Social Data Lab

Summit 2011.Q2 -- Future of Insurance

Summary and Purpose

An examination of how social data is impacting the insurance industry and how practices are adapting to those changes. Participants were introduced to entrepreneurs, startups and innovators influencing the way people are sharing data, and examined many issues facing the ever-changing landscape of insurance.


Social Data Revolutionary/SDL Director

- Andreas Weigend (

Allstate Representatives
- Diana Yi

- Marty Epstein

- Pauline Velez

- Peggy Brinkman

- Steve Yun (Workshop only)

- Tom Warden (Workshop only)

Tapiola Representatives
- Anna Tapio

- Anne Pallaste

- Matias Ruponen

- Ville Eskola

Social Data Lab Members (Summit Only)
- Alan Guo

- Aldo Briano

- Alex Cheng

- Jason Yu

- Karthik Venkateswaran
- Sebastiaan Boer

- Suthinand (Pao) Jirakulpattana

External Domain Experts (Summit Only)
- Alicia Morga (Reflecta)

- Greg Wolff (UnaMesa)

- Johannes Suikkanen (Gemic)

- Mike Sha (Wikinvest)

- Phebe Wong (City of SF)

- Sakari Tamminen (Anthropology PhD from York)

Detailed Agenda

Friday, May 13th, 2011 -- SAFARI & SUMMIT

  • 0900 Departure from Crowne Plaza Palo Alto
  • 0930 23andMe: Health Insurance, Genetics and Social Data, Marisa Nelson (Business Development Associate)
  • 1045 Depart for HealthTap
  • 1100 HealthTap: Medicine and How People Express Themselves, Ron Gutman (Founder and CEO)
  • 1230 Lunch at Stanford
  • 1400 Social Data Summit at Stanford
  • 1430 Introductions and Expectations, Social Data Lab Members (Stanford students)
  • 1500 Workshop
  • 1715 Presentations by Lab
    • The Power of Social Data (Karthik)
    • Social Data Intelligence Test (Aldo,Tim)
  • 1830 Dinner at Three Seasons

Monday, May 17th, 2011 -- WORKSHOP

  • 1450 Andreas Weigend
  • 1540 Tom Warden - Overview of Social Data work at ARPC & Allstate
  • 1600 Tapiola
  • 1620 Coffee / Cookies-Pastries (defining break out topics)
  • 1630 Break out
  • 1715 Rapporteurs present findings
  • 1745 Dr. Weigend's Summary
  • 1815 Transfer to San Francisco for dinner


Participants were introduced to some of the exciting startups that are influencing the way insurance is doing business. The startups were focused predominantly around the topic of health and data.


  • 23andMe
    • They aim to help people understand what their genes mean by indexing them and highlighting significant findings. 23andMe allows its clients/users to study their ancestry, genealogy, and inherited traits. The company also markets to researchers and scientists, for whom they provide neatly categorized and easily searchable data.
  • Healthtap
    • HealthTap is an Interactive Health company passionately dedicated to improving your health and well-being, and to improving the overall process of care for you and your doctor, while reducing costs.



  • What is the value of data?

How do we measure the value?

“Equation of the business”

  • Why do people share?

Self expression

Attention seeking

Need to belong

  • What does true customer centricity mean, and how do social data make a difference?

CRM as part of the managerial economy, vs customer centricity?

  • Behavior change

What makes people change their behavior (emotions)?

How can we get people to understand trade-offs?

In the digital world, you can share information with people you don’t know.

  • Trust

What builds trust?

What destroys trust?

Asymmetry / Balance


  • Social norms

Digital vs physical (O2O: online to offline)

  • What has changed?






Startups that might embed learning that we can tease out for insurance

Color, Hunch: generalized model of consumer prefs

App to communicate with the agent

What could be shared that is useful for agent and not negative for customer

What be cases where customers share with an institution?

