Big Data

www.pwc.com/it/digitaltransformation

Milan,4th December 2014

PwC headquarter How can Big Data help anInsurance Company?

Massimo Iengo – Director PwC

Speaker

Massimo Iengo joined PwC in September 2012 from Hewlett-Packard where he was working inthe WW Portfolio team as the Lead Architect for the global Customer Analytics Program.

Prior to this role Massimo worked at Generali group, Bain, Accenture and HP, where he had a keyrole in numerous technology innovation initiatives, and developed a strong experience andthought leadership on CRM, Customer Analytics, and Front Office Innovation, with a special focuson Financial Services.

Massimo Iengo

Director, Digital StrategyTel: +39 (02) 80646438massimo.iengo@it.pwc.com

1

PwC – How Big Data can help Insurance Company?

• Big Data defines a goal of transforming datainto insights and intelligence that are deliveredto those who need it, when and where it's needed, tomake and implement better strategic andoperational decisions.

• Big Data represents a new way of doingbusiness – one that is driven by data-baseddecision making and new types of product andservices enriched with data.

• Big Data means massive volumes ofenterprise-generated and third party data nowavailable (including real-time data streamed frommobile devices).

PwC – Point of view on Big Data (1/2)

By applying analysis of Big Data to pressingbusiness issues, companies are reshaping theiroperations and accelerating their businessresults. As its potential becomes more evident, BigData will transform every aspect of the organization,from strategy and business model design to marketing,product development, HR, operations and more.

In our Opinion:

Italian Companies that move quickly to capitaliseon the potential of Big Data will have theopportunity to gain ‘first mover’ advantage,enabling them to innovate in ways that aredifficult to replicate.

Our definition of Big Data

Value of Big Data

Market's opinion

Font: PwC’s 5th Annual Digital IQ survey was conducted on 1,108executives participated in the global survey.

PwC – How Big Data can help Insurance Company?

2

3

PwC – Point of view on Big Data (2/2)

Our latest global CEO survey found that insurers as a whole are just behind thetechnology, communications and entertainment sectors in their readiness to embracebusiness model innovation

A major life insurance company recognised that as the Internet advances – Along withcustomer needs and preferences – Its traditional direct model of insurance underwriting anddistribution could be threatened. Using Big Data, the company used predictive analytics to‘model’ the life insurance market, resulting in dramatic changes in how it markets,sells, underwrites and distributes its products.

New Business Model

Co

stre

du

ctio

nto

exis

tin

gp

roce

sses

PwC – How Big Data can help Insurance Company?

Big Data Use Cases

Case 3

Health Insurance – Fraudidentification

A global health insurance company hasexperienced a very high number of claims topremium ratio and with a high percentage offraudulent claims. Most of the fraud cases hadbeen through fraud rings indicating organizedfraud. Client felt that if fraud indicators aremonitored at industry level it gives them thescope to act in a more robust way.

Case 2

Standard Insurance – Dataappliance selection

A global insurance company was looking toevaluate data appliances as alternatives totheir current data warehousing environment.The main objectives were to lower costswhile maintaining the current service levelsand business continuity requirements of theexisting warehouse infrastructure

Case 1

Life insurance – Big Datacan see tomorrow today

A large life insurance company needed topredict the future in order to position itselfoptimally in the rapidly evolving onlineinsurance sales market.The company neededto know if the life insurance sales that werecentral to its growth were destined to followthe same evolution to direct distribution.

4

1 2 3 4

Issue Approach Project ActionsImpacts and

Outcome

Standard Big Data Project phases

PwC – How Big Data can help Insurance Company?

Case 1Life insurance – Big Data can see tomorrow today (1/2)

5

• A Fortune 500 insurance company that servesits customers with a wide variety of insuranceproducts, determined that fast-movingmarketplace changes would require a hard-hitting and sweeping analysis of its salesand marketing strategies in the onlinespace.

• It was critical for the company to avoidfinding itself flat-footed if the market shiftedsuddenly, and it planned to build a high-performance direct distribution operating modelto support future growth of online sales.

• Estimate the potential for selling individuallife insurance through the direct channel andto forecast sales for the next three to fiveyears. To find the answers, we collected andanalyzed vast Big Data sets that addressed threequestions:

How would new healthcare regulations andthe proliferation of electronic medical recordsimpact online sales?

How much marketing effort would it take tomake consumers feel comfortable shoppingfor life insurance in this new way?

And what kinds of upcoming technologychanges would make online sales moreviable?

Life Insurance Issue Project Approach

1 2 3 4

Issue Approach PwC ActionsImpacts and

Outcome

PwC – How Big Data can help Insurance Company?

Case 1Life insurance – Big Data can see tomorrow today (2/2)

PwC is assisting the company in developing amore data-driven decision culture as it combs

through more Big Data sets to predict othermarket changes that may shake up the

insurance industry in the future.

• PwC helped analyzed macroeconomic, consumerdata and technology advancement data

• Modelled the processed data within a timewindow of five to ten years, and discoveredthree potential barriers to market growth:

life insurance applications often requiresome kind of medical underwriting(medical reports)

consumers tend to be reluctant to sharetheir most personal medical informationonline

the complexity of some of the lifeinsurance products.

