Session 73 PD, Predictive Modeling for the Marketing Actuary Moderator: Maria Patricia Marcelo Arellano, FSA, CERA, MAAA Presenters: Andy Ferris, FSA, FCA, MAAA Sarah R. Hinchey, FSA, CERA Patrick Sugent
Session 73 Panel Discussion Predictive Modeling for the Marketing Actuary Sarah Hinchey FSA, CERA, MAAA Patrick Sugent AndyFerris FSA, CFA, MAAA
Case Study: Cross-sell campaign Sarah Hinchey 2
We are Wombat Life Insurance Co. We want to use our customer data in a more intelligent way Where do we start? Step by step approach to turn insights into action 3
Step 1 Identify and involve key stakeholders Business Drivers Product Experts Data Gatekeepers
Step 2 Gather and aggregate data 5
Step 3 Explore data for insights Wombat : New contracts by business line 120,000 100,000 80,000 60,000 40,000 20,000 0 2011 2012 2013 2014 2015 Term Life Disability Universal Life Fixed Annuity 529 Savings $100,000,000 $80,000,000 $60,000,000 $40,000,000 $20,000,000 $- 2015 APE (Annualized Premium Equivalent) $500 $180 $2000 $5000 $2400 Term Life Disability Universal Life Fixed Annuity 529 Savings Cross-sell 2011-2015 Product No other contract Term Life Disability Universal Life Fixed Annuity 529 Savings Term Life 90% 1% 2% 0% 0% 7% Disability 85% 10% 0% 5% 0% 0% Universal Life 85% 0% 10% 0% 3% 2% Fixed Annuity 85% 10% 3% 2% 0% 0% 529 Savings 75% 15% 5% 4% 0% 1% 6
and define your campaign goal. Situation Term life sales are declining Complication Difficult to sell: 1% conversion ratio Resolution Data-driven lead generation for Term life cross-sell on 529 base
Step 4 Build and select predictive model Clean & prep data Train models on training data; evaluate on testing data Select best model Random Forest Cumulative Gains Chart Logistic Regression Cumulative Gains Chart 100% 100% 80% 80% Buy % 60% 40% Buy % 60% 40% 20% 20% 0% 0% Selection % Selection % Random Selection Modeled Selection Random Selection Modeled Selection By targeting the 30% of customers with highest propensity to buy, the Random Forest model captures 72% of the customers who actually bought, while the Logistic Regression model only captures 59% of the customers who actually bought. 8
Step 5 Design & execute campaign Score customers Identify target customer base Customer propensity to buy (50,000 customers with 529 plan) Propensity to buy 14% 12% 10% 8% 6% 4% 2% 0% Average propensity to buy for entire base 1 2 3 4 5 6 7 8 9 10 Decile (customer ranking) 9
and deliver leads to sales force Customer Score (Ranking) Group A 0.95 High B 0.85 High C 0.75 High D 0.65 Control E 0.35 Control Be mindful of exclusions Set control group Wait for results 10
Step 6: Collect and Monitor results Target Group (HIGH propensity to buy) Control Group (random propensity to buy) 9,000 Leads 1,000 Leads 9,000 Contacted 100% 1,000 Contacted 100% 4% Overall Conversion Rate 4,500 Reached 1,000 Meetings 400 Applications 50% 22% 40% 1% Overall Conversion Rate 500 Reached 100 Meetings 15 Applications 50% 20% 15% 360 Conversions 90% 10 Conversions 67% 11
and evaluate the impact Target group Eligible contacted Sales Before model 50,000 40,000 400 1% conversion After model 15,000 Top 3 deciles 10,000 9,000 High 1,000 random 370 4% high conversion 1% random conversion Wombat Life Insurance Co. realized nearly the same level of sales while reducing acquisition costs by 75% 12
Summary 1. Identify & involve stakeholders 6. Monitor 2. Gather results and and impact aggregate data 5. Execute campaign 4. Build model 3. Explore data for insights
Key thoughts Start small Involve key stakeholders early on Stay focused on impact 14
Section 73 Panel Discussion: Predictive Modeling for the Marketing Actuary Andy Ferris, FSA, FCA, MAAA, CFA
The Opportunity Set Applications of predictive analytics can be deployed to significantly improve a wide variety of core operations for life insurance companies. Sampling of Applications of Predictive Analytics in Core Operations for Life Insurers Producer Optimization Product Design & Pricing Sales and Marketing New Business & Underwriting Inforce Management Claims and Fraud Producer Recruitment Identification of individuals most likely to become a successful producer for a given manufacturer Producer-Client Matching Identify behavioral patterns and personality attributes associated with successful, lasting producer-client relationships; deploy tactics to optimize matches Target Marketing / Lead Generation Improve quality of leads by identifying those most likely to qualify & most likely to buy Producer Retention Segmenting existing producers and deploying customized tactics to support success and retention Up-Sell Programs Identify existing customers whose need for life insurance has increased, and who remain healthy. Offer increased face amount with limited underwriting Application Triage Identifying certain healthy individuals for which certain medical exams can be waived Underwriting Predicting mortality experience on a seriatim basis, using new data sources to supplement or replace certain