Different levels of anonymized data (readings Cynthia Dwork paper + video,, esp How To Break Anonymity of the Netflix Prize Dataset : "The dataset is intended to be anonymous, and all customer identifying information has been removed. We demonstrate that an attacker who knows only a little bit about an individual subscriber can easily identify this subscriber's record if it is present in the dataset, or, at the very least, identify a small set of records which include the subscriber's record." [Narayan and Shmatikov, 2006]

matching “anonymized” data to non-annonymized with some degree of confidence so that the information ratio in the match overcomes the noise the lack of good matching creates

Designing incentives / business model such that people give truthful answers
Behavior change: Drivers young and old

How do behavioral beliefs change over lifecycle (psych principles): What motivates young and old ppl to change their behavior? (In our case, to share data)

e.g., postitive vs negative rewards

given by whom?

time scale: minutes or years

Students to think one level higher up

In general terms (e.g., to find ways for customers to friend their agent)

general behavioral trade vs specific

Insurance specific questions:

Why data is important (information asymmetry)

Data = raw material

Data within the company

Collected by company

Is the data correct?

Cost of data: Cost of false data, cost of correcting data

Collected by user

Data that sit elsewhere

Social norms I.

Thought experiments

Assume everything recorded... how would it change your behavior?

What does it mean to own data?

Data creation / collection / sharing / distribution / consumption...

Incentives for user to contribute

esp truthfully


Mechanics of processing

“Big data” -- is that an issue?


Car fleets, etc....

Social norms II


Agent based

Financial engines (simulation), 23andMe,

Amazing variety of products.... vs essentially one

Illusion of choice?

Trust, self-service



Impact on the organization

New product development processes?

Marketing / market research

What skillset do they need?

Hypothesis generation about customer behavior

Decision map


Give them hope

Teach them how to fish



  • How can we create stuff consumers truly value, from their data? Getting them to truthfully create and share data
  • What now, what later?
  • How can social data be used to reduce claims: change behavior of existing customers, reduce p(bad customers)
  • Reactive --> proactive
  • Descriptive --> predictive
  • Knowing --> doing
  • Measuring behavior --> Changing behavior
  • Key: GOALS ARE ALIGNED in the global pic
  • Where are opps for true innovation (increasing the size of the pie), vs zero sum game (“rearranging deck chairs”)

What really has been the main revolution: Economics of communication

  • Past: Means of production owned by gatekeepers
  • Present: means of data production and distribution commoditized.

What is scarce?

  • Storage?
  • Creativity?
  • Attention!
  • Scarcity defines the currency à attention economy

Digital social fabric

  • Weaving individuals together
  • Sensors for insurance

Design principles for the attention economy

How to get it done in the organization

How to evaluate a social data initiatives

  • Customer centricity / Value back to customer / Helping cust to make better decision
  • Does your prod get better or worse over time?
  • Does your Customer relations dept feed into product /

Do experiments

  • Know your metrics
  • Apply the learnings


  • Design metrics that actually matter based on social data



  • Help people collaborate / coordinate in new ways (discovering people)
  • How can your product help people make better decisions
  • How do you design your product to make it easy for people to contribute
  • Feedback / learning from mistakes (highly granular, ads on amazon to competitors)

Social playgrounds

  • Nike+ vs Nike
  • Build social playgrounds so people create and distribute “their” data

Old-style customer centricity à transparent, true customer centricity

  • The revolution happens in the minds of the customers
  • Impact on how the individual customer is treated by companies:
  • Transparency / one-way mirror à two way mirror
  • Based on social data (geolocation, likes etc), SDR enables new ways of customer centricity

Influencing the customer / Recommender systems:

  • Mobile
  • Social (connection, social commerce)
  • Context (had a kid, ready to move)
  • Find regularities, modeling intention, relationships

New possibilities for behavior change

  • Based on social data, people behave differently as individuals, and societal norms are shifting
  • Interviewing / hiring process shifting (requiring facebook)
  • Frame in context of SDR. Sharing data changes behavior
  • Reason: feedback, attention, audience
  • Easy goals, TINY STEPS
  • Faster feedback loops where short terms is possible

What can individuals achieve on their own vs collaboration: How insurance can leverage the social graph


  • Group of friends / my group of friends
  • Group reinforcement / social dynamics
  • 1) Social contract with a group where people know each other
  • 2) Socialized commitments
  • 3) Fairer pricing given the characteristics of the group
  • 4) Social marketing / social filters
  • Level of granularity
  • What needs to be personally identifiable?
  • What can be done in the aggregate?
  • What do we need to do to facilitate this?
  • What can we do to create a social playground?
  • Control
Last modified: July 1st, 2012 at 21:48

Social Data Lab

Powered by GroupSpaces · Terms · Privacy Policy · Cookie Use · Create Your Own Group