Ibm james kobielus power of graph analysis_open analytics dc summit_march_ 25 2013


Published on


Published in: Technology
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Ibm james kobielus power of graph analysis_open analytics dc summit_march_ 25 2013

  1. 1. The Power of Graph Analysis: Graspingthe Shape of Business InfluenceJames Kobielus []Big Data EvangelistOpen Analytics DC Summit, Arlington VA, March 25, 2013 © 2013 IBM Corporation
  2. 2. Graph analysis can help you see deeply into influence patterns that you might in turninfluence for the better. © 2013 IBM Corporation
  3. 3. Influence is often invisible, but has concrete value Getting things accomplished demands effective use of influence. Organization charts dont tell the full story of business influence:  Internally: they flatter the chain of- command  Externally: they are next to useless Turning influence into a business asset demands that first you identify:  who has influence  over whom  with respect to what range of topics and decisions, and  with what potential impact on your own bottom line © 2013 IBM Corporation
  4. 4. Grasping influence = graphing engagement patterns • 360-degree view of individuals in full social context • Hidden, latent, nonobvious connections among people, groups, and organizations • Shifting patterns of connectedness, proximity, centrality, influence, status, and importance • Cooperation and collusion, coalition and co-dependency, influence and deference, and affiliation and isolation Dissect the interpersonal dynamics of communities. © 2013 IBM Corporation
  5. 5. Graph analysis: what is it?•Established branch of advanced analytics•Mining, categorizing, and predictinginteractions and influence patterns in theirfull social context•Leveraging many sources of data, including: •Behavioral •Transactional •Clickstream •Geospatial •Messaging •Smartphone •Usage •Call detail records •Social media posts•Discovering, segmenting, exploring,mapping, visualizing, forecasting, drilldown,what-if analysis•Sometimes known as “social networkanalysis” or “small world” analysis•Widely applied in social sciences © 2013 IBM Corporation
  6. 6. Graph analysis: visual modeling of influence patterns Same project flow as other advanced analytics modeling disciplines Data discovery Data extraction Data preparation Variable assessment and selection Classification Clustering Statistical analysis Regression Scoring Model deployment Can leverage specialized graph analysis toolsand graph databases, or more general-purpose tools and data platforms © 2013 IBM Corporation
  7. 7. Graph modeling: engagement proximity = influence “Nodes” = individuals “Links” = interactions and relationships “Graphs” = social networks of interactions and relationships among individuals “Attributes” = characteristics of individuals, relationships, and social networks 4 “Proximity” = number, type, and weight of 5 links that connect one individual to another 1 2 6 3 © 2013 IBM Corporation
  8. 8. Graph analysis: customer engagement Leaders? Followers? Influencers? Mavens? Mavericks? Cliques? Clusters? Outliers? Vanguards? Laggards? Late adopters?Gauge customer influence and engage differentially to boost customer lifetime value © 2013 IBM Corporation
  9. 9. Using influence graphs in customer engagements  CHANNEL INFLUENCE: Which customers are more influential in some channels than others?  PEER INFLUENCE: Which customers have the greatest impact on awareness, sentiment, and propensities of their peers?  CAMPAIGN INFLUENCE: Which customers promise the biggest bang for your marketing bucks through their influence on their peers?  LOYALTY INFLUENCE: Which customers influence the most peers to stay with you and grow the relationship — or churn and jump to the competition?  PROMOTER INFLUENCE: Which customers are net promoters and potential brand ambassadors?  ADOPTER INFLUENCE: What can influence users to encourage late adopters to accelerate their adoption of new products?  SUPPORT INFLUENCE: Which customers have valuable expertise and should be enlisted as community resources for inquiries and support?  NEGATIVE INFLUENCE: Which customers are the squeakiest wheels who are most likely to spread bad feelings if they are dissatisfied, or are most likely to keep it bottled up? © 2013 IBM Corporation
  10. 10. Graph analysis: employee engagementCOMMUNITY-OF-INTEREST INFLUENCE:How do communities of interest emerge,endure, and grow?KNOWLEDGE-SHARING INFLUENCE: Howdo cross-functional knowledge-sharingrelationship take root?PEAK-PERFORMER INFLUENCE: Whichtypes of individuals, subject-matter experts, orrelationships have the greatest influence onteam productivity?MENTOR INFLUENCE: How well arementors are fostering relationships betweenmentees and other employees?IDEATION INFLUENCE: Which people aremost effective at introducing, disseminating,and gaining adoption for new ideas?NEGATIVE INFLUENCE: Which people in adistributed team are acting as bottlenecks tocollaboration, knowledge-sharing, innovation,and ideation? Source: managementpocketbooks.wordpress.comAssess employee influence and create environment to boost performance, innovation, knowledge sharing, and ideation © 2013 IBM Corporation
  11. 11. Graph analysis: partner engagement STRATEGIC INFLUENCE: Which actual or potential partners have the the right degree of influence for realizing desired business outcomes? AGENDA INFLUENCE: Who within our organization or our partner has the influence necessary to disseminate and gain adoption of key messages, ideas, or agendas among key internal or external stakeholders? TEAMING INFLUENCE: Who within or organization or the partners has the influence needed to establish, strengthen, and sustain the teaming arrangement at the heart of the alliance? Source: Assess partner influence and create environment to boost strategic goals, agendas, and teaming arrangements? © 2013 IBM Corporation
  12. 12. How to use graph analysis to influence the influencers? •CUSTOMER ENGAGEMENT: •Engage with influencers on social media •Use graph analysis to assess degrees of customer influence •Establish differentiated engagement approaches and incentives based on degrees of influence •Experiment with mixes of all of these established and emerging influence channels, including but limited to social media •Compile the data, metrics and analytics needed to assess whether you achieved the expected return on your investments across various influence channels •Have an agile decision framework to help guide continual adjustments of your marketing mix among channels commensurate with their relative effectiveness EMPLOYEE ENGAGEMENT: •Use social graph analysis is to feed back results to the entire network in a structured setting in which discussion can be productively facilitated. •Viewing the graph analysis diagrams seems to be particularly powerful •Revealing these previously hidden barriers constructively can be a powerful tool for creating new opportunities for cross-group relationship development. •Graph analysis can also suggest ways of restructuring organizational charts or implementing new processes. •PARTNER ENGAGEMENT: • Use social graph analysis on partner organizations to assess where they and their employees stand in degrees of strategic, agenda, and teaming influence •Use these analyses as inputs into strategic partnering decisions © 2013 IBM Corporation
  13. 13. Thank You!13 © 2013 IBM Corporation