Settlement Optimization Maximizing Returns in Late Stage Collections Paul Robinson Associate Vice President Canadian Tire Bank Matt LaHood Sr. Director FICO © 2014 Fair Isaac Corporation. Confidential. This presentation is provided for the recipient only and cannot be reproduced or shared without Fair Isaac Corporation’s express consent. Agenda ►Canadian Tire Overview ►Optimization ►Settlement ►CTB ►Key 2 Process Overview Optimization Analytics Experience and Early Results Learnings © 2014 Fair Isaac Corporation. Confidential. Canadian Tire Overview Canadian Tire is one of the most recognized brands in Canada comprised of a diversified retail segment and a robust financial services business. Canadian Tire Corporation (CTC) Canadian Tire Retail Segment Canadian Tire REIT Segment Financial Services Segment Glacier Credit Card Trust © 2014 Fair Isaac Corporation. Confidential. Canadian Tire Overview Canadian Tire operates over 1,700 stores and its Family of Companies employ over 85,000 Canadians. CTC Quick Facts Founded in 1922 by the Billes family who still own a controlling stake Diversified retail banners including Canadian Tire Retail, PartSource, Sport Chek, Sports Experts, Pro Hockey Life and Mark’s 90% of Canadians are within 15 minutes of a CTR store; 60% of Canadians visit a CTR store each month 31 million square feet of retail space under multiple banners Market leadership in Automotive, Fixing, Living and Playing categories, as well as Men’s Industrial and Casual Apparel Named Marketer of the Year (2013) Named one of North America’s Top 50 Brands (2014) One of Canada’s Best Managed Companies (2013) Canada’s Second Most-Trusted and Loved Brand (Ipsos) © 2014 Fair Isaac Corporation. Confidential. Optimization Strategy Complexity CTB’s Optimization Journey Authorizations Optimization V1 Settlement Optimization V1 Initial Credit Limit Assignment V1 Settlement Optimization V2 Credit Limit Management V3 Early Stage Collections V1 Late Stage Collections V1 Credit Limit Management V1 2007 2008 © 2014 Fair Isaac Corporation. Confidential. Credit Limit Management V4 Credit Limit Management V2 2009 2010 2011 2012 2013 2014 Optimization Process Overview 6 © 2014 Fair Isaac Corporation. Confidential. Unlocking Value from Optimization Four Key Steps Accelerated Learning ► Test efficiently to learn beyond historical operating regions to further increase future performance Design Framework Implementation & Tracking ► Gain insight into key profit drivers and opportunity pockets through diagnostics and final strategy engineering © 2014 Fair Isaac Corporation. Confidential. Establish mathematical relationships between customer treatment options, their reactions and profitability Simulation and Optimization Reporting & Analysis 7 ► Optimization Software (Decision Optimizer) Interpretation ► Decision Modeling Identify optimal strategy scenarios subject to your multiple goals, business constraints and forecasts for the future Step 1: Model Customers’ Reactions to your Actions Customer Risk score = 680 Rev Balance = $10,000 Rev Util = 61% Time in File = 132 Segment = A Action Offer 80% Settlement 1 E(Pay) = 35% E(Loss) = 65% E(NPV) = -$3,750 Offer 60% Settlement 2 E(Pay) = 50% E(Loss) = 50% E(NPV) = -$2,000 Offer 40% Settlement 3 E(Pay) = 70% E(Loss) = 30% E(NPV) = -$200 As Action Effects built into Decision Model it also allows you to run thousands of scenarios very quickly 8 © 2014 Fair Isaac Corporation. Confidential. Reaction Step 2: Consider Scenarios and Select Operating Point Efficient Frontier Choose the optimal operating point from multiple choices Projected Profit $$$$$$ Scenario G Increase profitability without incurring additional losses $$$$$ F G H I Efficient Frontier J E D Current Operating Point Where you are today $$$$ C $$$ B FICO Optimization helps you understand all options Scenario B Maintain profitability per account and decrease settlements $$ A $ $ $$ $$$ Total Dollars Settled 9 © 2014 Fair Isaac Corporation. Confidential. $$$$ $$$$$ Step 3: Operationalize Benefits ► Optimized treatments can be converted into decision trees and loaded into collection system in order to execute consistent decisions every month, week, day, etc. ► Trees should have interpretability, robustness and ease of implementation ► Provide 10 visibility into strategies and pockets of opportunity © 2014 Fair Isaac Corporation. Confidential. Step 4: Accelerate Learnings 11 © 2014 Fair Isaac Corporation. Confidential. Settlement Optimization Analytics 12 © 2014 Fair Isaac Corporation. Confidential. Examples of Where FICO Has Applied Decision Modeling and Optimization in Collections Area Prediction Example Objective Pre-delinquency Potential for customer to become delinquent Minimize