traditional medical exams Cross-Sell Programs Identify existing customers who are likely to need and likely to buy a second product an annuity, a P&C product, etc. Deploy customized, targeted offers. -2- Customer Lifetime Value Enable calculation of customized individual CLV; deploy customized proactive tactics for retention, second offers, etc. Post-Level Term Offers Segment population based on current health risk, current life insurance needs, likelihood to buy. Deploy customized, targeted offers Retention Strategy Use customized, individual estimate of lapse likelihood to enable customized proactive and reactive tactics to improve retention effectiveness LTC Claims Management (Active Lives) For each active life, estimate the likelihood of developing certain cognitive or physical impairments, then proactively encourage healthy policyholder behavior to enable prevention LTC Claims Management (Disabled Lives) For each disabled life, estimate the likelihood of transitions between type of impairment (physical vs. cognitive) and associated level of care required (home health care, assisted care facility, nursing home), then proactively encourage healthy policy holder behavior Fraud Detection Identify potential overpayments of claims for LTC or related products
Current & Future Use of Applications Predictive Analytics in Core Operations A 2015 study indicated increasing use of analytics in core operations for insurers in the next 3-5 years. 100% 90% 94% 97% 94% 94% 91% 88% 80% 70% 60% 79% 84% 79% 78% 79% 78% In 3-5 years 50% 40% 30% 41% 42% 41% 42% 30% 36% 40% 38% 32% 20% 10% 18% 12% 12% Today 0% Sales Life P&C Marketing Product Design & Pricing Fraud Underwriting Claims Source: Global Digital Insurance Benchmarking Report 2015, Bain & Company. -3-
Observations on Common Challenges and Leading Practices Observations on leading practices will be shared around the following five categories: Categories of Leading Practices 1) Conducting a holistic assessment of the opportunity set 2) Prioritizing the applications and developing an overall roadmap 3) Developing a holistic project plan for each prioritized application 4) Developing a business cases for each prioritized application 5) Anticipating data challenges -4-
#1 Conducting a Holistic Assessment of the Opportunity Set The opportunity set varies by manufacturer, and is a function of the manufacturer s target markets, distribution channels, product offerings, and internal operations. A customized assessment can be completed in a relatively short period, and can be used to strategically prioritize tactics to optimize short term and long term benefits. Selected Leading Practices Conduct interviews with each functional group to understand current operational process, any planned or inflight initiatives, and to brainstorm together to visualize how the process could be re-designed, by introducing predictive analytics as a new tool. Approach each potential application as a classic business process redesign initiative, focused on deploying predictive analytics as a new tool to the enhance the existing business process to save time, save money, and/or increase efficiency. Be inclusive in gathering a robust set of opportunities / ideas none is too simple or too complex at this point; Seek to identify a robust set of ideas that span both short term and long term improvements; Create a friendly, engaging environment to stimulate thinking and challenge conventional processes. Be an ambassador for predictive analytics; Appreciate that some individuals haven t yet been exposed to the strategic value that predictive analytics can bring - educate and engage those folks Document the in the full set of opportunities identified in the assessment and share back with functional leadership for confirmation of the opportunity set. -5-
As Part of the Assessment, Document the Listing of the Potential Initiatives The full set of opportunities identified in the assessment should be briefly documented, and shared back with functional leadership for confirmation. ILLUSTRATIVE -6-
#2 Prioritizing the Applications and Developing an Overall Roadmap Assemble an overall roadmap showing the portfolio of analytics initiatives, including those in-flight and those newly proposed. Selected Leading Practices Seek a diverse portfolio of initiatives, ideally including some early easy wins as well as some longer term transformational impacts, balanced across functional and business units Ladder initiatives to build upon each other, and to collectively generate exponential benefits Leverage data sources, data subscriptions, as well as increasingly mature analytics capabilities Be sure you have functional and/or business unit buy-in for the sequencing, prioritization, and overall roadmap Illustrative Overall Roadmap 2016 2017 2018 2019 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Risk-Based Target Marketing v1.0 Risk-Based Target Marketing v2.0 Cross Sell Life Product to Existing P&C Base ILLUSTRATIVE Application Triage Producer Recruitment CLV Phase 1 - Lapse CLV Phase 2 - Profitability CLV Phase 3 Proactive Tactics LTC Claims (Disabled Lives) LTC Fraud Detection LTC Claims (Active Lives) Ongoing Initiative Phase One Initiative Phase Two Initiative Phase Three Initiative -7-
