Description
Tools for Performance Metrics and Business Intelligence
Tools for Performance Metrics
and Business Intelligence
October 12, 2007
Mo Masud
Lisa Wester
2
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Agenda
Business Intelligence and Reporting In Action
Steps to Realize Business Intelligence Opportunities
Measuring Model Impact through Business Intelligence
Rationalizing Performance Metrics and Business Intelligence
3
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Key Discussion Themes
Practical Ways to Measure Predictive Modeling Impact
Designing New Ways to Measure Strategic Choices in a Predictive
Modeling Enabled World
The Role of IT and Decision Support Systems in Predictive Modeling
Aligning Strategy with Predictive Modeling Results to Maximize Benefits
Steps to Get Started
Predictive Modeling is a Continuous Process
Going Beyond Model Development to Gain a Competitive Edge
4
Copyright © 2007 Deloitte Development LLC. All rights reserved.
The Time is Now for Going Beyond Just Model Development
Looking into the future
Analytics is allowing organizations to stay ahead of the competition and ahead of the market
The predictive modeling story resonates well with all parties – consistent, objective, reduced
expense, accurate
Analytics is becoming a core competency and organizational strategy – competing on analytics
Data driven
• Ability to store, process and distribute information faster than
competitors will differentiate organizations and give them a
competitive advantage
• Data quality and data governance is a top IT priority
• Developing fluid business strategies based on predictive
modeling results will set part the market leaders
Rationalizing Performance Metrics and BI
5
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Rationalizing Performance Metrics and BI
Attributes of companies that are analytically-oriented
One or more senior execs who are strongly advocating analytics and fact-based decision making
Widespread use of descriptive statistics, predictive modeling and complex optimization techniques
Substantial use of analytics across multiple business functions and processes
Movement toward an enterprise-level approach to managing analytical tools, data, and organizational
skills and capabilities
Stages of analytic competition
1. Facing major barriers (organizational, technical)
2. Analytic progress is made only locally, and not spread to other parts of the company
3. Value and promised realized, but unsuccessful implementation of analytic strategies
4. Vision has been realized, but enterprise-wide adoption has not been undertaken
5. Analytic competition has been built into company strategy
Source: Competing on Analytics, Thomas H. Davenport, Don Cohen, and Al Jacobson
6
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Rationalizing Performance Metrics and BI
Predictive Modeling in the Property & Casualty Insurance Industry is a Now a Core Business
Strategy
Several Carriers have already
implemented Predictive Modeling for
commercial business
– Primarily targeted at non-renewals and
right pricing
– Customized models
– Large financial benefits
Several carriers that have already
developed modeling capabilities are
rapidly advancing their capabilities
– Broader adoption in the industry
– Numerous production systems
– Expansion into account models
– Broader business applications
– Tighter linkage to business rules
– New players are entering the space
Companies must go beyond just model
development
– Standard practice in leading companies
– Will be seamlessly integrated with business
rules to drive results
– Early adopters will have leveraged capability to
gain market share
– Extension to non-standard lines and more
creative applications
– Generic LOB tools will emerge
– Integration of models across functions (Claim,
U/W, Marketing)
– Ability to process results faster than
competitors will distinguish market leaders
Time
3+ Years Last 3-5 years 1 – 3 Years
Innovators
Early Adopters
Early Majority Late Majority
Laggards
Yesterday Today Future
7
Copyright © 2007 Deloitte Development LLC. All rights reserved.
“85% of our new products
are automated with
predictive modeling, which
enhances underwriting
consistency and makes it
easy for our commercial
agents to do business with
Safeco”
“Fully launched Customized
Pricing, our predictive pricing
model that automatically
provides the most appropriate
price for a new small business
submission”
“Achieved overall written
premium growth of 4%.”
AIG’s Mr. Purdy who
describes predictive modeling
as “evolutionary” said, “I
expect that every insurer will
be using WC data they
collect to do predictive
modeling at least by 2010.”
“Modeling is now moving to a
higher level of sophistication”
“Commercial lines NPW grew
3% for Q2, driven by $83M in
new commercial business,
up 13%compared to Q2
2006.”
“The dramatic improvement is
a direct result of our
multidisciplinary WC
improvement strategy and
predictive modeling”
“Because our comprehensive
use of predictive modeling
is now analyzing 80% of our
commercial lines business,
both new and renewal
accounts, we are able to
better match price to risk. As
result we feel we can continue
to improve our profit margin
[…]”
“As Maryland continued to
experience an increasingly
soft market, we retained 86%
of expiring premium”
“Predictive modeling /
automated underwriting were
implemented to enhance
service delivery.”
“The quality of our property,
WC, auto, and specialty mix
is continuing to improve as
we use predictive models”
“Agents are giving us
preferred shelf space vs.
weaker competitors”
“Increased ease of use
through faster decisions,
streamlined processing, and
expanded account rounding”
“Reduced system quote to
issue time through dynamic
questions, improved agent
interface, and automated
UW.”
Rationalizing Performance Metrics and BI
Several companies that have implemented Predictive Models are raving about the impact
that it has had on their business.
8
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Banking - With SPSS, HSBC
Bank USA effectively mines
an ever-growing file of
customer data, creating
predictive models to uncover
cross-selling and "roll over"
sales opportunities. Focusing
on the best prospects for each
product helps maximize sales
and minimize marketing costs,
and SPSS' ease of use helps
researchers deliver
intelligence faster to decision
makers.
