who are in your zipcode. I remember Stowe did a post on this some time back. “Where’s the people?”

You only get one first launch. What’s the point of missing it by doing it before you got to the social tipping point?

Blinksale: where’s the market? (invoicing thing)

Explorations

Where is all this going? All commerce on the internet in the future will be social. Put in context of social recommendations etc (perfect little black dresses). A social iTunes — what would it look like? They could acquire Last.fm and integrate them to iTunes, for example. I could recommend music to my friends via iTunes…

Calendars are hard! We’re still waiting for the perfect (at least good) calendar-sharing system.

Social browsing… “What should I look at today, based on recommendations of these n people I really find smart?”

Safety/privacy concerns: solutions we have in the offline world need to be emulated online.