#3 Developing an Integrated Project Plan For Each Prioritized Application The project plan must include the business process change aspect in the business implementation phase, which is often the most challenging aspect in achieving business value creation in these initiatives. Selected Leading Practices Emphasize that applications of predictive modeling are valuable only after they have been implemented and enable more efficient internal process and/or otherwise generate tangible business value. When analytics initiatives fail, it is most frequently to do lack of appreciation for the business implementation and business process change aspects, not the statistical model build phase. The integrated project plan helps to manage expectations for the full set of activities required to achieve business value from a given predictive model. The integrated project plan helps you solicit and obtain the right level of resources and funding to ensure project implementation and success. Illustrative Integrated Project Plan ILLUSTRATIVE -8-
#4 Developing a Robust Business Case For Each Prioritized Application Create a robust business case, which encompasses costs and benefits throughout all phases of the project plan. Selected Leading Practices A robust project plan and business cases helps to ensure that funding remains ongoing for initiatives that span multiple phases If your company uses a standard structure or format, use that. If not, then show standardized project evaluation metrics, including NPV, IRR, Payback Period, etc. For cases that involved actuarial estimates (mortality, risk, etc.) be such to engage the relevant actuarial group, to use existing experience studies, validated assumptions, etc. where possible; leverage existing actuarial pricing assumptions where possible Engage finance to ensure consistency of finance-related assumptions Be transparent, and solicit buy-in and belief in the business cases from all stakeholders Illustrative Business Case $40 M $30 M $20 M $10 M $0MM $10 M $20 M $30 M ILLUSTRATIVE Projected Costs and Benefits $8 $14 $20 $3 $5 $4 $5 $3 $10 $7 $9 $6 $0.4 $2.0 $12 $14 $16 $19 $0.2 $0.2 $0.2 $0.2 1 2 3 4 5 6 Projected 5 Year Net Cash Flow Summary (Estimates in $M) Benefits Levers Increase Application Volume & Placement Rate $30.9 Decrease Medical Requirement Costs ILLUSTRATIVE $12.6 Decrease Underwriting Operations Cost $23.2 Total Benefits $66.7 Cost Levers Provision for deterioration in mortality experience ($43.5) Initial and External Modeling Costs ($2.7) Total Costs ($46.2) 5 Year Project Summary Net Present Value of Net Cash Flows (NPV) $20.4 Internal Rate of Return (IRR) 161% Simple Payback Period 2.0 years -9-
#5 - Anticipating Data Challenges Data considerations must be proactively pursued and explored to avoid potential roadblocks and delays. Selected Leading Practices Anticipate that extraction of data from internal systems (admin, commission, claims, reinsurance, etc.) for use in model build will require more time and resources than expected Actively explore how far internal data goes back in time, how it is formatted, as well as completeness and accuracy of such data Proactively consider current data collection / storage conventions, and whether there is a different level of granularity, format, and structure to deploy going forward to better enable future data & analytics initiatives Actively explore data sources from external vendors, which can often be obtained relatively quickly and inexpensively, and can enable quick wins, perhaps while more robust models using internal data are developed over a longer horizon Actively consider contractual and permitted use restrictions on data purchased from external vendors Anticipate a growing set of data sources in the future vendors, wearables, electronic health records, social media, location data, etc. It is sad that nowadays there is so little useless information. Oscar Wilde, Irish playwright, novelist, and poet The goal is to turn data into information, and information into insight. Carly Fiorina, Former CEO of HP An approximate answer to the right problem is worth a good deal more than an exact answer to a proximate problem. John Tukey, American Mathematician -10-
Deloitte refers to one or more of Deloitte Touche Tohmatsu Limited, a UK private company limited by guarantee, and its network of member firms, each of which is a legally separate and independent entity. Please see www.deloitte.com/about for a detailed description of the legal structure of Deloitte Touche Tohmatsu Limited and its member firms. Please see www.deloitte.com/us/about for a detailed description of the legal structure of Deloitte LLP and its subsidiaries. Copyright Deloitte Consulting LLP. All rights reserved -11-