Retail - Using the Wal-Mart Model, Paul Westerman points
out Wal-Mart's use of store and product traits to guide
replenishment. Simplified, 'beach' products are assigned to
'beach' stores. But how many snorkels should go to a
particular store? Paul advises that "a data warehouse can
provide a good estimate based on another, similar product
that has the same distribution."
Market basket analysis reveals these complementary
products. Combined with demographic data, you
understand the market forces at work (e.g., a correlation
with young affluent families) and send masks and flippers to
stores in neighborhoods with swimming pools, too.
Through careful planning, your past and future converge to
tackle today's problems.
Healthcare - The new way is
predictive modeling, and it is being
implemented at health plans left and
right.
"The goal of a clinical strategy is to
find the right intervention for the right
person at the right time," says Carol
McCall, vice president of Humana's
Center for Health Metrics, "and
ultimately, you want to be able to
understand enough about people to
be able to literally custom-tailor the
exact intervention so that you can
meet them in a way that they will be
the most receptive to it." McCall is a
fellow of the Society of Actuaries and
a member of the Academy of
Actuaries.
But if everyone in the industry is doing it, how can companies differentiate themselves
The answer may lie at looking at leading-edge companies in other industries
Rationalizing Performance Metrics and BI
9
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Rationalizing Performance Metrics and BI
• The strongest statistical models will not offer clients any benefits to their bottom line unless it is
seamlessly integrated within their technical and business infrastructure
• In conjunction with a integrated predictive modeling solution, clients must realign their business
strategy to capitalize on their predictive modeling solution
• Operating in a predictive modeling enabled organization is a new way of doing business and
requires new performance metrics to measure the success of the predictive modeling investment
• Realizing predictive modeling benefits is a continuous process that is evolving to meet the demands
of the external market
• Business intelligence tools are the means to continuously monitor and measure the impact of
predictive modeling enabled strategies
Enhanced Decision Support Capabilities Gives Organizations a Competitive Edge
Measuring Model Impact
through Business Intelligence
11
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Achieving Continuous Predictive Modeling Improvement
Continuous
Predictive
Modeling
Improvement
Predictive Model Development
The foundation of workers’ compensation analytics is
the design and development of predictive models
Business Process Redesign -
Redesigning underwriting
workflows, business rules and
business processes to enable
predictive modeling results based
decision-making
Scoring Engine Development
The technical manifestation of the
predictive models transforming raw
data into predictive model scores and
underlying reason codes
Technical Integration
The integration of the
predictive modeling results
produced by the scoring
engine within front-end
underwriting applications
Performance Monitoring
Decision Support Framework
A framework to enhance
enterprise wide
information processing,
data capture, and
performance
Defining reporting requirements and aligning
requirements across business units, and
performance metrics to measure the
effectiveness of the predictive modeling
enabled business decisions
12
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Aligning strategy with predictive modeling results is the first step to maximize benefits
D
e
p
l
o
y
A
n
a
l
y
z
e
D
e
v
e
l
o
p
S
e
l
e
c
t
IT
U
n
d
e
r
w
r
i
t
i
n
g
M
a
r
k
e
t
i
n
g
Operations
Predictive Modeling Enabling Business Strategies for Success
Underwriting Excellence
• Improve pricing precision
• Increase objectivity throughout the
underwriting process
• Enhance risk selection and risk
avoidance capabilities
• Improve pricing competitiveness in
profitable segments
• Improve underwriter negotiation
capabilities
Underwriting Excellence
• Improve pricing precision
• Increase objectivity throughout the
underwriting process
• Enhance risk selection and risk
avoidance capabilities
• Improve pricing competitiveness in
profitable segments
• Improve underwriter negotiation
capabilities
Operational Efficiency
• Reduce transaction costs
• Straight through processing of select risk
segments
• Improve ease of doing business with
agents
• Improve claims management activities
• Improve customer service at all levels
Operational Efficiency
• Reduce transaction costs
• Straight through processing of select risk
segments
• Improve ease of doing business with
agents
• Improve claims management activities
• Improve customer service at all levels
Marketing and Retention
• Target the right risks for non-renewals
• Improve retention of profitable risks
• Increase cross-sell opportunities
• Identify geographic and product
expansion opportunities
• Enhance recruiting of profitable
producers
Marketing and Retention
• Target the right risks for non-renewals
• Improve retention of profitable risks
• Increase cross-sell opportunities
• Identify geographic and product
expansion opportunities
• Enhance recruiting of profitable
producers
Enhanced Decision Making
• Increase fraud detection capabilities
• Improve monitoring of underwriting
performance
• Enhance ability to react to market forces
sooner
• Increase information processing
capabilities and data governance
Enhanced Decision Making
• Increase fraud detection capabilities
• Improve monitoring of underwriting
performance
• Enhance ability to react to market forces
sooner
• Increase information processing
capabilities and data governance
13
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
A practical approach to align business strategy with predictive modeling results
Segmentation of policies along the lift curve by decile and by decile groupings (green, yellow, red), can
be used to develop strategies and tactics for pricing and business rules that address different parts of the
lift curve. The use of deciles to affect underwriting results and meet operational goals (e.g., increasing
straight-through processing) is termed “decile management”. There are 5 impact categories where
deciles drive rules.