collection expense Early stage collections Priority and treatment of accounts Maximize profit Mortgage workouts Restructure terms Maximize NPV Minimize loss Credit card settlements Which accounts to settle pre-charge off Settlement after charge-off Who is the most likely candidate for a settlement and for what $ amount Maximize money collected Maximize profit Agency placement Which accounts to place with which agency Maximize profit Maximize money collected Collections optimization What is the overall potential impact Minimizing costs Maximizing return 13 © 2014 Fair Isaac Corporation. Confidential. Minimize loss Settlement Offers Are a Powerful, Yet Underutilized, Arrow in the Collection Manager’s Quiver ► Too much time, effort, and resource is spent pursuing borrowers who won’t fully pay ► Settling with the right people for the right amount can accelerate dollars collected while freeing resource to work accounts where their efforts make a difference ► Identifying the right settlement offer for the right borrower at the right time is immensely difficult 14 © 2014 Fair Isaac Corporation. Confidential. Key Collections Decisions Early Stage Collections 30 60 Late Stage Collections 90 120 150 Recovery 180 Primary Secondary Decisions Charge-off 15 Which accounts to contact How to contact the accounts When to contact the accounts © 2014 Fair Isaac Corporation. Confidential. How to work out the accounts (settlement, payment plan options) Which accounts to send to early outs Where to place the account (which agency or attorney) Which accounts to sell (for how much) How Does Your Institution See the Problem? ► Profit ► Revenue ► Loss Rates ► Revenue “Give-back” ► Payment Program Loss Rates ► Collection Costs But much of this potential is lost when settlement policies rely on a simple rules based approach… 16 © 2014 Fair Isaac Corporation. Confidential. With Optimization, Conflicting Goals Are Understood and Balanced 17 © 2014 Fair Isaac Corporation. Confidential. With Consumer Response Included Illustrative Example of Action-effect Model for a Settlement Plan In this oversimplified example for Segment 2, settlement percent of 40% has a higher NPV E(NPV | 70%) = 70% * $10,000 * 41%-$10,000 * (1–41%) = -$3,030 E(NPV | 40%) = 40% * $10,000 * 70%-$10,000 * (1–70%) = -$200 While optimization is running extremely sophisticated mathematics behind the scenes; it is NOT a black box 18 © 2014 Fair Isaac Corporation. Confidential. Settlement Optimization Influence Diagram Settlement offer decisions range from 100% (No Offer) to 20% of balances outstanding, with impact on Net $Paid captured by action-effect models Offer Not Paid Call Net $ Paid No Contact Offer No Call 19 Settlement © 2014 Fair Isaac Corporation. Confidential. Collector (In-house, Agency) P(Charge-off) CTB Experience and Early Results 20 © 2014 Fair Isaac Corporation. Confidential. CTB Settlement Optimization Objectives ► Premise: Circumvent future collections efforts and agency recovery fees through proactive settlement on impaired accounts across all balance ranges and past due statuses ► Objectives: ► Gather Data: Target actions to past due accounts that align with the model development population (high charge-off rates and probability of listing), but collect response information to enable improved future model precision ► Execute Well: Leverage offer response information to improve future settlement offer positioning and perfect operational execution ► Repeat: Create a sustainable business based on experiences from pilot and revised models 21 © 2014 Fair Isaac Corporation. Confidential. CTB Settlement Optimization: Getting Started ► Decision Modeling ► CTB had little experience with proactive, in house settlement offers ► Used limited agency data to help build models ► Simulation and Optimization ► Target accounts with high probability of charge off and listing ► Action space: 20% to 90% of balance ► Interpretation ► Test vs. Control framework leverages existing TRIAD implementation to select accounts for Settlement Optimization (Action) and business as usual (Control) ► Created a dedicated outbound/inbound collections team from scratch ► Accelerated ► 22 Learning Closely monitor the results to ensure enough successful settlements for future model development © 2014 Fair Isaac Corporation. Confidential. CTFS Settlement Optimization Testing Framework CTB Settlement Optimization Testing Framework ► NOT business as usual as there are inherent differences in tactic execution ► Action ► ► (liquidate): Attempt to sell settlement, if no success list to agency for full balance The lifetime stream of payments has been brought forward to a shorter window ► Control ► ► 23 (cure): Attempt to cure the account using normal practices Write offs and recoveries will come over a longer period in the future © 2014 Fair Isaac Corporation. Confidential. CTB Settlement Optimization: Early Results ► The Action Group clearly brings write offs forward in time when compared to the Control group… ► …and the Action Group has higher overall payments (internal and Agency) when compared to the Control Group Cumulative % of Balance Paid by Month queued, Monthly Cohort Average Cumulative % of Balance Written-off by Month queued, Monthly Cohort Average 1 2 3 4 5 6 7 8 Action 24 © 2014 Fair Isaac Corporation. Confidential. 