RETURN FRAUD IDENTIFICATION JOE HENDERSON CFI REGIONAL LOSS

RETURN FRAUD IDENTIFICATION JOE HENDERSON, CFI REGIONAL LOSS PREVENTION MANAGER

AGENDA • Overview of the Auto. Zone Loss Prevention Team • Returning Merchandise – Understanding the customer experience • Store Level Tools • Field Level Tools • Case Study • Q&A AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

THE AUTOZONE LP TEAM • LP Process Improvement Team • 3 project managers led by LP Process Improvement Manager • Identify gaps in store process and procedure • Work cross functionally with Store Operations, Supply Chain, IT, etc. • Develop and implement methods to close gaps • Reverse engineer shrink related scenarios (questionable shrink results, billing issues, etc. ) • LP Analytics • 2 Analysts led by LP Analytics Manager • Responsible for all things analytical from an LP perspective • Reporting tools • Predictive Modeling • Regression Analysis AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

THE AUTOZONE LP TEAM • Divisional Loss Prevention Managers • Team of six (6) DLPM’s led by Director of LP Field Ops, Analysis, and Process Improvement • Responsible for approximately 1, 000 stores and 8 -10 RLPM’s each • Divisional Investigators • Fast paced team of five (5) experienced Loss Prevention professionals led by the Director of LP Field Ops, Analysis, and Process Improvement • Responsible for coaching, training, and development of the field LP team • Primary focus on high-risk regions • Regional Loss Prevention Managers • Team of 54 RLPM’s led by one of the six (6) DLPM’s • Responsible for all aspects of Loss Prevention, Safety, and Environmental compliance AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

RETURNING MERCHANDISE – THE CUSTOMER EXPERIENCE • Our returns processes are unique versus other specialty / small-box retailers in that we require customers to pass through our parts counter and then POS before receiving funds from a return • Directs all returned merchandise to be verified by a member of management at the initiation of a return • Condition of merchandise (Damaged vs Undamaged, Auto. Zone part vs OEM, etc. ) • Merchandise warranty is verified (if applicable) and testing is performed (is the specific part the issue? ) • Non-Receipted returns are validated through 3 rd party software • Customer identification is verified through warranty database and customer signature is captured • Random 2 nd level management password required on high dollar items AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

STORE LEVEL TOOLS • Same day verification of all returns processed by the Store Manager • Regardless of who “generated” a return, the Store Manager must verify the return using an RF Scanner • Aside from the Auto. Zoner who generated the return, this provides our 1 st opportunity to identify a potentially fraudulent return • Provides an electronic footprint that signifies: • The part was present in the store at the time of the return • The return is a legitimate return (right part, right package, etc. ) AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

STORE LEVEL TOOLS • Return Signature Report • As a part of the Return process customers signatures are captured • Based on the assumption that a customer was physically present at the time of the return • Can be used to compare signatures from prior returns • Identifies returns where the employee who generated a return was also the customer • Helps the store or field LP team identify patterns AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

STORE LEVEL TOOLS • Manager’s Next Day Review Process • 2 nd opportunity for a Store Manager to validate a specific return • Multiple step process performed the day following a return using the RF Scanner • Check QOH for return to stock merchandise • Verify damaged returns that are returned to DC for store credit • Provide electronic footprint that “manager dispose” merchandise has been destroyed and discarded per vendor agreements • Validate reasons for non-payment on special order merchandise • System logic will direct the manager to contact customers via telephone to further validate the return • Only select / high-risk returns are populated to this tool and are based on dynamic criteria • Item was / was not sold at returning store in last 90 days • Item was returned early in the day or late at night • Item was not found in our national warranty database (Forced Return) • Return was processed at point of sale more than 15 minutes after it was created • Return Denial was overridden by member of management based on Returns Authorization logic AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

STORE LEVEL TOOLS (CONT. ) • Uncredited Returns Reporting • Allows real-time visibility to merchandise that was returned and store credit was not given by the DC • Gives the field LP Team visibility to exceptions within their store hierarchy • Helps identify training / execution opportunities, as well as integrity issues AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

STORE LEVEL TOOLS (CONT. ) • LP Dashboard • Enterprise wide reporting / awareness tool that is visible to every Auto. Zoner in the organization • Updates weekly and provides details to shrink inducing activities • Store Manager is required to print and review with each Auto. Zoner AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

