Marketing Applications of Predictive Analytics Robert J. Walling III, FCAS, MAAA San Diego, CA October 6, 2008
Overview Who s Buying What? Who s Selling What? A Proactive Approach Monitoring Results
Who s Buying What?
Questions Where Are Our Products Competitive? What Should We Encourage Our Agents to Sell? Where Do We Have a Competitive Advantage?
Historical View - Competitive Position Competitive Quotes Market Position Based on rate differences If my rates are lower than the competitor, I must be more competitive If my rates are higher than the competitor, I must be less competitive Assumes insurance decision is solely based on price
Those Were The Days How We Got Competitive Information Insurance Department Manual Exchange Industry Meetings Directly from Competitors Agents
What We Did With the Competitive Information Rate Examples Explored New Factors Adjusted Rate Actions Accordingly
What We Did With the Competitive Information Rate Examples not sure if rates being calculated are right Explored New Factors if you know what they are Adjusted Rate Actions Accordingly now it is like taking a shot in the dark
That Was Then, This Is Now Increased Use of Tiering/Scoring Increased Use of New or Proprietary Factors Insurance Score Prior Limits Historical Non-chargeable Losses Increased Difficulty Getting Competitor Information Filings on Copy Resistant Paper Classifying Certain Information as Underwriting Agents Not Getting Manuals Increased Complexity of Rating Plans
Competitive Positioning Competitive Positioning Determining How Your Company Compares to Their Competitors Considerations Risk Characteristics Geography Price Brand Distribution channel
Changing View of Competitive Positioning Segmented Market Share Life Cycle Analysis Competitive Quotes Based on market realities Still retains analysis of price Market Position Agency Management Analysis Also includes market reality of actual results Allows insurer to then investigate cause of competitive issues
Business Life Cycle Marketing Effort Quote Sale Retention
Retention & Conversion Analysis Quoting Analysis: Likelihood of a Prospective Insured Obtaining a Quote Conversion Analysis: Likelihood of a Insured that Received a Quote Purchasing from You Retention Analysis: Likelihood of a Current Insured Renewing with You Cross Selling Analysis - Likelihood of a Current Insured Purchasing Additional Coverages/Policies Use of These Likelihoods as a Proxy for Competition Does not tell why you are losing/not writing risks
Quoting Analysis Measure characteristics of shoppers, quoters, purchases, and retained business Characteristics Internal company information External demographic information Credit profiles Shopping incentives Identify insureds to target
Conversion Analysis Example Example Relativit 2.000 1.800 1.600 1.400 1.200 1.000 0.800 0.600 0.400 0.200 0.000 1.766 1.408 1.431 1.483 1.000 1.000 <missing> 0 1 2 4 6 EDUCATION Relative Likelihood
Conversion Analysis Example Example Company 1.200 1.000 1.054 1.000 Relativity 0.800 0.600 0.400 0.683 0.335 0.200 0.000 1 2 3 4+ Number of Drivers
Retention Analysis X Claim History 0.57 18 0 0.56 16 0 0.55 14 0 12 0 0.54 10 0 0.53 80 0.52 60 0.51 40 0.5 20 0.49 0 1 2+ 0
Segmented Market Share Overall Market Share is the Focus Many Risk Segments can be Segmented by ZIP/County Personal Vehicle Registrations Homes Businesses Commercial Vehicles Some Registration Information Can be Tied to Demographic Information Age Marital Status Homeowners Occupation Etc.
Detailed Market Share Segments 1. Detailed Penetration by: Geography (ZIP, Census Block) Risk Characteristics (Make/Model, Industry, etc.) Insured/Driver Characteristics 2. Penetration Levels Based on Characteristics 3. Penetration Levels by Company Risk Characteristics 4. Models of Projected Market Penetration Based on Historical Data
Penetration by Model Year ABC Insurance Company 4.00% 3.50% 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 1973 & OLDER 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 0.00% Vehicle Year 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 0.0% Market Penetration Incurred Loss Ratio Market Penetration Incurred Loss Ratio
Segmented Market Share 40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% 18-24 25-34 35-44 45-54 55-64 65-74 75+ Customers Non-Customers
Demographic Characteristics X Penetration by Percent Married 4.00% 100.0% 3.50% 3.00% 90.0% 80.0% 70.0% Market Penetration 2.50% 2.00% 1.50% 60.0% 50.0% 40.0% Incurred Loss Ratio 30.0% 1.00% 20.0% 0.50% 10.0% 0.00% Less Than 44.9 45.0-49.1 49.2-51.1 51.2-52.3 52.4-53.5 53.6-54.0 54.1-55.8 55.9-56.5 56.6-57.2 57.3-57.9 58.0-58.4 Percent Married 58.5-58.9 59.0-60.1 60.2-60.8 60.9-61.4 61.5-62.4 62.5-63.0 63.1-64.0-63.9 65.4 65.5 + 0.0% Penetration Incurred Loss Ratio
Who s Selling What?