9 10 Control 11 12 13 14 15 1 2 3 4 5 6 7 Action 8 9 10 Control 11 12 13 14 15 CTB Settlement Optimization: Contact Strategy ► Contact tactics play an important role in obtaining a Settlement arrangement Contact Rate By Letter Type By Cohort 35% ► Dialing campaigns varied; outbound, managed dial, limited skip etc. campaigns and follow up calls to outbound calling, tested three types of letter contact: 25% ► 20% ► Concurrently Specified Settlement Amount vs. ► Generic ‘Helpful Solutions’ vs. ► No letter ► Ongoing concern that a Settlement offer in writing may impact value on future sale of receivables 25 30% © 2014 Fair Isaac Corporation. Confidential. No Letter Specific Letter Generic Letter CTFS Settlement Settlement Optimization – Valuation Valuation CTB Optimization: ► Timing Dynamic Value Metric Write Offs Action Control Brought Forward Eventually catch up? The ‘cash flow’ dynamics of the Action group are fundamentally different vs. the Control group, leading to negative valuation in the short term Average Difference in Cumulative Net Loss Ratio (Action - Control) (Less) Recoveries Brought Forward Eventually catch up? 16% (Less) Brought Forward Eventually catch up? /Divided by Receivables = Equals Net Loss Ratio Higher in Short Term Higher in Long? Term Write Offs – Recoveries – Agency Fees = Net Loss Ratio Receivables Cumulative Net Loss Ratio Difference Agency Fees 14% 12% 10% 8% 6% 4% 2% 0% -2% 1 2 3 4 5 6 7 8 9 Months after Queued 26 © 2014 Fair Isaac Corporation. Confidential. 10 11 12 13 14 CTFS Settlement Settlement Optimization – Contingent Liability CTB Optimization: Contingent Liability ► The valuation method does not entirely capture the reduction in high risk contingent liability ► The ‘remaining balance at risk’ reflects the proportion of the original balance that has yet to be paid, recovered or charged off—the Control group has much larger contingent liability vs. the Action group Action Receivables 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% 1 27 Control Receivables 2 3 Receivables at Risk © 2014 Fair Isaac Corporation. Confidential. 4 Total Losses 5 6 7 8 Total Payments & Recoveries 1 2 3 Receivables at Risk 4 Total Losses 5 6 7 8 Total Payments & Recoveries Key Learnings 28 © 2014 Fair Isaac Corporation. Confidential. CTFS Settlement Optimization – Key Learnings CTB Settlement Optimization: Key Learnings ► CTB can find customers willing to accept and pay a Settlement offer, using good modeling techniques and effective operational execution ► Settlements generates significant improvements in overall recoveries, for both in-house and agency collections ► Earlier ► actions drive earlier listings and write-offs Timing of actions has potential to significantly disrupt normal portfolio cash-flows ► Test/Learn/Rollout ► approach Start with existing response information, and plan data collection up front to enable better models and coverage for the future ► Early and clear communication with internal stakeholders regarding expected impact of new settlement strategy on both recoveries and write-offs is key to ongoing program success 29 © 2014 Fair Isaac Corporation. Confidential. Thank You! Paul Robinson Paul.Robinson@ctfs.com Matt LaHood mlahood@fico.com 415.690.0768 © 2014 Fair Isaac Corporation. Confidential. This presentation is provided for the recipient only and cannot be reproduced or shared without Fair Isaac Corporation’s express consent. Learn More at FICO World Related Sessions ►Collection Platform Success in a Transitioning Economy ►The New Normal: Adopting and Adapting to Drive Sustainable, Profitable Growth Products in Solution Center ►FICO® Decision Optimizer ►FICO® Debt Manager ►FICO® Customer Communication Services: Collections Experts at FICO World ►Cheryl Miller ►Ana Marcos ►Matt Stanley ►Matt LaHood White Papers Online ►Five Imperatives in a Shifting Collections Landscape ►Harnessing the Speech Analytics Advantage Blogs ►www.fico.com/blog 31 © 2014 Fair Isaac Corporation. Confidential. Please rate this session in the FICO World App! Paul Robinson Paul.Robinson@ctfs.com 32 © 2014 Fair Isaac Corporation. Confidential. Matt LaHood mlahood@fico.com