STORE LEVEL TOOLS • • • Monthly communications that target Return Fraud Quarterly awareness posters placed in areas visible to Auto. Zoners Period training (compliance tracked by field LP team) “Shrink. Buster” recognition program New-hire Shrink. Buster certification / training material AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS The Field LP Team takes a multi-layered approach to identifying Return Fraud. In addition to the store level tools, the Field LP team uses the following resources: • Proact • P&L Exception Reporting • LP Dashboard • Returns Authorization • Predictive Modeling AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS - PROACT • Dynamic toolbox to identify Return Fraud • Return Rate, Return Dollars, Average Return Dollars by CSR • Number of Days Between Sale and Return • Returns Over 15 Minutes • Multiple Returns of the same SKU by the same CSR • Forced Returns • SKU / Activity specific Returns by CSR AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS - PROACT • Proact is dependent on user ability to extract and interpret data • Reviewing changes in return dollars, return count, Avg Ticket etc. won’t identify the employees with experience in Return Fraud • Could lead to “data overload” • Proact is set up to drive the Field LP Team to low-level data / exceptions • Targeted queries that are highly correlated to Return Fraud AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS - PROACT Days Between Sale & Return • Identifies undamaged returns that fall outside of the 90 day returns policy • Identifies undamaged returns on high dollar items that could indicate “warranty surfing” • Items like starters, alternators, batteries, fuel pumps, and ignition parts have a warranty registered to them • Over time, the likelihood of the part failing reduces significantly • Auto. Zoners could “surf” through warranty files looking for old parts that have not been returned • The higher the number of days between the sale and return / the higher likelihood of fraud AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS - PROACT Returns Over 15 Minutes • Returns that take longer than 15 minutes to process at POS increases the likelihood of fraud • We can identify Returns Over 15 Minutes through Proact, the Managers Next Day Review, and other store level reporting • If the same CSR generated and processed the PCI, the return is highly correlated to fraud AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS - PROACT Multiple Returns / Same Part / Same CSR • Highly indicative of Returns Fraud • Identify patterns of low-risk / low-visibility items • Items commonly found around the parts counter, POS, and items that commonly have QOH issues AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS - PROACT Forced Returns • Returns processed on items that could not be found in our National Warranty Database • Identifies potentially fraudulent returns where a part was taken from the shelf and no customer was present AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS – P&L EXCEPTION REPORT • Finding a needle in a stack of needles in a needle factory • 5, 000+ stores across continental United States, Mexico, and Brazil • Exception Report designed to identify the bottom 15 of a Division or Region in regards to: • Average Ticket Change vs Last Period • Return Rate Change vs Last Period • Undamaged / Damaged Return Rate vs Last Period • Gross Through the Register vs Last Period • Ranks each individual store (benchmark against peers) • Lists the number of exceptions by store • Stores with higher number of exceptions = potential opportunity AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS – LP DASHBOARD • At the field level, the LP Dashboard is a tool that helps drive the Field LP Team where the exceptions exist • Stores with exceptions at the individual Auto. Zoner level are highlighted in red • Also helps identify execution issues around Recalls, Uncredited Returns, Returns Authorization Overrides, and Training compliance AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS – RETURNS AUTHORIZATION • Auto. Zone uses TRE for an “off the shelf” solution to non-receipted returns • Focused on external return fraud • Store has the ability to override a “denial” • Password is required • Activity is tracked at all levels • Historical view available based on consumer ID number • Reporting is dynamic and offers both a real-time and YTD view AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

FIELD LP TEAM TOOLS – PREDICTIVE MODELING • Predictive Modeling allows the Field LP Team to “intercept” a high trending store before a high shrink result is booked • Based on Item Adjustments and Uncredited Returns since store’s last inventory • Offers a 50 bps range to help identify the underperforming stores (i. e. 0. 75% - 1. 25%) • Identifies stores that are trending over company threshold which drives the Field LP Team to initiate a Predictive Model Assessment • Series of audits, thorough review of exception reports, general loss interviews, etc. AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

CASE REVIEW • Based on the Predictive Model, it was found that a specific store was trending abnormally high; completely out of character for this store • The usual suspects of reporting were researched and multiple employees were identified as having a high dollar amount of “Express Part” Returns (special order items) over the course of several weeks; also out of character for this store AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

CASE REVIEW • Our PCI signature report gives the Field LP Team the ability to review signatures signed to the electronic debit pad • The majority of the returns in question had a legitimate name and phone number, however, the signature was an indicator of a potential problem • Additional signatures were reviewed and it was identified that the Auto. Zoner in question had actually slipped up and generated a return in his own name (not visible to protect identity) AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

• The Managers Next Day Review for the returns in question were worked per policy • Unfortunately, this store did not have a Store Manager and a 2 nd level manager, or PSM, worked the process (or so it appeared) • Digital video revealed that when the Managers Next Day Review was being worked, it was being worked by the person processing the questionable refunds (Wolf watching the hen house) using another managers password AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

CASE REVIEW • The returns in question drove us to review the electronic journal on specific business dates. • There was a total of 5 MSD Ignition Controllers returned, and a total of 2 Reman Cylinder Heads; all fraudulently • The funds from the refunds were placed on to a merch card (store credit) AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

CASE REVIEW • After making note of the multiple returns where the return tender was a merch card, we tracked the transactions where the dollars were redeemed • There was a total of eight merch cards activated within the same week • Same store, typically within the same day, and the transaction was processed by the same person who generated the refund • For the transaction shown, the parts were warrantied to the Auto. Zoner who generated the returns, processed the redemption transactions, and he even used his own Auto. Zoner discount card. • Despite the reporting telling us multiple people were involved, the details in this transaction proved different AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

CASE REVIEW • At the conclusion of the investigation we conducted an interview with the Auto. Zoner in question • $1, 900 admission • Admission of using another managers confidential password to generate fraudulent refunds • Admission of forging customer signatures on the debit pad • Admission of using a 2 nd managers password to work the Managers Next Day Review • Admission of taking unpaid Auto. Zone merchandise from the store • Auto. Zoner was terminated and prosecuted • 2 additional Auto. Zoners were terminated for giving out their management password • Close the gap! AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

MORAL OF THE STORY • Returns Fraud cases can be very simple or they can be very complex • The case example shown was complex and it involved a lot of moving pieces from an inventory perspective, a POS perspective, and it appeared that there were multiple people involved. • Don’t limit yourself to one report or just one system for the reporting • For instance, this case required us to use five different sources to complete the investigation • Predictive Model • POS • Proact • KPI Reporting • Digital Video • Rocket Shuttle (details for merch card usage) • Step out of your comfort zone and trust the data / follow it where it takes you • Don’t hesitate to question the data when necessary; it helps boost your confidence as an investigator • Reverse Engineer every investigation that challenged you or was out of the norm vs a routine investigation • Develop a method to identify the behavior faster next time AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED

CLOSING Q&A AUTOZONE INC. COPYRIGHT 2016. ALL RIGHTS RESERVED