RRD And Edward Don Net30Score Case Studies April 2010
- 1. © 2010 PredictiveMetrics, Inc. Confidential & Proprietary Cashing In With Statistical Portfolio Scoring Edward Don & Company and RR Donnelly Case Study Summary Maximizing results by understanding risk
- 2. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 2 Edward Don & Company Case Study
- 3. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 3 About Edward Don & Company Distributor of equipment & supply to foodservice industry. “Everything But The Food” National distribution network – some international Approximately 35,000 active customers with an average balance less than $2500 Centralized credit office with 5 direct reports with a staff of 40 credit and collection analysts Sell to restaurants, lodging, healthcare and institutions through a sales staff of 350 Most profitable accounts are the Independent operators (mom & pop’s), but are also the most challenging from credit and collection perspective Owned and operated by the Don family since 1921, Edward Don & Company (Don) is the world's leading distributor of foodservice equipment and supplies
- 4. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 4 Why Did Don Embrace Risk-Based Collections? Don averages close to 1 collector for every 1,500 customers – it is difficult for the collector to determine which customers they should focus on, when to focus on them and what treatment to apply using just aging information. Don found that using only aging to prioritize collections, many of our low dollar, high risk customers were being ignored. Don also reviewed many orders that were placed on hold, but those that stopped were often contacted too late and many that were put on hold were low risk and the orders were being released anyway. Collection activity was managed and prioritized solely on the aging and highest dollar customer accounts. Was looking to gain collection efficiencies due to the size of the customer portfolio, the number of customers going on hold and the limited number of collectors
- 5. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 5 Don – Collection & Portfolio Management Process Prior to March 2007 Portfolio Risk Analysis was determined by a judgmental scorecard based on a combination of bureau and internal data. Obtained statistical credit scores from credit bureau (Experian) on a quarterly basis and created a a “blended score” with an A to E matrix, based on 70% payment to Don measured by DBT (Days Beyond Terms) and 30% based on the credit bureau score. Collection activity was prioritized based on the customer’s risk score. System generated credit holds were based on Risk Class Score, C & D accounts would go on hold if a violation occurs in any one of 9 credit rules. The process resulted in excessive holds which credit analyst had to frequently review resulting in 75% of holds being released without contacting sales and 98% of accounts eventually being released from credit hold. In addition since only 10% of accounts were ranked A - low risk, Don thought they may be missing opportunities to increase sales with that customer segment. Don implemented a risk-based collections methodology more than 6 years ago based on bureau scores
- 6. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 6 Don – Collection & Portfolio Management Process Prior March 2007 Bureau based model produced the following customer segments 13.2% of Accounts are Ranked A (9.4% of $) 21.4% Ranked B (17.2% of $) 32% Ranked C (32.6% of $) 33.4% Ranked D (40.8% of $) Don was using a bureau based model to assess risk and accounts could not be rolled up to the parent level and had to be scored at the ship-to level
- 7. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 7 Don – Collection & Portfolio Management Process Implemented March 2007 Model focused on using only Don’s internal A/R and performance data, data which has proven to be more predictive for small and mid-sized customers. After several tests, final model validated based on Don’s customized “BAD” definition - more than 50% of monthly outstanding balance goes 61 or more days past due at sometime during the 6 month period after the scoring date. For Don it increased the number of A&B (low risk) accounts and reduced the number of C & D (high risk) accounts allowing collectors to truly focus on problems and collection opportunities based on risk assessment that was proven to be accurate. End result is that Don’s collectors are spending more time collecting and less time releasing orders. After having PMI conduct a Net30Score Validation Analysis and being satisfied with the predictive results Don made the decision to use Net30Score and ScoreMiner, PMI’s hosted web-based reporting and querying application
- 8. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 8 Don Customer Segmentation Prior vs. Post Statistical Portfolio Scoring Statistical based collection / portfolio score with roll-up to parent level 25% Risk Class A 39% Risk Class B 21% Risk Class C 15% Risk Class D 36% of customers rank risk class C and D Bureau based score that could not score at the parent level 13.2% Risk Class A 21.4% Risk Class B 32% Risk Class C 33.4% Risk Class D 65.4% of customers rank risk class C and D
- 9. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 9 Summary of Benefits Gained From Statistical Portfolio Scoring Significantly improved collection department performance by prioritizing collection resources based: first upon inherent risk, and second upon account aging 25% increase in phone contacts, which drove cash flow improvement and lower DSO. DSO reduction of 8% during first 24 months Significant reduction in the number of accounts going on credit hold, increase customer satisfaction Reduction in credit bureau costs, no longer purchase the quarterly scoring updates Using ScoreMiner for customer reporting needs, including filtering and querying, segmentation, comparative and rate of change analysis Most important we saw a productivity gain as collectors spent less time reviewing and releasing orders and more time productively contacting customers. During the worst economic downturn it has ever seen, Don has been able to reduce the overall level of risk in it’s customer portfolio
