Analysis of Internet Purchasing Behavior October 3, 2011 CAS In Focus Seminar Kevin Levitt Senior Vice President comscore, Inc. Roosevelt Mosley, FCAS, MAAA Principal Pinnacle Actuarial Resources Nick Kucera Consultant Pinnacle Actuarial Resources Experience the Pinnacle Difference! Discussion Topics Current State of Insurance Marketing Customer response analyses comscore background and data Description of research Characteristics of shoppers Model development Analysis of quotes submitted Analysis of policies purchased Impact of price on conversion Unstructured data Additional Research 2 1
Current State of Insurance Marketing Explosion in the investment that insurers are making in marketing Many different advertising mediums are being used Traditional, internet, social media, etc. Has led to the understanding that customer behavior is about more than price Customer service, reputation, convenience Different elements are important to different market segments 3 Customer Response Analyses Marketing Effort Sale Renewal Renewal Quoting Analysis Conversion Analysis Lapse / Cancelation Analysis Retention Analysis 4 2
Customer Response Analyses Quoting analysis: analysis of the likelihood of a prospective insured obtaining an insurance quote from you Conversion analysis: analysis of the likelihood of a prospective insured that has received a quote purchasing insurance from you Lapse/Cancelation analysis: likelihood of an insured not lapsing or canceling the policy mid-term Retention analysis: analysis of the likelihood lih of a current insured renewing with you 5 Customer Response Modeling Challenges for Insurance Companies Model structure and parameterization New territory learning curve Priority Internal data availability Internal data applicability Availability of price change data Measuring market competitiveness Applications 6 3
comscore Background & Data 7 comscore is a Global Leader in Measuring the Digital World NASDAQ SCOR Clients 1700+ worldwide Employees 900+ Headquarters Global Coverage Local Presence Reston, VA 170+ countries under measurement; 43 markets reported 32 locations in 23 countries V0910 8 4
What We Do. We provide digital marketing intelligence that helps our customers make better-informed business decisions and implement more effective digital business strategies We measure the continuous online activity of 1 million people in the U.S. who have granted us explicit permission to confidentially measure their Internet usage patterns. Our consumer panel is a representative cross-section of the U.S. population, worldwide regions and individual countries We also have permission to: Survey panelists Match to third-party databases Append offline data 9 The Trusted Source for Digital Intelligence Across Vertical Markets 9 out of the top 10 INVESTMENT BANKS 4 out of the top 4 WIRELESS CARRIERS 47 out of the top 50 ONLINE PROPERTIES 45 out of the top 50 ADVERTISING AGENCIES 9 out of the top 10 MAJOR MEDIA COMPANIES 9 out of the top 10 AUTO INSURERS 9 out of the top 10 INTERNET SERVICE PROVIDERS 9 out of the top 10 PHARMACEUTICAL COMPANIES 9 out of the top 10 CONSUMER FINANCE COMPANIES 9 out of the top 10 CPG COMPANIES V0910 10 5
Auto Insurance Detail Data is captured from what panelists see using scraping technology 11 Auto Insurance Detail Data for 5 Top Auto Insurance Company Sites ZIP code Bodily injury liability limits Coverage package Premium quoted Final Purchased Premium Company name Homeownership Whether SSN entered Primary driver education Current Insurance Information Whether currently insured Length of gap in coverage Length of current carrier Length continuously insured Prior BI Limit Drivers Age Gender Marital Status Industry/Occupation Vehicles Vehicle year/make/model/type Vehicle use Annual mileage Comprehensive deductibles Collision deductibles Incidents Incident Type Incident Description 12 6
Auto Insurance Detail Data Source of Traffic Data Source of traffic is broken out into the following categories: Paid search Natural search Webmail sites Other referred Non-referred For search, we know the search engine, click type (paid/natural), and the search phrase For webmail sites and other referred, we know the referring site 13 Pinnacle & comscore Research Marketing Effort Population Target Description Auto Insurance Website Visitor Initiated Of those that visit the website, people with what characteristics are more likely to start the quote process? Sale Initiated Submitted Of those that initiate the quote process, what is the likelihood that they complete it? Of those that actually submit the quote, how many complete the purchase? Policy Bind Of those that actually submit the quote Submitted 14 7
Characteristics of Shoppers 15 Data Summary Means of Entry N u m b e r o f 40.0% V i 30.0% s i t 20.0% o r s 10.0% Model - Submitted Means of Entry 80.0% 600 570 559 546 70.0% F 500 r e 60.0% 463 q u 411 400 e 50.0% 46.8% n c y 9.7% 32.6% 5.4% 5.5% 300 200 100 o f S u b m i t t e d p e r Q u 1 o, t 0 e 0 s 0 0.0% Natural Search Non-Referred Other Referred Sponsored Search Webmail Means of Entry Exposure Percentage Frequency per 1000 0 16 8
