This document discusses how social graphs can transform enterprise applications. It begins with an outline and provides context on the shift from traditional software to social software. A case study on IBM's ThinkPlace collaboration platform is presented, showing how incorporating social graph features like user identities and connections improved idea sharing and discovery. The concepts of social and interest graphs are introduced. Five steps are proposed for transforming enterprise applications, including examining user data and relationships, creating an internal social graph, customizing applications, and integrating social features into existing workflows. Throughout, examples show how social graphs are already used by companies for applications like recommendations, talent management, and business intelligence.
(Original format) How Enterprise Social Graphs Can Transform Enterprise Applications
1. From Software …
to Social Software
How Enterprise Social Graphs Can
Transform Enterprise Applications
Kapil Gupta
2. Standard Disclaimer
Views and opinions discussed in
the following presentation are mine
alone, and do not reflect the views
of JP Morgan Chase or IBM
Where companies or vendors are
mentioned, there is no implied
endorsement by me or my
employer
3. Outline
Business & Technology context, and
why we should care
Case study: IBM ThinkPlace
Brief primer on Social Graphs
Implications for the Enterprise
context
5 Steps on the road to transforming
14. … all of which are in the sweet
spot for Social
Social Graphs and
Social Network Analysis
are the architectural foundations
for the use of Social technologies
21. When used together, applications
can start with who you know and
surmise what your connections like;
properly-designed interfaces and
interactions can provoke the
desired outcome
24. This is the promise of Social
Software
aka Social-enabled software
aka Social-aware software:
Software that can enable
discovery – and promote
outcomes – through
Recommendations
25.
26. Example: IBM ThinkPlace
A worldwide collaboration
platform to drive innovation
across the enterprise
My role at IBM (2005-2010) on this effort:
Product Manager/Product Owner
28. 1 2 3 4 5
Year
Businessimpact
1 2 3 4 5
Year
Numberofideas
Our problem wasn’t too few ideas: it
was actually too many
29. But
This raised the bar for finding
good ideas
Photo by Flickr user MayaEvening, licensed under Creative Commons, Attribution 2.0 Generic
30. Some good ideas
were seen, but
many remained
hidden
Photo by Flickr user giladr, licensed under Creative Commons, Attribution 2.0 Generic
31. Identity - who you are
Contacts - who you know
Activities - what you are
So we designed and implemented a
set of features across the system
designed around people
(vs. designed around ideas)
32.
33. The approach taken for ThinkPlace
can be applied to applications
across the enterprise, but
requires a set of infrastructure
components to support it
35. Enterprise Social Graphs can
bridge the gap between
“Systems of Engagement”
(interactions inside, outside firewall)
and “Systems of Record”
(data stores inside the firewall)
37. Visualization of my Linkedin Graph
Linkedin has a “Professional contacts
graph”
38. Just like them, every Enterprise has a
unique Social graph
… and
unique interest graphs
39. Relationships between Employees,
Clients, Partners, Competitors
Relationships can also be inferred
from enterprise meta data
(directory, email, collaboration tools)
or be user-submitted
40. Relationship information can be
augmented from external sources
(public graphs: Linkedin, Facebook,
Twitter)
Analogous to Integrated Supply
Chains, but for connections &
relationships: Integrated Supply
Graph
(of course, view of data is not shared
41. Most commonly cited enterprise
uses: (ad) targeting, retargeting
Customer journey: identifying
influencers; empowering advocates;
predicting customer behavior (via
interest graph changes)
Improving web experience: product
discovery, user communities
42. Talent management: identifying
“connectors”, leaders; as a
basis for referrals, evaluation
Business development &
Business Intelligence:
employee alumni in other
companies; new hires from
potential clients & competitors
Org charts of suppliers, clients,
43. 5 steps on the road to
transforming Enterprise
Applications into
Social Software
44. 1) Examine:
a) how your customers use your
products, how they interact with each
other and you
b) what kind of data already
exists in enterprise data stores
45. 2) Review your application
design and infrastructure stack
Your approach will have
significant implications for
application design, and the
software stack to support such
applications
46. 3) Create your own graph with all
known internal and external
relationships and data sources
47. 4) Customize application design
depending on the kinds of ties, and
types of interactions desired
Define social objects and interactions
Redesign UI experience
Leverage activity streams
48. 5) Seek to make existing
interactions easier and more
effective – avoid the trap of creating
a set of social-driven interactions
that exist separately
Two places to target: where goal of
user interaction involves
Discovery and
Collaboration