Download - 201206 IASA Session 673 - Mining Social Data
June 2012 1
Mining Social Data to make informed Risk EvaluationsSession 673 Tuesday June 5, 2012 3:30 PM
June 2012 2
Survey Results (Thank you to those who participated)
Does your company have a formal Social Media strategy?
Does your company have a formal Social Media use policy for employees?
June 2012 3
Survey Results (Thank you to those who participated)
How well does your company currently use each of the following? 0 = Not at All, 1 = Fairly Well, 2 = Well, 3 = Very well
June 2012 4
Survey Results
What is the main purpose for your company’s website?
What is the main purpose for your company’s use of LinkedIn?
June 2012 5
Survey Results
What is the main purpose for your company’s use of YouTube?
100% of respondents answered “Other/Na” for Pinterest.What is the main purpose for your company’s use of Blogs?
June 2012 6
0% 20% 40% 60% 80%
AdvertisingRelationships
LeadsServicingOther/Na
0% 20% 40% 60% 80%
AdvertisingRelationships
LeadsServicingOther/Na
Survey Results
What is the main purpose for your company’s Facebook page?
What is the main purpose for your company’s use of Twitter?
June 2012 7
0% 20% 40% 60% 80%
AdvertisingRelationships
LeadsServicingOther/Na
0% 20% 40% 60% 80%
AdvertisingRelationships
LeadsServicingOther/Na
Survey Results
How are you capturing your online chats or posts?
NOTE: Negative posts / blogs / tweets can be considered to be “complaints” under insurance department regulations and could require the same logging and reporting as if written or called in.
June 2012 8
Survey Results
Support / Staff availability hours for online presence:
The average number of staff supporting social media strategy was 1 after discounting an answer of 160 that was probably total service staff.
June 2012 9
40%
20%
20%
20%
No Presence
Multiple BusinessTime Zones
Local Business TimeZone
24 Hours
Survey Results
Capturing analytics on social media site(s) usage:
Capturing demographics about users of social media site(s):
June 2012 10
0% 20% 40% 60% 80% 100%
Marketing
Underwriting
Customer Service
Claims
Survey Results
Tracking ROI for the use of Social Media:
Capturing customer information from social media site(s):
June 2012 11
0% 20% 40% 60% 80% 100%
Marketing
Underwriting
Customer Service
Claims
Survey Results
Monitoring company reputation across the internet:
Which departments in your company use information collected from social media?
June 2012 12
0% 20% 40% 60% 80% 100%
Marketing
Underwriting
Customer Service
Claims
0% 20% 40% 60% 80% 100%
MarketingUnderwriting
Customer ServiceClaims
Survey Results
How does your company use mobile apps:
100% of Respondents stated they attempt to collect email addresses from customers.
27% of Respondents stated they attempt to collect Facebook account name.
June 2012 13
0% 20% 40% 60% 80% 100%
Marketing
Underwriting
Customer Service
Claims
LIMRA’s Research on Social Media use by Insurance Carriers
June 2012 14LIMRA 2012 Life Insurance Conference
85%
65%
54%
65%
0%
0%
98%
95%
81%
88%
45%
15%
0% 50% 100% 150%
YouTube
Google+
Tumblr
2011 2010
Currently On or Plan To Be
65%56% 59%
47% 43%
0%10%20%30%40%50%60%70%
EXPAND BRANDAWARENESS 56%
Companies Expanding Presence
Social Data (not Social Media and not Social Networks) comes from all over
June 2012 15
AcxiomDun & BradstreetISOLexisNexisMerckleMIBMillimanNeustarPolkRiskmeter
Grocery store rewards programsFrequent guest and Frequent Flyer programsCredit Card purchasingOnline purchasing – books, movies
Social Data is constantly evolving
June 2012 16
Social Data is Being Added in Immense Volumes daily:
§ 66% of adults and 75% of teens are content creators on the internet § 66% of internet users are social networking site users § 55% share photos § 37% contribute rankings and ratings § 33% create content tags § 30% share personal creations § 26% post comments on sites and blogs § 15% have personal websites § 15% are content re-mixers § 14% are bloggers § 13% use Twitter § 6% use location services–9% allow location awareness and 23% use
maps etc. Source: Pew Research
Social data is more than the data, it is the data and the relationships – that’s what makes it “social” data, why it is complex and unstructured, and how it differs from simple data.
