Emerging Issues: Social Media Analytics in Insurance. April 14, 2015

Transcription

1 Emerging Issues: Social Media Analytics in Insurance April 14, 2015 Insurance Services Office, Inc.,

2 Socially informed P&C decision-making Are they telling me the truth? What are they saying about me? What does it all mean? Where do customers socialize? Who is my ideal customer? How can I make more money? How do I connect with him or her? Marketing and Fraud Lifetime Value and Insurer Brand 2

3 What s now? What s next? 100% 90% 80% 27% 35% 37% 70% 55% 60% 50% 40% 22% 22% 37% 84% 30% 20% 51% 43% 26% 10% 26% 18% 12% 0% Brand and influence Marketing and acquisition Product strategy and actuarial Underwriting and point-of-sale 5% Claims management Other Greater than six months Immediately or within six months Source: ISO Social Media Survey (August 2014), n = 44 3

4 Case study: Personality-driven options Persona cluster analysis Personality/preference quiz Price sensitivity Service preference Carrier type Payment preference Awareness/interest in UBI Matching formula Source: InsuraMatch Online Super Agency Offers Personality-Driven Options 4

5 Likes predict better than friends/family 5

6 Your personality in the cloud 6

7 Usefulness based on quality, quantity 40 Number of respondents indicating usefulness Geolocation (e.g. Foursquare) Reviews (e.g. Yelp) LinkedIn Bookmarking (e.g. Reddit) Pinterest Instagram Twitter Forums (e.g. Tumblr) Google+ Facebook* YouTube* Podcasting Short-form video (e.g. Vine) Snapchat Emoji.li/Emojicate Approximate number of active monthly users Source: ISO Social Media Survey (August 2014), number of respondents = 44 * - Facebook and YouTube s user bases have been scaled down in order to fit on graph 7

8 Big data fits on a single page James J. Basic Info Sex: Male Birthday: MYOB Hometown: Anywhere Status: Just Married! (This example is completely hypothetical and intended for illustrative purposes only.) Groups House Hunters Proud Pug Owners Straphangers Fanilows! Midnight Munchies Friends Jim has 314 friends including: Christopher P. Daniel D. Anitha T. Hilary J. Shalini W. see more Contact Info jj@ .com Phone: Personal Information Activities: Fine dining, hunting (for bargains) Interests: Technology, sartorialism, cinema Work History Employer: Insuranceplex Position: Actuarial Ninja Tenure: 2003 Present Recent Activity James J. said: Just installed a gadget in my car to save $ on car insurance will need the $ for new houses we re looking at! Chris P. commented: Your a dork. 8

9 Model-ready social media data Hypothetical machine learning example Engage with insurer on social media? Yes No Positive Neutral Negative Posts on social review sites? e(d) = 96 months e(d) = 18 months What are hobbies and interests? No Yes Active lifestyle e(d) = 63 months Sedentary lifestyle e(d) = 38 months Yes Reviewed on commerce sites? No Mostly positive reviews Mostly negative reviews e(d) = 78 months e(d) = 32 months Mostly positive reviews Mostly negative reviews e(d) = 44 months 9

10 Data uniquely suited for LTV Source: ISO Social Media Insurer Focus Group (December 2014) 10

11 Commercial lines case study Danny s Diner Reviews: 860 $$ Italian-Mexican Fusion Most Helpful Reviews James J.- Danny s Diner devastatingly disappoints! Anitha T.- The two restauraunts that surround this place are much better. Don t waste your $$!! Shalini W.- There must have been a kitchen fire in this place because when I went it REEKED of smoke! Business Info Hours: 5:00 pm 2:00 am 2013 Sanitary Inspection Grade: F 2014 Sanitary Inspection Grade: C Show more info (This example is completely hypothetical and intended for illustrative purposes only.) 11

12 Personal touch in commercial lines 7 6 Respondents Indicating Usefulness Marketing and distribution Premium audit and underwriting Ratemaking and Pricing Loss control and claims Commercial Property and Auto Workers Compensation Medical Professional Liability Liability (General, Cyber, D&O) Source: ISO Social Media Insurer Focus Group (December 2014) 12

13 Good, bad, or indifferent? Day 182 of our claim with [Insurer]. Jinkies78 Day 246 of our claim with [Insurer]. Jinkies78 I think today is day 291 of our claim with [Insurer] since we all lost our house and belongings. A perfect driving record, very few miles put on my car and you RAISE my insurance rates?! What kinda logic is that?! -- ericburdette 13

14 Brand in a bubble Number of tweets per year Insurer A Insurer B Insurer C Insurer D Insurer E Powered by Unmetric 14

15 Relating brand to business performance Facebook likes (columns) and shares (points) per $1,000 of advertising expense A F G B D C Powered by Unmetric 15

16 What did they say to get on this list? Insurer annual revenue for most-engaged P&C Academy Award tweets 5+B 5+B 50+B 5+B 5+B Specialty 5+B 20+ B South American Insurer annual revenue for most-engaged P&C Super Bowl tweets 50+B 50+B <5M Regional 25+B <5M Regional 50+B 25+B Source: Unmetric event pulse ( 16

17 Insurer value lost in translation Percentage of followers A F G D C Powered by Unmetric 17

18 Unmetric: the only social media intelligence platform focused on brands Unmetric is not a tool for listening to what audiences are saying, nor is it a tool for publishing content. Instead, Unmetric works along with such tools to create what we call an Intelligent Workflow. 18

19 Average life of social media data 7 6 Respondents Source: ISO Social Media Insurer Focus Group (December 2014) 0 < 6 months 6 months - 1 year 1-2 years 2+ years POS Fraud Customer Lifetime Value UW Tiers Claims Fraud 19

20 Social media data is fast data Insurer case study Number of Tweets mentioning Insurer Day Number 20

21 Social media after the storm Geo-location features Hot spots and scattering Cluster analysis Field crew mobilization Impassible roads Power failures Other issues affecting service Source: Grabbing Hurricane Irene by the tail: Insurers around the world use interactive mapping and modeling to understand impacts of hurricanes and other disasters. 21

22 First movers get first predictions Industry not leveraging full potential of social media data Lifetime value and advanced brand analytics appear next on horizon Industry specification necessary for tools to result in optimal decisions Potential first mover advantage for insurers who can build fastest pipelines to data 22

23 Questions? No part of this presentation may be copied or redistributed without the prior written consent of ISO. This material was used exclusively as an exhibit to an oral presentation. It may not be, nor should it be relied upon as reflecting, a complete record of the discussion.. Insurance Services Office, Inc., 2015