Management approval is required
for all override transactions on
deciles >7
Override
Approvals
Ask underwriting rule group #3
only for risks in decile 8 or greater
Underwriting
Exposure Rules
Only a 10% pricing range will be
provided for deciles 4-6
Underwriter
Pricing Flexibility
All risks in deciles 7 – 10 are
automatically referred for
underwriter review
Underwriter
Referral Rules
For deciles 1-3 initial pricing will be
at a 20% discount of manual price
Pricing
Hypothetical Business
Rules
Decile
Management
Impact
Categories
14
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Turning insight into action – rules based decision making
• In order to achieve profitability in this zone, the recommended price may be be beyond what the market
can bear or outside filed pricing deviations. In these cases, the pricing correction may take place over
several years, due to filing limitations.
Overview:
The predictive model, through the pricing matrix, will provide a suggested ‘right-price’ for specific risks. There are many
important implementation decisions that will drive how this price recommendation is utilized.
Green
Yellow
Red
• Since these risks are profitable and desirable, your goal is to write and retain as many as possible.
• These risks give you the greatest flexibility in achieving competitive market pricing; you should only give
enough pricing credit to write/retain the account.
• Market conditions in relation to model recommended pricing should be monitored to avoid policy attrition
of most desired accounts.
• The use of subjective credits and company placement options should be limited on these risks as they
are already generally priced adequately. Minor adjustments may be necessary but primarily, these
risks can be automatically processed subject to underwriting satisfaction.
Sample Business Rule:
• Low scoring policies can be placed in any rating tier. No referral necessary if price is less than expiring and quoted at
within 10% above matrix recommendation.
• Moderate scoring policies can be placed in all but the most competitive tier. No referral necessary if price is more than
10% off expiring but not greater than 20%.
Sample Business Rule:
• Low scoring policies can be placed in any rating tier. No referral necessary if price is less than expiring and quoted at
within 10% above matrix recommendation.
• Moderate scoring policies can be placed in all but the most competitive tier. No referral necessary if price is more than
10% off expiring but not greater than 20%.
Sample Underwriting Business Rules
15
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Examples of new performance metrics to make sure your strategies are working
New Business Underwriting
• Recommended Total Premium
• Deviation from Recommended vs. Premium Written
• Distribution of Recommended Scheduled Modifications
• List of most frequent positive contributing reason messages
• Comparison of premium year over year
Renewal Underwriting
• Recommended Total Premium
• Deviation from Recommended vs. Premium Written
• Distribution of Recommended Scheduled Modifications
• Distribution of recommended Non-Renewals
• Deviation from Recommended vs. Applied Non-Renewals
• List of most frequent positive contributing reason messages
• Comparison on schedule modification from prior year
Marketing and Distribution
• Agent Distribution by Decile
• Individual agent book distribution across deciles
• Hit ratio of leads generated by model
Claims • Deviation between predicted and actual claim frequency/ severity
16
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
• Dynamic reports will drill down capabilities in
a “real time” information processing
environment
• Reporting needs are aligned with enterprise
wide strategy
• Top down approach to data quality and data
governance with support from senior
executives
• Analysis conducted on new performance
metrics designed to maximize predictive
modeling ROI
• Flexibility at the end-user level to customize
reports and develop scenarios and perform
“what-if” analysis
• An decision support framework designed to
enhance information processing and analysis
capabilities across multiple business units
• IT resource intensive monthly production
reports
• Reporting needs are localized at the
business unit or functional level
• Limited corporate wide initiative towards
data quality and data governance
• Based on traditional underwriting and
actuarial metrics
• Limited flexibility for sensitivity analysis and
simulation of impact of business action
• Not designed to measure business process
improvements and customer satisfaction
• Information processing and decision support
capabilities are not at the end user or
business unit level
Enhanced Decision Support Framework Current State Decision Support Systems
IT and Decision Support Systems are key components that enable ongoing measurement
17
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Predictive Modeling changes the way companies do business…
Ongoing measurement allows them to make sure their new strategies are working
• Develop performance metrics to measure model performance in a predictive model
enabled business environment
• Implement reporting tools to measure the impact of predictive modeling on key
performance metrics
• Re-adjust business strategy and operations based on business intelligence derived
from performance metrics
Performance
Metrics and
Monitoring
• Examine marketing model results and identify geographic territories for possible
market expansion and profitable growth opportunities
• Identify profitable agents based on modeling results and target for relationship
management
• Identify profitable territories for new agency appointments
Improved
Distribution
Channel
Management
• Identify non-predictive application data currently being captured by Utica Mutual in an
effort to streamline and reduce unnecessary information currently being captured
during the underwriting process
• Develop initial business rules to leverage model results during the underwriting
process
Ease of Doing
Business
Contributing to ongoing measurement of predictive modeling impact
Contributing to a improved customer service and marketing activities
