Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 1 of 12 UNITED STATES DISTRICT COURT EASTERN DISTRICT OF ARKANSAS (LITTLE ROCK) NATIONAL TRUCKING FINANCIAL RECLAMATION SERVICES, LLC, BRUCE TAYLOR, EDIS TRUCKING, INC., JERRY FLOYD, MIKE CAMPBELL, PAUL OTTO, TOWNES TRUCKING, INC. and R&R TRANSPORTATION, INC., individually, and on behalf of all others similarly situated, Plaintiffs, vs. PILOT CORPORATION, PILOT TRAVEL CENTERS, LLC d/b/a PILOT FLYING J, FJ MANAGEMENT, INC., CVC CAPITAL PARTNERS, JAMES A. "JIMMY" HASLAM, Ill, MARK HAZELWOOD, MITCH STEENROD, SCOTT WOMBOLD, JOHN FREEMAN, VINCENT GRECO and BRIAN MOSHER, ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) Case No. 4:13-cv-00250-JMM Defendants. REPORT OF THE INDEPENDENT ACCOUNTANT. HORNE LLP Introduction As provided in the Revised Settlement Agreement dated July 23, 2013 (the "Agreement"), and approved by the Court in the Amended Class Settlement Preliminary Approval Order dated July 24, 2013, HORNE LLP ("HORNE") was appointed Independent Accountant and assigned certain responsibilities under the terms of the Agreement. These responsibilities are set forth in Sections 34 through 40 of the Agreement. This report to the Court is required by Section 38 and addresses the accuracy of the Defendants' Internal Auditors' ("DIA") calculation of the Settlement Payments made to Class Members and other information related to our work. Specifically, we tested DIA's calculations to obtain assurance to a 95% confidence level Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 2 of 12 that there was a 5% or lower error rate in the amounts that were calculated by DIA. The Agreement allowed a 5% precision rate. Our testing work began on August 20, 2013, and concluded on May 15, 2014. The project required approximately 8,375 team member hours to complete. Approximately 18 team members worked over 100 hours each on the project. Our team members were at Pilot Flying J's ("PFJ" or "Pilot") headquarters in Knoxville, Tennessee, for approximately 15 weeks, and worked remotely from HORNE offices for the balance of the project. In our opinion, the process of DIA's calculation of customer payments and the oversight and review function performed by HORNE has worked as intended. Pilot management, DIA, and HORNE have worked through an extremely difficult and tedious process. The time and work required to complete this project have taken Pilot and HORNE significantly longer than either of us initially expected. However, from our perspective, a large portion of the additional time and work was incurred to deal with issues that would have negatively impacted the accuracy of its customer payments had they not been addressed. In our view, Pilot placed a priority on accurate and complete payments to customers. We believe that Pilot has fulfilled its duty and responsibility under the Agreement for the areas that we were charged with reviewing. On May 15, 2014, we completed the required testing and determined that DIA's calculations of Settlement Payments met the accuracy criteria required by the Agreement. Scope of Our Work The scope of our work included: • Review and determination of the adequacy of the processes and procedures used by DIA; -2- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 3 of 12 • Information technology and database integrity testing; • Evaluation of DIA's capability and objectivity; • Determination that DIA properly identified Eligible Class Members that are entitled to compensation; • Inspection on a test basis of documentary evidence of the procedures performed by DIA; and • Testing the results of DIA's procedures and the accuracy of the results of those procedures. The scope of our testing was limited to the components of the Discount Program addressed in the Agreement. For a particular customer, this could include a number of cents off the retail price, a number of cents over or under the diesel cost, and, for some customers, a change in the specific rack or loading facility to be used to set the cost. Overview of the Audit Process The audit period was from January 1, 2005, to July 15, 2013, and covered a total of 7,853 PFJ accounts. Payments to Eligible Class Members totaled $68,703,184.08, which included $9,615,441.50 of interest at 6%. DIA began their audit work in April 21, 2013, and had essentially completed all payments by April 30, 2014. During the peak of their work, they deployed 20 to 25 members of their internal audit staff and contracted with area accounting firms to obtain 20 to 30 additional accountants to work on the audit. Many of the more complex accounts took several days to audit. The classifications of accounts that were part of the Eligible Class Members included four types: • 783 Manual Rebate accounts; • 6, 778 Direct Bill accounts; - 3- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 4 of 12 • 35 Hedge accounts; and • 257 Point of Sale customers. To make the audit possible, at least eight years of data had to be accumulated, which included: • Daily retail diesel fuel prices for over 600 locations; • Daily diesel cost from all loading locations; • Transaction information, including customer, gallons purchased, date purchased, and location; • Discount Program information by customer, with the more complex