- 10. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 10 RR Donnelley Case Study
- 11. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 11 About RR Donnelley (RRD) Provides print and related services, including business process outsourcing, print fulfillment, transactional print-and-mail, print management, online services, digital photography, color services, and content & database management. Products and related service offerings include magazines, catalogs, retail inserts, books, directories, commercial print, financial print, direct mail, forms, labels, office products, pre-media, and logistics services. Client Financial Services are centralized domestically in Chicago and Reno with a six person management team and 51 credit and collection analysts, 3 credit coordinators and 6 collectors. Have an average 65k open bill-too’s each month, 75% of which have an average balance of less than $2,500 Use SAP and have complex invoicing solutions and multiple systems feeding target platforms. RRD was founded in 1864, based in Chicago and is a public company with 2008 revenue of 11.5B
- 12. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 12 Why Did RRD Embrace Risk-Based Collections and Portfolio Management? RRD was using systemic rules, aging analysis, and dunning response, which seemed reactionary, to drive day-to-day collection activity Our team composition varies greatly in collector to account responsibility ratio and it was very difficult for our team to maximize the time allotted to proactively collect Outbound collection activity was being diverted to handling: inbound calls from unnecessary dunning, accounts where relationships were ‘known’ as well as customers where risk of loss was low As our ERP uses a formula to put account on hold that didn’t use risk, we were reactionary as we were trying to collect the last order before producing the next. This was time consuming, and not within the partnering spirit we aim to forge with sales and clients We were looking for a proactive solution to compliment to our system driven methodologies for collections
- 13. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 13 RRD – Collection & Portfolio Management Process Prior to July 2008 The target platform relied on aging based segmented dunning, systemic holds and our collections software The on-boarding platforms relied on relationships and highly skilled personnel Implemented collection software to manage and automate collector work flow. Collection software used a rules based scorecard to drive daily work lists that was based on aging metrics, customers segmentation and collector workload expectation The collection software was never fully embraced Our target A/R platform (SAP) was undergoing an upgrade to a newer version Partially integrated strategies existed due to acquisitions.
- 14. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 14 RRD– Portfolio & Collection Management Process After July 2008 After Net30Score go-live and before our SAP upgrade PMI validated and deployed Net30Score quicker than our systems upgrade and we were able to start to take ‘advantage of’ and ‘experiment with’ the definitions, expectations and scoring outputs PMI provided an on-line management tool (ScoreMiner) PMI allowed for ad hoc reporting capabilities right away PMI was able to customize fields to enable easier reporting PMI created search fields useful for us PMI included all provided data which allowed for comprehensive and detailed usability PMI provided training for our staff After having PMI conduct a Net30Score Validation Analysis and being satisfied with the predictive results, RRD made the decision to use Net30Score and ScoreMiner to help manage the collection process
- 15. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 15 How RRD Leveraged Statistical Collection & Portfolio Score In The Collection Process RRD improved our new worklist tools by using the PMI scoring into our workflow model RRD was able to work with our partners at D&B to provide additional eRAM scorecard data that fit our needs. The Net30Score outputs are now weighted variables in our D&B eRAM scorecards. RRD senior management uses PMI’s ScoreMiner for ad-hoc reporting and querying. Analysts and collectors who use the scoring day to day now access the data through SAP and eRAM but also have access to ScoreMiner After the SAP upgrade Net30SCore was embedded into our systems and solutions
- 16. Confidential & Proprietary © 2010 PredictiveMetrics, Inc. 16 Summary of Benefits Gained From Statistical Portfolio Scoring For the portion of our portfolio using Net30Score from July 08 until the SAP upgrade in Aug 09, RRD saw the following: Total past dues (TPD) improved from $138.1M to $110.1M 120+ days dues went from $20.2M to $5.5M Since Aug the trends continued TPD are $101.6M & $1.1M is over 120 days Major improvements in the 1-30 area and aged 120 areas attributable to focusing those hours of the day on the right customers based on risk Reduction in reliance on systemic holds for past dues and dunning to drive collections (shift from reactionary to proactive management – working smarter not harder) Final Comment: Getting in early allowed us the time to work out kinks, get the team familiarized with the product, and include the model into other processes and system solutions. It also allowed us to start to see results even before the SAP upgrade and full effect took root, which helped with organizational buy-in with management and the functional user community. RRD has been able to integrate the PMI scoring into our dunning and collection work list processes (SAP FSCM), as well as into eRAM and has helped us improve each of those areas.