Questions Which Agents Place Our Products in a Preferred Marketing Position? Higher Conversion Ratio Superior Underwriting Results Which Agents Use Us To Keep Their Preferred Markets Honest? High Quote Counts, Low Conversion High Underwriting Expenses Adverse Selection Which Agents are Efficient Sources of New Business?
Issues It is Important to Separate the Role of the Agent from the Role of the Rating Plan How Does an Agent s Book Compare to a Comparable Book Across All Agents? What Incremental Value Does the Agent Add? Higher Hit Ratio Higher Retention Ratio Lower Claim Frequency Lower Loss Ratio Questions Ideally Suited for Predictive Modeling
Profiles by Agency Focus 0.0370 Pre dicte d V alue s 90% 0.0365 80% 0.0360 70% 0.0355 0.0350 0.0345 60% 50% 40% 30% 0.0340 20% 0.0335 10% 0.0330 Personal Balanced Comm/Specialty Missing Age nt Focus 0%
Agency Management Example WC Agency Analysis Example 2.5 600,000,000 2 203% 500,000,000 1.5 400,000,000 1 100% 94% 0.5 0-0.5 37% 9% 12% 9% 3% -11% -12% -52% -41% 59% 16% 19% 2% -7% -28% -46% -47% 32% 19% -2% -8% -2% -11% -33% -7% 6% 41% -6% 300,000,000 200,000,000-1 -99% -71% -78% 100,000,000-1.5 70 200 202 204 207 209 213 214 216 220 225 237 241 242 243 247 249 251 255 259 260 261 265 266 271 275 276 278 280 281 284 286 287 296 Agency Number (Magency_number) 0 Exposures Relative Contribution
Factors That Matter Rank in Agency Volume in Agency Number of Carriers Represented Geography (Urban, Suburban, Rural) Licensing Status Years of Appointment Agency Life Cycle
A Proactive Approach
Where Do I Want to Improve? Profitability High Profit, Low CP High Profit, High CP Competitive Position Low Profit, Low CP Low Profit, High CP
How to Improve Competitive Position Marketing to Quote Target Marketing Lists Demographic Analysis Brand Ease of Doing Business Quote to Sale Service Price Value Ease of Doing Business Sale to Retention Service Price Value Relationship Ease of Doing Business
Marketing to Quote Goal Increase Desirable Quote Traffic Targets Customer We Expect to Have Superior Performance Lifetime Value? Independent agent with demonstrated success Tactics Create awareness of company Identify shoppers Target marketing lists Demographic analysis Provide shoppers with incentive to contact Ease of Doing Business
Marketing to Quote Example 1 Analysis Reveals Lower than Average Penetration in State X Targeted solution: List of households in the market to purchase a new or used car in the next 6 months, for direct marketing: State X: 525,000 records 10 zip codes: 11,400 records Further Refine for Territories/Current Vehicle Types You Have Been Having Success
Marketing to Quote Example 2 Conversion Analysis Reveals a Competitive Advantage Above Average Hit Ratio Superior Underwriting Results Appears Related to Class Plan/Underwriting Score Garage Service Operations Non-Franchised In Business More than 3 Years Above Average Credit Score (Blended Commercial/Personal) Buy Marketing List of Businesses that Meet Criteria Provide to Agents with Proven Success with Program
Monitoring Results
How are we doing? Many of the Modeling Techniques also Measure Changes/Trends Hit Ratios/Conversion Ratio Analyses Segmented Market Penetration How Do We Measure Risks That Are In the Market Instead of the Total Market?
Analysis of Quote Activity Batch quoting is not an exact science May not reflect mix of business in marketplace Because of insurance scoring and other proprietary elements, absolute rate level is difficult to determine More likely to get relative performance between risks correct Use Performance of Different Insurers in Common Agency Management Platform to Determine Effective Competitor Relativities Difference Between Your Program and Other Companies by Risk Characteristics
Effective Competitor Relativities Example Company 1.20 Relativity 1.00 0.80 0.60 0.40 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.90 0.84 0.75 0.75 0.80 0.86 0.88 0.20 0.00 No Yes multiline
Credit Scores Conversion Example Relativity 2.00 1.80 1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00 Insurance Score
Parting Thoughts When asked to divulge the secret of his playing strategy, ice hockey great Wayne Gretzky used to say, "I skate where the puck is going to be, not where it's been."