Social Data Quadrant Map for Use
June 2012 17
Everyone is getting into Social Data
June 2012 18
Social Intelligence –Insurance Solutions for Social Data
June 2012 19
Risk Areas Key: Prospect, Underwriting and Claims
June 2012 20
Recent research indicates that 24 percent of insurance companies are evaluating using social data in claims and 26 percent are evaluating it for underwriting.Source : SMA
Case Study: Prospect Scoring
June 2012 21
PredictiveAnalysis
and Modeling
Low Medium HighPropensity to Convert
High value,Low
conversion, 2nd Priority
High value, Medium
conversion, Top Priority
High value, High
conversion, Top priority
Good value, Low
conversion, Low Priority
Good value, Medium
conversion, 2nd Priority
Good value, High
conversion, Top Priority
Low value, Low
conversion,Low Priority
Low value, Medium
conversion, Low Priority
Low value, High
conversion, 2nd Priority
Potential Value
Low
Med
ium
H
igh
Potential Future Value of Customer
Scoring of prospects based on conversion and value, marketing strategy developed to match
Survey Data
Web LogData
TextData
Purchased Data
Psycho-graphic Data
The Addition of Social Data to Score Prospects
Hobbies and Extreme SportsRelationshipsActivities and CalendarTravel CommentsHome Repair / Construction UpdatesPersonal Family UpdatesGPS Coordinates of daily tripsTweets on political and organizational affiliationsBlog comments – what blob as well as contentReligious and community affiliations
June 2012 22
Web LogData
Psycho-graphic Data
Case Study: Target Retention StrategiesStep 1: Determine Life time Value
June 2012 23
Time of Purchase Demographics -Loses predictive value over time as relevance is superseded by inforce behaviors
Customer behavior shifts focus from current to future value
Predictive Analysis
Current Value
Future Value
Post Purchase Activity –Increases in predictive value over time as behavioral patterns develop –IntegrateSocial datahere
Case Study: Target Retention Strategies Step 2: Predict Potential Lapse
June 2012 24
Predictive Analysis –
Model Channel and
Consumer Behaviors
Source of Business influences lapse tendencies based on channel behaviors
Transaction behavior influences lapse tendencies based on consumer behaviors
Web LogData
Supplement withSocial data
Case Study: Target Retention StrategiesStep 3: Develop Strategy Matrix
June 2012 25
Match effort to risk and value –
• High value low risk gets medium effort, save money on retaining low risk customers
• Low value customers get low cost efforts across the board
• Targeted high efforts on high value / high risk
Case Study:Life Underwriting via App + Social Data
June 2012 26
Second child born last yearHigh investment risk toleranceLived in home 2 yearsOwns homeCommuting distance 1 mileReads Design and Travel MagazinesUrban single clusterPremium bank cardGood financial indicatorsActive lifestyle: Run, Bike, Tennis, AerobicsHealth food choicesLittle to no television consumption
Actively pursue for issuance of a preferred policy without requiring fluids or medical records.Use strong retention tactics.
Life UW using a GLM predictive model to assess risk:§ Use info on app plus social data, No fluids or files§ Integrate 3rd party publicly available information.
Case Study:Life Underwriting via App + Social Data
June 2012 27
Do not send offers. Do not pursue aggressive retention strategies. If applies, pursue additional medical records and tests.
In a test over 30,000 applicants, behavioral and lifestyle factors provided 37% of the risk assessment influence and performed as well as additional, more intrusive medical tests and fluids.
Current residence four yearsLived in same hometown 15 yearsCurrently rentingCommuting distance 45 milesWorks as administrative assistantDivorced with no childrenForeclosure/bankruptcy indicatorsAvid book readerFast food purchaserPurchases diet, weight loss equipmentWalks for healthHigh television consumptionLow regional economic growthLight wine drinker
Types of third party marketing data
June 2012 28Deloitte Predictive Model for Life
Life Underwriting Savings:Using 3rd Party Data versus Medical Data
June 2012 29Deloitte Predictive Model for Life
Workers Comp already has a track record of using Social Data
June 2012 30
Claims Analytics:Fraud Red Flag Dashboard
June 2012 31Courtesy of Attensity
Social Analytics: Customer Engagement Dashboard
§ Automatically monitor social conversations
§ Filter out irrelevant posts
§ Analyze posts to extract key insights
§ Engage customers with proactive outreach
§ Improve the experience customers are having on the site
§ Improve brand image and emphasize the legitimacy of business
June 2012 32Courtesy of Attensity
Social Analytics: Conversation Sentiment Tracking
June 2012 33Courtesy of Attensity
Social Analytics:Website Sentiment by LOB
June 2012 34Courtesy of Attensity
Social Analytics:Overall Sentiment Ratings Dashboard
June 2012 35Courtesy of Attensity
Social Analytics:Competitive Sentiment Dashboard
June 2012 36Courtesy of Attensity
Contact Information
Robert E. Nolan CompanyManagement Consultants
www.renolan.com
Steven M. Callahan, CMC®
Practice Directorwww.linkedin.com/in/stevenmcallahan
June 2012 37