Contributing to agency focused operations and reduced expenses
Business Intelligence and
Reporting in Action
19
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
• The ability to measure, monitor and analyze the performance of your predictive modeling
initiative and related business decisions will allow carriers to distinguish themselves from
competitors in today’s competitive market
• New metrics and monitoring tools are required to measure the effectiveness of your
predictive modeling initiative and how the market is reacting to your business strategy
• New metrics and monitoring tools are critical in measuring the impact of your business
decisions and to help determine if your business process is aligned with your predictive
modeling initiative
• An enterprise wide data collection and management strategy will maximize the long term
performance of your predictive modeling initiative and improve your reporting and
information processing capabilities
• Choosing a reporting model that meets your short-term and long-term objectives will ensure
that key metrics and information is available at all levels of your organization
A framework for P&C companies to stay ahead of the competition in today’s
predictive modeling enabled business environment
20
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Applying Performance Metrics at critical steps in the predictive modeling process
Model Build Process
Enabling IT infrastructure
and applications
• Model Performance across Business
Segments
• Disruption Analysis from New to Renew
• Producer Management Analysis
• Pricing Matrix Impact Analysis
• Leads and New Business Production
Analysis
• Market Sensitivity Analysis
Business Implementation
• Analysis of Risk Characteristics Driving
Model Results
• Underwriting Rules Sensitivity Analysis
21
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Pre- Final Model Development – Measuring robustness of the predictive model
22
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Pre-Business Rules Development – Measuring underlying risk characteristics
driving model results
23
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Tying it all together – Pricing Impact Dashboard
Steps to Realize Business
Intelligence Opportunities
25
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Steps to Realize BI Opportunities
• Design reporting and data warehousing framework
• Develop performance metrics
• Implement reporting framework
• Analyze results
• Realign strategy and metrics based on performance metric results
Primary
Activities
• Data warehousing and ETL tools
• Business Objects, Crystal Reports, Cognos among others
• Java, .Net or other application development tools
Tools and
Methodologies
• A framework that aligns technology, analytical and business services to maximize return
on your predictive modeling initiative
• Designed to achieve seamless integration across technology platforms, business
applications and functional business units
• Increases an organization’s access to timely information and dissipates information
across all levels of an organization
• Improves data quality and data collection efforts to further enhance future predictive
modeling recalibration efforts
• Provides a mechanism to communicate results to key internal and external stakeholders.
Benefits
26
Copyright © 2007 Deloitte Development LLC. All rights reserved.
– Prepare the organization for the next stage of
the underwriting journey where models will be
used as the central risk selection and
underwriting tool
– Develop performance metrics and framework
to continuously monitor the impact of the
predictive modeling initiative
Model Value Capture Lessons P&C Company Implications
Lessons learned through the model build and conceptualization work should be considered
as P&C companies develop their implementation plans.
– Companies often focus on the individual
variables rather than the aggregate power
– Relationship between variables drives the
major components of value creation
Understand the
Predictive Power
– Ready the organization to rigorously apply,
rather than deconstruct, the model
– Importance of communication and education
on the usage of the model output should not
be understated
– The model’s lift often exceeds the company’s
capacity for rating action based on filed rating
plans
– Companies often waste months looking to
“fine tune” current variables and uncover new
variables
Pursuit of “Great”
Models Often
Obstructs Immediate
Value Capture
– Validate and implement the current model as
soon as possible to enable value capture
– Prepare/research new variables for future
model updates - Data Quality and
Governance
– Modeling additional market segments will
drive significant value
– Modeling sometimes remains an actuarial
exercise, rather than an enterprise
implementation priority
– Companies fail to recognize the sweeping
nature of the changes required to realize full
benefits and that predictive modeling is a
continuous process
It’s All About
Implementation
– Companies often fail to move beyond the
interim workarounds that apply to the best or
worst scoring policies, rather than addressing
the book holistically
– Short term “quick fix” changes can encumber
rather than empower underwriters – which
can serve to obstruct buy-in
Avoid the
“Interim Trap”
– Don’t let immediate value capture obstruct the
full implementation
– To achieve full growth and efficiency gains full
implementation is required
Steps to Realize BI Opportunities
27
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Speaker Contact Information
Mo Masud, Senior Manager
Deloitte Consulting LLP
[email protected]
860-725-3341
Lisa Wester, Manager
Deloitte Consulting LLP
[email protected]
312-486-1994
28
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Member of
Deloitte Touche Tohmatsu
© Deloitte & Touche LLP and affiliated entities.
Deloitte, one of Canada's leading professional services firms, provides audit, tax, consulting, and
financial advisory services through more than 6,200 people in 50 offices. Deloitte operates in Québec
as Samson Bélair/Deloitte & Touche s.e.n.c.r.l. The firm is dedicated to helping its clients and its
people excel. Deloitte is the Canadian member firm of Deloitte Touche Tohmatsu.
