The Autopsy of Friendster - ACM Conference on Online Social

Chair of Systems Design
www.sg.ethz.ch
Social Resilience in Online Communities:
The Autopsy of Friendster
David Garcia
Conference in Online Social Networks. Boston, US
October 7th, 2013
David Garcia
Chair of Systems Design
The Autopsy of Friendster
www.sg.ethz.ch
Outline
1
When Online Social Networks Die
2
Measuring Social Resilience
3
The Autopsy of Friendster
Conference in Online Social Networks.
Boston, US
October 7th, 2013
2 / 19
David Garcia
Chair of Systems Design
The Autopsy of Friendster
www.sg.ethz.ch
Chair of Systems Design at ETH Zurich
Main Research Areas
Economic Networks & Social Organizations
e.g. ownership networks, R&D networks, financial networks, ...
e.g. online communities, OSS projects, animal societies, ...
Netherlands
Belgium
Ireland
Sweden
Germany
Denmark
United Kingdom
Finland
Greece
Spain
Portugal
France
Methodological Approach: Data Driven Modeling
economic databases: Bloomberg, patent and ownership databases
digital traces: user interaction, OSN, activity volumes
Conference in Online Social Networks.
Boston, US
October 7th, 2013
3 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
When Online Social Networks Die
1
When Online Social Networks Die
2
Measuring Social Resilience
3
The Autopsy of Friendster
Conference in Online Social Networks.
Boston, US
October 7th, 2013
4 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
Researching on the death of OSN
When Online Social Networks Fail
What are the reasons that drive large amounts
of users away from an Online Social Network?
Conference in Online Social Networks.
Boston, US
October 7th, 2013
5 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
Researching on the death of OSN
When Online Social Networks Fail
What are the reasons that drive large amounts
of users away from an Online Social Network?
Internet Archaeology
Analysis of the non-written traces of a
disappeared society, aiming at understanding the
way it worked and its demise.
The Onion on Friendster:
“Millions of profiles left utterly untouched”
“The users prized photos of themselves drinking”
“Friendster was a primitive society lacking video
comments, status updates”
Conference in Online Social Networks.
Boston, US
October 7th, 2013
5 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
Social resilience and users leaving an OSN
Social Resilience
The ability of the community to withstand external stresses,
disturbances, and environmental changes
Stresses for OSN include changes in the user interface, technical
problems, threats to privacy, rumors, competing sites...
Conference in Online Social Networks.
Boston, US
October 7th, 2013
6 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
Social resilience and users leaving an OSN
Social Resilience
The ability of the community to withstand external stresses,
disturbances, and environmental changes
Stresses for OSN include changes in the user interface, technical
problems, threats to privacy, rumors, competing sites...
Users have a cost c of using the network (global
constant)
Time to learn to use, standing ads, fees...
User benefit b is a function of its social
environment
Information and attention from social contacts
Conference in Online Social Networks.
Boston, US
October 7th, 2013
6 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
Social resilience and users leaving an OSN
Social Resilience
The ability of the community to withstand external stresses,
disturbances, and environmental changes
Stresses for OSN include changes in the user interface, technical
problems, threats to privacy, rumors, competing sites...
Users have a cost c of using the network (global
constant)
Time to learn to use, standing ads, fees...
User benefit b is a function of its social
environment
Information and attention from social contacts
If c/b < 1 a user will stay active, or leave
otherwise
Conference in Online Social Networks.
Boston, US
October 7th, 2013
6 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
Finding the users remaining in an OSN
Cascades of users leaving the OSN
When a user becomes inactive, its friends lose benefit
The cascade stops at a generalized K -core decomposition
Our assumption: b is proportional to the amount of active friends
A
B
C
D
Under certain cost c, there is a critical K for which the K -core of the
social network will remain active
Conference in Online Social Networks.
Boston, US
October 7th, 2013
7 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
When Online Social Networks Die
The Autopsy of Friendster
Cascades of users leaving a social network
Conference in Online Social Networks.
Boston, US
October 7th, 2013
8 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
Measuring Social Resilience
The Autopsy of Friendster
Measuring Social Resilience
1
When Online Social Networks Die
2
Measuring Social Resilience
3
The Autopsy of Friendster
Conference in Online Social Networks.
Boston, US
October 7th, 2013
9 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
Measuring Social Resilience
The Autopsy of Friendster
We measured the resilience of five OSN
name
Facebook
Livejournal
Myspace
Orkut
Friendster
date
2004
1999
2003
2004
2002
status
successful
in decline
in decline
in decline
failed
users
3M
5.2M
100K
3M
117M
links
23M
28M
6.8M
223M
2580M
source
(Wilson, 2009)
(Mislove, 2007)
(Ahn, 2007)
(Mislove, 2007)
Internet Archive
Friendster
Founded in 2002
Reached 20M users in the United States
Became popular in Southeast Asia, reaching 117M accounts
Friendster discontinued its OSN service, deleting all profile data
The Internet Archive crawled as much as possible
Now Friendster is a flash games site
Conference in Online Social Networks.