Data Summary Session Day/Time Model: Purchase Made During Session 30.0% 120.00 ure (Sessions with completed quote) Expos 25.0% 20.0% 15.0% 10.0% 5.0% 100.00 80.00 60.00 40.00 20.00 Purchase Frequency (per 1,000) 0.0% - Weekday Morning Weekday Day (9-5) Weekday Evening Weekday Late (10PM+) Session Day/Time Weekend Night Saturday Sunday Data Summary Driver1 Age Model: Purchase Made During Session re (Sessions with completed quote) Exposur 18.0% 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 100.00 90.00 80.00 70.00 60.00 50.00 40.00 30.00 20.0000 10.00 Purchase Frequency (per 1,000) P 0.0% - 65+ 60-64 55-59 50-54 45-49 40-44 35-39 30-34 25-29 24 23 22 21 20 19 <=18 Age Driver 1 18 9
Data Summary Number Vehicles/Drivers Model: Purchase Made During Session re (Sessions with completed quote) Exposur 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 90.00 80.00 70.00 60.00 50.00 40.00 30.00 20.0000 10.00 Pu urchase Frequency (per 1,000) 0.0% - 1 / 1 1 / 2+ 2 / 1 2 / 2 2 / 3+ Policy Number Vehicles / Drivers 3+ / < 3 3+ / 3+ 19 Data Summary Bodily Injury Limit 20 10
Model Development 21 Modeling Techniques Neural Network Decision Tree Regression Analysis Ensemble 22 11
s Submitted Analysis 23 Submitted Likelihood Model Decision Tree 12
Decision Tree English Rules Low Estimate IF total_pages < 11.5 AND total_ssl_page < 9.5 Initiated: 0.4% High Estimate IF aig EQUALS 1 AND 11.5 <= total_ssl_page AND 4.5 <= visit_time AND means_of_entry IS ONE OF: WEBMAIL NON-REFERRED SPONSORED SEAR NATURAL SEARCH AND total_pages < 15.5 Initiated: 71.4% Submitted Number of Prior Visits 1.200 Number of Prior Site Visits R e l a t i v e 1.000 0.800 L i 0.600 k e l i Q 0.400 h o o d S u b m i t t i n g u o t e 0.200 1.000 0.889 0.779 0.668 0.557 o f 0.000 0 1 2 3 4 Number of Prior Visits Relative Likelihood of Submitting 26 13
Policy Purchased Analysis 27 Age of Driver Model: Purchase made during session 1.40 18% Relativity 1.20 1.00 0.80 0.60 0.40 41% 47% 67% 69% 76% 72% 131% 16% 117% 118% 111% 14% 104% 104% 103% 100% 12% 10% 62% 8% 6% Exposure Percentage 4% 0.20 20% 2% 0.00 0% Sessions with Completed Modeled Relativity One Way Relativity Age Driver 1 28 14
Education 2.50 Model: Purchase made during session 50% 45% 2.00 196% 40% Relativity 1.50 1.00 70% 100% 126% 131% 146% 155% 149% 35% 30% 25% 20% 15% Exposure Percentage 0.50 10% 5% 0.00 0% Sessions with Completed Modeled Relativity Primary Driver Education 29 Price Sensitivity One Session 1.20 Model: Purchase made during session 60% 1.00 100% 106% 50% Relativity 0.80 0.60 0.40 76% 79% 71% 52% 76% 40% 30% 20% Exposure Percentage 020 0.20 10% 0.00 0% Sessions with Completed Modeled Relativity Percent Final d Premium greater than Minimum d Premium 30 15
Price Sensitivity One Month 31 Regression Preliminary Results 1.20 Model: Purchase made during session 60% 115 1.15 115% 50% Relativity 1.10 1.05 1.00 0.95 100% 97% 40% 30% 20% Exposure Percentage 0.90 10% 0.85 0% Sessions with Completed Modeled Relativity Did user come to insurer s site from another insurer s site? 16
Unstructured Data 33 Search Phrase Analysis Categories Specific companies s auto, homeowners, life, insurance Value cheap, affordable, etc. Specific agents Circumstance - teen Other TurboTax, hydrogen fuel cell Process Remove punctuation and non-characters [./!#?*$%()-@] Parse phrase into words Summarize the frequency of each word Correct for misspellings Build cluster analysis to determine key word associations Insuranced insurance: insurance0 insurance3 insurances insurancer insurancej insurancel insuranceg insurrance inssurance insurannce insurancce iinsurance innsurance insuraance isnurance insuarnce inusrance insuranceco 34 17
Cluster Text Analysis 35 Summary & Additional Research 36 18
Cumulative Lift 37 Lift Chart 38 19
Additional Research Marketing Effort Population Target Description Auto Insurance Website Visitor Initiated Of those that visit the website, people with what characteristics are more likely to start the quote process? Sale Initiated Submitted Of those that initiate the quote process, what is the likelihood that they complete it? Of those that actually submit the quote, how many complete the purchase? Policy Bind Of those that actually submit the quote Submitted 39 Data Considerations Voluntary panel Only captures internet shopping trends Does not marry online shopping with offline shopping May not track same user on separate computers 40 20
Potential Applications 1. For potential insureds with a higher likelihood of purchasing a policy, steps can be taken within the quote and purchase process to ensure that t the customer does not drop out 2. For potential insureds with a higher than average likelihood of purchase, insurance companies would be able to market more aggressively to these potential insureds 3. For segments of potential insureds that have a lower than average likelihood of purchasing a policy, an insurance company can identify and investigate where these lower than average likelihoods are and try to increase the likelihood of purchase 41 Questions? Experience the Pinnacle Difference! 21