Deloitte refers to one or more of Deloitte Touche Tohmatsu, a Swiss Verein, its member firms, and
their respective subsidiaries and affiliates. As a Swiss Verein (association), neither Deloitte Touche
Tohmatsu nor any of its member firms has any liability for each other's acts or omissions. Each of the
member firms is a separate and independent legal entity operating under the names "Deloitte,"
"Deloitte & Touche," "Deloitte Touche Tohmatsu," or other related names. Services are provided by
the member firms or their subsidiaries or affiliates and not by the Deloitte Touche Tohmatsu Verein.
doc_944995890.pdf
Tools for Performance Metrics and Business Intelligence
Tools for Performance Metrics
and Business Intelligence
October 12, 2007
Mo Masud
Lisa Wester
2
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Agenda
Business Intelligence and Reporting In Action
Steps to Realize Business Intelligence Opportunities
Measuring Model Impact through Business Intelligence
Rationalizing Performance Metrics and Business Intelligence
3
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Key Discussion Themes
Practical Ways to Measure Predictive Modeling Impact
Designing New Ways to Measure Strategic Choices in a Predictive
Modeling Enabled World
The Role of IT and Decision Support Systems in Predictive Modeling
Aligning Strategy with Predictive Modeling Results to Maximize Benefits
Steps to Get Started
Predictive Modeling is a Continuous Process
Going Beyond Model Development to Gain a Competitive Edge
4
Copyright © 2007 Deloitte Development LLC. All rights reserved.
The Time is Now for Going Beyond Just Model Development
Looking into the future
Analytics is allowing organizations to stay ahead of the competition and ahead of the market
The predictive modeling story resonates well with all parties – consistent, objective, reduced
expense, accurate
Analytics is becoming a core competency and organizational strategy – competing on analytics
Data driven
• Ability to store, process and distribute information faster than
competitors will differentiate organizations and give them a
competitive advantage
• Data quality and data governance is a top IT priority
• Developing fluid business strategies based on predictive
modeling results will set part the market leaders
Rationalizing Performance Metrics and BI
5
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Rationalizing Performance Metrics and BI
Attributes of companies that are analytically-oriented
One or more senior execs who are strongly advocating analytics and fact-based decision making
Widespread use of descriptive statistics, predictive modeling and complex optimization techniques
Substantial use of analytics across multiple business functions and processes
Movement toward an enterprise-level approach to managing analytical tools, data, and organizational
skills and capabilities
Stages of analytic competition
1. Facing major barriers (organizational, technical)
2. Analytic progress is made only locally, and not spread to other parts of the company
3. Value and promised realized, but unsuccessful implementation of analytic strategies
4. Vision has been realized, but enterprise-wide adoption has not been undertaken
5. Analytic competition has been built into company strategy
Source: Competing on Analytics, Thomas H. Davenport, Don Cohen, and Al Jacobson
6
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Rationalizing Performance Metrics and BI
Predictive Modeling in the Property & Casualty Insurance Industry is a Now a Core Business
Strategy
Several Carriers have already
implemented Predictive Modeling for
commercial business
– Primarily targeted at non-renewals and
right pricing
– Customized models
– Large financial benefits
Several carriers that have already
developed modeling capabilities are
rapidly advancing their capabilities
– Broader adoption in the industry
– Numerous production systems
– Expansion into account models
– Broader business applications
– Tighter linkage to business rules
– New players are entering the space
Companies must go beyond just model
development
– Standard practice in leading companies
– Will be seamlessly integrated with business
rules to drive results
– Early adopters will have leveraged capability to
gain market share
– Extension to non-standard lines and more
creative applications
– Generic LOB tools will emerge
– Integration of models across functions (Claim,
U/W, Marketing)
– Ability to process results faster than
competitors will distinguish market leaders
Time
3+ Years Last 3-5 years 1 – 3 Years
Innovators
Early Adopters
Early Majority Late Majority
Laggards
Yesterday Today Future
7
Copyright © 2007 Deloitte Development LLC. All rights reserved.
“85% of our new products
are automated with
predictive modeling, which
enhances underwriting
consistency and makes it
easy for our commercial
agents to do business with
Safeco”
“Fully launched Customized
Pricing, our predictive pricing
model that automatically
provides the most appropriate
price for a new small business
submission”
“Achieved overall written
premium growth of 4%.”
AIG’s Mr. Purdy who
describes predictive modeling
as “evolutionary” said, “I
expect that every insurer will
be using WC data they
collect to do predictive
modeling at least by 2010.”
“Modeling is now moving to a
higher level of sophistication”
“Commercial lines NPW grew
3% for Q2, driven by $83M in
new commercial business,
up 13%compared to Q2
2006.”
“The dramatic improvement is
a direct result of our
multidisciplinary WC
improvement strategy and
predictive modeling”
“Because our comprehensive
use of predictive modeling
is now analyzing 80% of our
commercial lines business,
both new and renewal
accounts, we are able to
better match price to risk. As
result we feel we can continue
to improve our profit margin
[…]”
“As Maryland continued to
experience an increasingly
soft market, we retained 86%
of expiring premium”
“Predictive modeling /
automated underwriting were
implemented to enhance
service delivery.”
“The quality of our property,
WC, auto, and specialty mix
is continuing to improve as
we use predictive models”
“Agents are giving us
preferred shelf space vs.
weaker competitors”
“Increased ease of use
through faster decisions,
streamlined processing, and
expanded account rounding”
“Reduced system quote to
issue time through dynamic
questions, improved agent
interface, and automated
UW.”
Rationalizing Performance Metrics and BI
Several companies that have implemented Predictive Models are raving about the impact
that it has had on their business.
8
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Banking - With SPSS, HSBC
Bank USA effectively mines
an ever-growing file of
customer data, creating
predictive models to uncover
cross-selling and "roll over"
sales opportunities. Focusing
on the best prospects for each
product helps maximize sales
and minimize marketing costs,
and SPSS' ease of use helps
researchers deliver
intelligence faster to decision
makers.