Boston, US
October 7th, 2013
10 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
Measuring Social Resilience
The Autopsy of Friendster
100
MySpace
1
Friendster
Empirical K-core decompositions
304
Conference in Online Social Networks.
Facebook
200
Livejournal
ks
Boston, US
October 7th, 2013
11 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
Measuring Social Resilience
The Autopsy of Friendster
The social resilience of the five OSN
100
P(ks > K)
10-1
10-2
10-3
Friendster
Facebook
Orkut
MySpace
LiveJournal
10-4
100
101
ks
102
The CCDF of the k-core distribution measures the amount of active
users under certain cost and benefit conditions
No network is the best under every condition
The survival of an OSN does not just depend on its topology
Conference in Online Social Networks.
Boston, US
October 7th, 2013
12 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
Measuring Social Resilience
The Autopsy of Friendster
Not power-law degree distributions
Prob[degree = d]
100
MySpace
MySpace
Facebook
10-2
10-4
10
A rumor can survive forever in a network
with a Power-Law degree distribution
ML estimation of Power-Law fit
Evaluation: Kolmogorov-Smirnov test
-6
Null hyp: degree distribution follows PL
100 101 102 103 104 105
Prob[degree = d]
100
degree
LiveJournal
Friendster
Orkut
Orkut
10-2
10-4
10-6
10-8
100 101 102 103 104 105
degree
Conference in Online Social Networks.
FS
LJ
FB
OK
MS
b
dmin
1311
88
423
171
2350
α̂
3.6
3.3
4.6
3
3.6
ntail
290K
81K
5K
280K
623
D
4.59
0.02
0.14
0.02
0.03
p
< 10−15
< 10−15
< 10−15
< 10−15
0.22
Kolmogorov-Smirnov on ML estimates
contradict eyeballing on PDF
Boston, US
October 7th, 2013
13 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
The Autopsy of Friendster
The Autopsy of Friendster
The Autopsy of Friendster
1
When Online Social Networks Die
2
Measuring Social Resilience
3
The Autopsy of Friendster
Conference in Online Social Networks.
Boston, US
October 7th, 2013
14 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
The Autopsy of Friendster
The Autopsy of Friendster
Limited connectivity patterns
4 x10 7
distance
2 x10 7
0
- 2 x10 7
- 4 x10 7
10 7
2 x10 7
4 x10 7
6 x10 7
8 x10 7
10 8
time [id]
10M user id bins - connections to future (red) and to past (blue)
Connectivity to future users is limited in time distance
Initial community was cohesive but did not bring many new users
Conference in Online Social Networks.
Boston, US
October 7th, 2013
15 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
The Autopsy of Friendster
The Autopsy of Friendster
P (L | Nact)
Microdynamics of users becoming inactive
10
0
10
-1
10
-2
10
-3
10
-4
10
We estimate user as inactive
when it does not create new
friendship links
Once a user starts losing
friends: when does it become
inactive?
0
10
1
10
Nact
2
10
Likelihood of becoming
inactive given amount of
active friends
3
Active friends as benefit
Users with less active friends were more likely to leave
Conference in Online Social Networks.
Boston, US
October 7th, 2013
16 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
The Autopsy of Friendster
The Autopsy of Friendster
Time series of OSN risk
P(ks < 6)
0.9
0.7
0.5
0.3
0
40M
t [id]
120M
80M
Amount of nodes below median of coreness (100K bin)
Shock at 20M (US collapse), decay after 80M
Conference in Online Social Networks.
Boston, US
October 7th, 2013
17 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
The Autopsy of Friendster
The Autopsy of Friendster
40
60
78% (52M)
20
15% (10M)
0
Google search volume (%)
80
Simulating Friendster’s collapse
2Jul2009
1Jan2010
date
2Jul2010
1Jan2011
Estimation of active users through Google search trend
Simulation of cascades with deteriorating conditions
R 2 = 0.972
Conference in Online Social Networks.
Boston, US
October 7th, 2013
18 / 19
David Garcia
Chair of Systems Design
www.sg.ethz.ch
The Autopsy of Friendster
The Autopsy of Friendster
Summary
1
We designed a rational model for users leaving an OSN
Decision depending on costs and benefits
Changes in costs and benefits create cascades that stop at k-cores
2
We measured the social resilience of five OSN
K-core decompositions show that no network is the best
Costs and benefits analysis is necessary to predict OSN success
Statistical analysis rejects power-law degree distributions
3
We analyzed the life and death of Friendster
Validated our assumption on user benefits related to active friends
Time evolution of coreness shows risky periods
Collapse is reproduced by the cascades of our model
Conference in Online Social Networks.
Boston, US
October 7th, 2013
19 / 19