Credit Scoring Models ICTF – Ft Lauderdale, FL November 13, 2012 Scott Tillesen Vice President, US Credit Services Tech Data Corporation
ICTF November 13, 2012 Agenda Today’s three topics: Background on Tech Data credit organization. Tech Data history of scoring models. Recent transition to integrated scoring tool.
About Tech Data Corporation Global distributor of computer hardware, software, and consumer electronics. $27+ Billion annual revenue. Product brands include: HP, Cisco, Apple, IBM, Lenovo, Symantec, Microsoft, NEC, Sony, and 100’s of others.
Tech Data Credit Operations - Europe 16+ credit offices Transitioning from insured risk: Trade references People Policies and practices Systems (including scoring)
Tech Data Credit Operations – The Americas 9 credit offices Almost exclusively owned risk. Centralized US credit office. Scoring (of various kinds) in use for fifteen years.
Scoring Transformation Uplift Analysis Off-Line Scheduled Reviews New Account Matrix Uplift Analysis with Sophistication Integrated and Automated Scheduled Reviews
1. Uplift Analysis Objective: Pro-actively increase of credit limits on worthy accounts. Elements Included: Current credit limit Average paydays Account status code
Uplift Analysis - Results Mass uplifts every six months (or so). Increase the credit limits of thousands of customers. Increased credit limits by millions of dollars. Virtually no associable bad debt or slow pay.
2. Off-line Scheduled Reviews Objective: Improve staff productivity. Elements Included: Basic “Uplift” elements – credit limit, status codes, and paydays. Added – Net worth External scores Previous payment plan Security
2. Off-line Scheduled Reviews – Results Drove next review date. Notes added to legacy system via macro. Handled over 50% of reviews with credit limits under $200K. Virtually no associable bad debt or slow pay.
3. New Account Matrix Elements Included: Personal Guaranty Objective: Improve consistency of decision and credit staff productivity during the review of new accounts. Elements Included: Personal Guaranty Signer’s personal credit score
3. New Account Matrix – Results Consistent decisions. Improved decisions. Faster processing. Avoided need for financial statements. Credit limit range: $5K to $20K Virtually no associable bad debt or slow pay.
4. Uplifts with Sophistication Objective: Pro-actively increase credit limits on worthy accounts. Elements Included: Basic “Uplift” elements – credit limit, status code, and paydays. Added – % past due, account age, net worth, external scores, previous payment plan, security, and credit utilization.
4. Uplifts with Sophistication – Results Added variables improved confidence. Recent uplift of 1,700 accounts by $25 million. Identification of 120 additional accounts for likely increase.
5. Integrated and Automated Scheduled Reviews Objective: Further staff productivity improvement and added sophistication. Elements Included: Traditional (38%): Age of account, F/S date, credit app signer credit score Payment portion (49%): NACM risk class score, Experian Intelliscore, Smyyth PQI score, previous payment plan, and paydays. Z score (13%) Kick out factors: stale f/s, bad external scores, net worth, status code
5. Integrated and Automated Scheduled Reviews - Results Required custom risk management system development. Handles over 50% of reviews with credit limits under $300K. Virtually no associable bad debt or slow pay.
Future Possibilities Socialize the scoring model score. Socialize the model’s recommended credit limits. Statistical identification of significant variables.
Scott Tillesen Vice President, U. S Scott Tillesen Vice President, U.S. Credit Services Tech Data Corporation Scott.Tillesen@techdata.com 727-538-5880 www.techdata.com/credit www.linkedin.com/in/tillesen