Retail - Using the Wal-Mart Model, Paul Westerman points
out Wal-Mart's use of store and product traits to guide
replenishment. Simplified, 'beach' products are assigned to
'beach' stores. But how many snorkels should go to a
particular store? Paul advises that "a data warehouse can
provide a good estimate based on another, similar product
that has the same distribution."
Market basket analysis reveals these complementary
products. Combined with demographic data, you
understand the market forces at work (e.g., a correlation
with young affluent families) and send masks and flippers to
stores in neighborhoods with swimming pools, too.
Through careful planning, your past and future converge to
tackle today's problems.
Healthcare - The new way is
predictive modeling, and it is being
implemented at health plans left and
right.
"The goal of a clinical strategy is to
find the right intervention for the right
person at the right time," says Carol
McCall, vice president of Humana's
Center for Health Metrics, "and
ultimately, you want to be able to
understand enough about people to
be able to literally custom-tailor the
exact intervention so that you can
meet them in a way that they will be
the most receptive to it." McCall is a
fellow of the Society of Actuaries and
a member of the Academy of
Actuaries.
But if everyone in the industry is doing it, how can companies differentiate themselves
The answer may lie at looking at leading-edge companies in other industries
Rationalizing Performance Metrics and BI
9
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Rationalizing Performance Metrics and BI
• The strongest statistical models will not offer clients any benefits to their bottom line unless it is
seamlessly integrated within their technical and business infrastructure
• In conjunction with a integrated predictive modeling solution, clients must realign their business
strategy to capitalize on their predictive modeling solution
• Operating in a predictive modeling enabled organization is a new way of doing business and
requires new performance metrics to measure the success of the predictive modeling investment
• Realizing predictive modeling benefits is a continuous process that is evolving to meet the demands
of the external market
• Business intelligence tools are the means to continuously monitor and measure the impact of
predictive modeling enabled strategies
Enhanced Decision Support Capabilities Gives Organizations a Competitive Edge
Measuring Model Impact
through Business Intelligence
11
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Achieving Continuous Predictive Modeling Improvement
Continuous
Predictive
Modeling
Improvement
Predictive Model Development
The foundation of workers’ compensation analytics is
the design and development of predictive models
Business Process Redesign -
Redesigning underwriting
workflows, business rules and
business processes to enable
predictive modeling results based
decision-making
Scoring Engine Development
The technical manifestation of the
predictive models transforming raw
data into predictive model scores and
underlying reason codes
Technical Integration
The integration of the
predictive modeling results
produced by the scoring
engine within front-end
underwriting applications
Performance Monitoring
Decision Support Framework
A framework to enhance
enterprise wide
information processing,
data capture, and
performance
Defining reporting requirements and aligning
requirements across business units, and
performance metrics to measure the
effectiveness of the predictive modeling
enabled business decisions
12
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Aligning strategy with predictive modeling results is the first step to maximize benefits
D
e
p
l
o
y
A
n
a
l
y
z
e
D
e
v
e
l
o
p
S
e
l
e
c
t
IT
U
n
d
e
r
w
r
i
t
i
n
g
M
a
r
k
e
t
i
n
g
Operations
Predictive Modeling Enabling Business Strategies for Success
Underwriting Excellence
• Improve pricing precision
• Increase objectivity throughout the
underwriting process
• Enhance risk selection and risk
avoidance capabilities
• Improve pricing competitiveness in
profitable segments
• Improve underwriter negotiation
capabilities
Underwriting Excellence
• Improve pricing precision
• Increase objectivity throughout the
underwriting process
• Enhance risk selection and risk
avoidance capabilities
• Improve pricing competitiveness in
profitable segments
• Improve underwriter negotiation
capabilities
Operational Efficiency
• Reduce transaction costs
• Straight through processing of select risk
segments
• Improve ease of doing business with
agents
• Improve claims management activities
• Improve customer service at all levels
Operational Efficiency
• Reduce transaction costs
• Straight through processing of select risk
segments
• Improve ease of doing business with
agents
• Improve claims management activities
• Improve customer service at all levels
Marketing and Retention
• Target the right risks for non-renewals
• Improve retention of profitable risks
• Increase cross-sell opportunities
• Identify geographic and product
expansion opportunities
• Enhance recruiting of profitable
producers
Marketing and Retention
• Target the right risks for non-renewals
• Improve retention of profitable risks
• Increase cross-sell opportunities
• Identify geographic and product
expansion opportunities
• Enhance recruiting of profitable
producers
Enhanced Decision Making
• Increase fraud detection capabilities
• Improve monitoring of underwriting
performance
• Enhance ability to react to market forces
sooner
• Increase information processing
capabilities and data governance
Enhanced Decision Making
• Increase fraud detection capabilities
• Improve monitoring of underwriting
performance
• Enhance ability to react to market forces
sooner
• Increase information processing
capabilities and data governance
13
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
A practical approach to align business strategy with predictive modeling results
Segmentation of policies along the lift curve by decile and by decile groupings (green, yellow, red), can
be used to develop strategies and tactics for pricing and business rules that address different parts of the
lift curve. The use of deciles to affect underwriting results and meet operational goals (e.g., increasing
straight-through processing) is termed “decile management”. There are 5 impact categories where
deciles drive rules.