• The company estimated that by 2015, under certainscenarios:

it could achieve $200 million in direct termlife insurance sales

see substantial growth in its direct whole lifeinsurance market share

• The company is acting on these conclusions byenhancing its direct distribution group.

PwC Actions Project Impact and Outcome

1 2 3 4

Issue Approach PwC ActionsImpacts and

Outcome

PwC – How Big Data can help Insurance Company?

6

Case 2Standard Insurance – Data appliance selection (1/2)

PwC – How Big Data can help Insurance Company?

• A global insurance company was looking toevaluate data appliances as alternatives to theircurrent data warehousing environment. The mainobjectives were to lower costs whilemaintaining the current service levels andbusiness continuity requirements of theexisting warehouse infrastructure.

• Following an RFI (Request for information) process,the client selected 2 leading parallel database vendorsto conduct the Proof of Concept. PwC was engaged tohelp manage and lead the execution of the POCfollowing our proposed phased approach.

• PwC facilitated the development of POCrequirements and success criteria, coordinating withclient teams (application development, architects,database administrators, Unix administrators, BIreporting team, engineering, operations and datacenter teams) and both vendors.

Life Insurance Issue Project Approach

1 2 3 4

Issue Approach PwC ActionsImpacts and

Outcome

7

Case 2Standard Insurance – Data appliance selection (2/2)

PwC – How Big Data can help Insurance Company?

PwC team also provided recommendations toevolve the client warehousing environment tomeet their future operational and analytical

needs.

• PwC helped with the development of test casessyndicated with the vendors and client teamsand aligned with the POC evaluationcriteria.

• PwC assisted the client in managing theinstallation of the POC appliances in thedatacenter, and the execution of aggressive 6-week test plans for each vendor.

• PwC team addressed the issues during testimplementation, and tightly tracked progressand status.

• PwC aggregated results to facilitate clientdecision-making and led review sessionswith client teams to help them score vendorsagainst the POC success criteria. The projectconcluded with a POC recommendation.

• The client was able to conduct 2 POCs inparallel, thereby meeting very aggressive executiontimelines set by senior management.

• Decision-making for the appropriate dataappliance was greatly facilitated by leveragingthe recommendation documents, scorecardsand integrated test results analysis developed.

PwC Actions Project Impact and Outcome

1 2 3 4

Issue Approach PwC ActionsImpacts and

Outcome

8

Case 3Health Insurance – Fraud identification and management (1/2)

PwC – How Big Data can help Insurance Company?

• A global health insurance company hasexperienced a very high claims to premiumratio with a high percentage of fraudulentclaims. Most of the fraud cases had beenthrough fraud rings indicating organized fraud.

• The main actors of fraud losses are:

Policy holders

- Concealment of pre-existing disease

- Failure to report vital information

- Duplicate and inflated bills

Health care providers

- Billing for services not rendered

- Preparation of forged claims

Internal employees

- Participating in fraud rings

- Facilitating policies in false names

- Fudging data in group health covers

• Client felt that if fraud indicators aremonitored at industry level it gives them thescope to act in a more robust way. In this regarda lot of data has been collected at transaction levelthrough its data wing from various insurers and thirdparties.

• Client wants to churn the data based on variousfraud indicators and predictive analytics toidentify fraudulent customers, agents, employees,hospitals, doctors, drug stores and others.

• The results would be presented in form of reports,dashboards, alerts at various levels (insurers,industry, bank and market regulators, governmentagencies) in order to ensure frauds does notpercolate through financial services.

Health Insurance Issue Project Approach

1 2 3 4

Issue Approach PwC ActionsImpacts and

Outcome

9

Case 3Health Insurance – Fraud identification and management (2/2)

PwC – How Big Data can help Insurance Company?

• Data exploration

• Business rule definition - identify fraudindicators

• Identification of fraud flag - a fraud flagwas assigned (dichotomous variable)based on a composite score which wascoming out on individual rules that wereset up

• Statistical modeling – iterated differentlogistic regression method

• User experience

reports and dashboards

Alert generation and casemanagement

PwC Actions

11

Project Impact and Outcome

1 2 3 4

Issue Approach PwC ActionsImpacts and

Outcome

The results would be presented in form of reports,dashboards, alerts at various levels (insurers,industry, bank and market regulators, governmentagencies) in order to ensure frauds does notpercolate through financial services.

Case 3 – Focus on Solution architecture

PwC – How Big Data can help Insurance Company?

10

Rule engine

Review transactions against Medical ProtocolDatabase

Cross checking medical procedures expensedagainst standard medical procedure

Compare claims amounts to industry billing standard

Analytical model

Scoring model to rank fraud risk level and prioritize review ofhighest risk cases

Track propensity for aberrant billing at each stage of submission Black List fraudulent addresses/companies to red flag for higher

scrutiny

Fraud reporting

Feedback loop toenhance predictivepower model with realtime changes intrends

Weekly monitoringreports andSegmentationDashboard of topcontributors

Standardizationand control of casemanagement