David Sadigh on user retention:
Intentional targeting: identifying search terms used to reach your site, and using these terms to inform the content of your site. e.g. displaying a family photo on a holiday site, if the visitor has reached it by searching for "family holidays in Italy".
It's not just about click-through, it's about sales.
What else can we use for personalising and targeting? Weather, stock market, time of day. On smaller screens, personalisation isn't a nice-to-have, it's essential (I'd say this might be true but doesn't logically follow).
David Marcus (Zong) on new business models enabled by mobile:
Today 1.3bn internet users, 3.2bn mobile users. Every second in China 4 babies are born, and they get 25 new mobile subscribers.
Much more content is sold on mobile than on the Internet. Why? Content isn't sold online, advertising is the model. Mobile is more commercial: Kanye West single is £0.79p on iTunes for the full track, or £3.50 for the ringtone.
Web and mobile are 2 different worlds. Advertising on cellphones won't work (Blyk might disagree). For web, one build ensures universal access. For mobile you have porting issues, carrier deals to distribute, etc.
It's traditionally been web people vs carriers. iPhone has changed things by inserting Apple into the value chain to take percentages of revenue. It also provides a route for mobile content (sync with desktop) that avoids over-the-air network use.
Qik: live streaming from mobile to web
2.3bn SMS messages will be sent this year.
Monetisation opportunities: mostly it's around building traffic then monitising it. What other examples?
- Swedish company Startle;
- FaceParty: text in to have your picture displayed at the top of the front page of the site (a bit like Flirtomatic's paid-for personal advertising)
- Facebook: an application Zong are working on to have live chats, extending to groups;
- The Cloud sell wi-fi access time by premium SMS;
All examples of mobile as a billing mechanism, no?
Going to talk about the thinking behind OpenSocial, rather than details: the social cloud.
We consider the Internet a cloud because we don't want to think about wires, cables and routers. It lets us activate a SEP field on these things. The younger generation sees the internet as oxygen, already there.
[Shows XKCD map of the social software world.]
Kids only use email to talk to The Man. Instead their social network identity or blog is them. Lots of URLs aren't documents, they're people. Social Graph API lets you expose the "me" and "friend" links between sites. New social software expects you to import all those old relationships. With the API, this can be automated.
Data, friends, activities drawn together (oooh) into clouds. OpenSocial maps the differences between these clouds away, so we don't have to worry about them.
Relationships: XFN proscribes 13 different relationship types, but it's more complicated than that. For instance, people don't break friend links on a site, except in the case of a messy breakup. At some point people get fed up with this and dump the identity, creating a new one: deliberately destroying profile and retaining plausible deniability.
[Shows map of relationships in Jane Austins Pride & Prejudice]
Douglas Adams: "you can't trust what people say to you on the web, any more than you can trust what people tell you using megaphones or in restaurants".
Our brains spend a lot of time and space dealing with social politics.
Francois Grey of CERN on grid computing and what it means for society and the web:
Web is changing the way science is carried out by enabling millions of citizens with no scientific training to contribute meaningfully, through volunteer computing.
Big event of 2008 may be the Large Hadron Collider being switched on: the largest producer of scientific data on the planet: 40m images/second, 15m GB of data/year. 100k CPU grid analysing it - SETI@home had 500k CPUs. Folding@Home got >1 petaflop of processing; it's been preloaded onto PS3s by Sony!
"Summer students are the power behind all scientific progress" :)