Management approval is required
for all override transactions on
deciles >7
Override
Approvals
Ask underwriting rule group #3
only for risks in decile 8 or greater
Underwriting
Exposure Rules
Only a 10% pricing range will be
provided for deciles 4-6
Underwriter
Pricing Flexibility
All risks in deciles 7 – 10 are
automatically referred for
underwriter review
Underwriter
Referral Rules
For deciles 1-3 initial pricing will be
at a 20% discount of manual price
Pricing
Hypothetical Business
Rules
Decile
Management
Impact
Categories
14
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Turning insight into action – rules based decision making
• In order to achieve profitability in this zone, the recommended price may be be beyond what the market
can bear or outside filed pricing deviations. In these cases, the pricing correction may take place over
several years, due to filing limitations.
Overview:
The predictive model, through the pricing matrix, will provide a suggested ‘right-price’ for specific risks. There are many
important implementation decisions that will drive how this price recommendation is utilized.
Green
Yellow
Red
• Since these risks are profitable and desirable, your goal is to write and retain as many as possible.
• These risks give you the greatest flexibility in achieving competitive market pricing; you should only give
enough pricing credit to write/retain the account.
• Market conditions in relation to model recommended pricing should be monitored to avoid policy attrition
of most desired accounts.
• The use of subjective credits and company placement options should be limited on these risks as they
are already generally priced adequately. Minor adjustments may be necessary but primarily, these
risks can be automatically processed subject to underwriting satisfaction.
Sample Business Rule:
• Low scoring policies can be placed in any rating tier. No referral necessary if price is less than expiring and quoted at
within 10% above matrix recommendation.
• Moderate scoring policies can be placed in all but the most competitive tier. No referral necessary if price is more than
10% off expiring but not greater than 20%.
Sample Business Rule:
• Low scoring policies can be placed in any rating tier. No referral necessary if price is less than expiring and quoted at
within 10% above matrix recommendation.
• Moderate scoring policies can be placed in all but the most competitive tier. No referral necessary if price is more than
10% off expiring but not greater than 20%.
Sample Underwriting Business Rules
15
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Examples of new performance metrics to make sure your strategies are working
New Business Underwriting
• Recommended Total Premium
• Deviation from Recommended vs. Premium Written
• Distribution of Recommended Scheduled Modifications
• List of most frequent positive contributing reason messages
• Comparison of premium year over year
Renewal Underwriting
• Recommended Total Premium
• Deviation from Recommended vs. Premium Written
• Distribution of Recommended Scheduled Modifications
• Distribution of recommended Non-Renewals
• Deviation from Recommended vs. Applied Non-Renewals
• List of most frequent positive contributing reason messages
• Comparison on schedule modification from prior year
Marketing and Distribution
• Agent Distribution by Decile
• Individual agent book distribution across deciles
• Hit ratio of leads generated by model
Claims • Deviation between predicted and actual claim frequency/ severity
16
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
• Dynamic reports will drill down capabilities in
a “real time” information processing
environment
• Reporting needs are aligned with enterprise
wide strategy
• Top down approach to data quality and data
governance with support from senior
executives
• Analysis conducted on new performance
metrics designed to maximize predictive
modeling ROI
• Flexibility at the end-user level to customize
reports and develop scenarios and perform
“what-if” analysis
• An decision support framework designed to
enhance information processing and analysis
capabilities across multiple business units
• IT resource intensive monthly production
reports
• Reporting needs are localized at the
business unit or functional level
• Limited corporate wide initiative towards
data quality and data governance
• Based on traditional underwriting and
actuarial metrics
• Limited flexibility for sensitivity analysis and
simulation of impact of business action
• Not designed to measure business process
improvements and customer satisfaction
• Information processing and decision support
capabilities are not at the end user or
business unit level
Enhanced Decision Support Framework Current State Decision Support Systems
IT and Decision Support Systems are key components that enable ongoing measurement
17
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Measuring Model Impact through Business
Intelligence
Predictive Modeling changes the way companies do business…
Ongoing measurement allows them to make sure their new strategies are working
• Develop performance metrics to measure model performance in a predictive model
enabled business environment
• Implement reporting tools to measure the impact of predictive modeling on key
performance metrics
• Re-adjust business strategy and operations based on business intelligence derived
from performance metrics
Performance
Metrics and
Monitoring
• Examine marketing model results and identify geographic territories for possible
market expansion and profitable growth opportunities
• Identify profitable agents based on modeling results and target for relationship
management
• Identify profitable territories for new agency appointments
Improved
Distribution
Channel
Management
• Identify non-predictive application data currently being captured by Utica Mutual in an
effort to streamline and reduce unnecessary information currently being captured
during the underwriting process
• Develop initial business rules to leverage model results during the underwriting
process
Ease of Doing
Business
Contributing to ongoing measurement of predictive modeling impact
Contributing to a improved customer service and marketing activities
Contributing to agency focused operations and reduced expenses
Business Intelligence and
Reporting in Action
19
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
• The ability to measure, monitor and analyze the performance of your predictive modeling
initiative and related business decisions will allow carriers to distinguish themselves from
