Athens Facebook Developer Garage

Athens Facebook Developer Garage
(live blogging from Athens, Zografou)

Evolution of the Web

Some stats on web, web2.0, adoption, AOL, yahoo, …, googlemaps, youtube

The facebook platform

  • Massive ditribution
  • Fast experimentation
  • but

  • Constant changes to policies

Architecture

php+java official libs (wasn’t official java support dropped???)

FQL – one table per query only (no join), access user, friend, group, group_member, event,event_member, photo, album, photo_tag

FBML – subset of HTML, some custom tags, Mock AJAX… Canvas belongs to you, apart from anything deceitful, you can put anything you want there (ads,..) – but ads like googleads don’t have context info to show related stuff, so???

Anatomy of An Application

  • Only one-line status messages can be added to user statuses from apps… bigger ones must be approved from users.
  • Now, limit to how many invitation an app can send to friends.
  • Presenter (Nikolas Kokkalis) took part to Stanford Facebook class
  • Cannot store male/female user info from facebook (for more than 24h), but if we ask users then we can!
  • Viral factor f = X*Y*Z, f>1 then Viral growth… For each user, at least one friend accepted invitation (for PerfectMatch, f was 1.4 – they had to rename to Best Match due to … lawyers!)
  • They’ve done experiments with the same exact app registered with 5 different names – and it does make A LOT OF DIFFERENCE (name matters)
  • Every concept can be made viral … with the right modifications
  • Other ways of income (apart from ads) : Virtual currency, virtual objects (VERY popular for teenagers to pay for those), surveys!!! (you ‘sell’ people filling surveys)
  • Also, users think your company is more legit if they see ads in your site.

Virality

  • It is an Art that requires a ton of Science
  • Invitation text is terribly important
  • Make it social, simple – use viral channels, ab test all
  • Viral channels in FB: invitations, notifications, newsfeed, email, profile box, profile action
  • Knowledge gap (we know the perfect match, do YOU know it? there’s a message for you, do YOU want to see it?)
  • Modify a channel, gather data, analyze results (& restart)
  • but… changing a channel can be time consuming, measuring + analyzing results difficult (reference to their http://viraltuning.com/ but not much info there yet)

Some apps

… developed by greeks

  • Greek Namedays
  • kerasmata
  • mySidebarBook

andyhot

Leave a Reply

Your email address will not be published. Required fields are marked *

Read previous post:
Tapestry-4.1.6 maven archetype

Tapestry-4.1.6 maven archetype There's an updated archetype for Tapestry-4.1.6-SNAPSHOT over at (the well known snapshot repo) http://people.apache.org/repo/m2-snapshot-repository All the hard...

Close