competitors in today’s competitive market
• New metrics and monitoring tools are required to measure the effectiveness of your
predictive modeling initiative and how the market is reacting to your business strategy
• New metrics and monitoring tools are critical in measuring the impact of your business
decisions and to help determine if your business process is aligned with your predictive
modeling initiative
• An enterprise wide data collection and management strategy will maximize the long term
performance of your predictive modeling initiative and improve your reporting and
information processing capabilities
• Choosing a reporting model that meets your short-term and long-term objectives will ensure
that key metrics and information is available at all levels of your organization
A framework for P&C companies to stay ahead of the competition in today’s
predictive modeling enabled business environment
20
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Applying Performance Metrics at critical steps in the predictive modeling process
Model Build Process
Enabling IT infrastructure
and applications
• Model Performance across Business
Segments
• Disruption Analysis from New to Renew
• Producer Management Analysis
• Pricing Matrix Impact Analysis
• Leads and New Business Production
Analysis
• Market Sensitivity Analysis
Business Implementation
• Analysis of Risk Characteristics Driving
Model Results
• Underwriting Rules Sensitivity Analysis
21
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Pre- Final Model Development – Measuring robustness of the predictive model
22
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Pre-Business Rules Development – Measuring underlying risk characteristics
driving model results
23
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Business Intelligence and Reporting in Action
Tying it all together – Pricing Impact Dashboard
Steps to Realize Business
Intelligence Opportunities
25
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Steps to Realize BI Opportunities
• Design reporting and data warehousing framework
• Develop performance metrics
• Implement reporting framework
• Analyze results
• Realign strategy and metrics based on performance metric results
Primary
Activities
• Data warehousing and ETL tools
• Business Objects, Crystal Reports, Cognos among others
• Java, .Net or other application development tools
Tools and
Methodologies
• A framework that aligns technology, analytical and business services to maximize return
on your predictive modeling initiative
• Designed to achieve seamless integration across technology platforms, business
applications and functional business units
• Increases an organization’s access to timely information and dissipates information
across all levels of an organization
• Improves data quality and data collection efforts to further enhance future predictive
modeling recalibration efforts
• Provides a mechanism to communicate results to key internal and external stakeholders.
Benefits
26
Copyright © 2007 Deloitte Development LLC. All rights reserved.
– Prepare the organization for the next stage of
the underwriting journey where models will be
used as the central risk selection and
underwriting tool
– Develop performance metrics and framework
to continuously monitor the impact of the
predictive modeling initiative
Model Value Capture Lessons P&C Company Implications
Lessons learned through the model build and conceptualization work should be considered
as P&C companies develop their implementation plans.
– Companies often focus on the individual
variables rather than the aggregate power
– Relationship between variables drives the
major components of value creation
Understand the
Predictive Power
– Ready the organization to rigorously apply,
rather than deconstruct, the model
– Importance of communication and education
on the usage of the model output should not
be understated
– The model’s lift often exceeds the company’s
capacity for rating action based on filed rating
plans
– Companies often waste months looking to
“fine tune” current variables and uncover new
variables
Pursuit of “Great”
Models Often
Obstructs Immediate
Value Capture
– Validate and implement the current model as
soon as possible to enable value capture
– Prepare/research new variables for future
model updates - Data Quality and
Governance
– Modeling additional market segments will
drive significant value
– Modeling sometimes remains an actuarial
exercise, rather than an enterprise
implementation priority
– Companies fail to recognize the sweeping
nature of the changes required to realize full
benefits and that predictive modeling is a
continuous process
It’s All About
Implementation
– Companies often fail to move beyond the
interim workarounds that apply to the best or
worst scoring policies, rather than addressing
the book holistically
– Short term “quick fix” changes can encumber
rather than empower underwriters – which
can serve to obstruct buy-in
Avoid the
“Interim Trap”
– Don’t let immediate value capture obstruct the
full implementation
– To achieve full growth and efficiency gains full
implementation is required
Steps to Realize BI Opportunities
27
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Speaker Contact Information
Mo Masud, Senior Manager
Deloitte Consulting LLP
[email protected]
860-725-3341
Lisa Wester, Manager
Deloitte Consulting LLP
[email protected]
312-486-1994
28
Copyright © 2007 Deloitte Development LLC. All rights reserved.
Member of
Deloitte Touche Tohmatsu
© Deloitte & Touche LLP and affiliated entities.
Deloitte, one of Canada's leading professional services firms, provides audit, tax, consulting, and
financial advisory services through more than 6,200 people in 50 offices. Deloitte operates in Québec
as Samson Bélair/Deloitte & Touche s.e.n.c.r.l. The firm is dedicated to helping its clients and its
people excel. Deloitte is the Canadian member firm of Deloitte Touche Tohmatsu.
Deloitte refers to one or more of Deloitte Touche Tohmatsu, a Swiss Verein, its member firms, and
their respective subsidiaries and affiliates. As a Swiss Verein (association), neither Deloitte Touche
Tohmatsu nor any of its member firms has any liability for each other's acts or omissions. Each of the
member firms is a separate and independent legal entity operating under the names "Deloitte,"
"Deloitte & Touche," "Deloitte Touche Tohmatsu," or other related names. Services are provided by
the member firms or their subsidiaries or affiliates and not by the Deloitte Touche Tohmatsu Verein.
doc_944995890.pdf