having frequent changes, and with the Discount Program arrangement varying by store location and/or region; and • Millions of emails, Salesforce entries, pricing documents, and other sources of information that could be used to establish the correct Discount Program arrangement. A high level summary of the DIA audit and customer payment process is stated below: 1. Obtain historical Discount Program pricing information for a customer account; 2. Obtain relevant documents from document database to determine the appropriate Discount Program historical pricing; 3. Correct historical Discount Program pricing in the customer's file; 4. Obtain relevant historical transactional information for the customer account along with pricing information and other relevant variables and process -4- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 5 of 12 through recalculation software to determine what correct pricing should have been; 5. Compare correct pricing discount to the pricing discount previously credited to the customer and determine the difference; 6. Calculate 6% interest on a transactional level basis; 7. Prepare a check to the customer for the additional discount and interest owed; 8. Mail check, along with detail on the corrected pricing information, to the customer; and 9. If no payment was owed to a customer, send a letter to the customer to that effect. HORNE Procedures. Staffing and Timing The following sections describe the procedures, reviews, and testing we performed to satisfy our responsibilities as outlined in the Scope of Our Work section above. Review and determination of the adequacv of the processes and procedures used by DIA On August 20, 2013, two of my partners and I traveled to Knoxville for our initial meeting with key members of DIA's audit team and other executives of PFJ. During this meeting we obtained an understanding of DIA's team structure, the professional background and experience of DIA's key team members (internal or outsourced), the critical information technology systems, the auditing processes, the status of DIA's progress, and the expected completion dates. After this initial meeting, we sent a team of HORNE staff members to the PFJ offices for several weeks to conduct a detailed - 5- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 6 of 12 analysis of the auditing processes by shadowing and interviewing DIA's auditors. Team members from our technology group conducted interviews with the technology professionals of PFJ and Logic-Force Consulting, LLC, an outside contractor retained by PFJ's legal counsel, to understand the data management control structure and how the integrity of the data used in the audit process was ensured. Based on this work, it appeared that, overall, the audit process was functioning as intended. However, there were several areas in which we needed additional clarification. On September 12, 2013, two of my partners and I attended a meeting in Knoxville with key DIA employees and other executives from PFJ. During this meeting, we discussed the results of our initial procedures. During our discussions, PFJ provided explanations to satisfy us concerning the areas that needed to be clarified and agreed to the modification of certain DIA audit procedures. lnfonnation technology and database integrity testing Our work in this area was designed to accomplish two key things. First, we tested to ensure that the data being searched and used in the audit process was complete and accurate and had been obtained from the proper systems and sources. Second, we performed testing on the "recalculation software" 1 to ensure that it was performing as intended. The testing of the IT systems in the database information took hundreds of hours of our IT auditors' time. The complexity of the work was caused by many factors, including legacy IT systems, volume of data, sources of data, and number of data sources. Our testing was designed to ensure that the data source file was the proper file 1 The "recalculation software" was designed by PFJ's IT department and was a software program used to calculate the proper pricing of all the corrected discount program agreements for all accounts ·that had revisions to their pricing agreements. This was an extremely complex program because of the length of time covered, data sources used, and pricing infonnation needed to properly compute what the customer's price should have been. -6- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 7 of 12 to use, that the information from that file was converted into the proper format for use in the audit, and that the conversion was complete. This process had to be performed for each data source. For example, this included customer transaction data, retail prices by day and location, and cost information by day and location. Because of the time period covered, the information source for each type of data could have come from one or more systems, some of which had been out of service for some time. Testing of the recalculation software was equally complex. Because the software was designed to be only as robust as it needed to be, it had to be revised as more complex accounts were encountered and/or errors in the software were discovered. This resulted in many versions of the recalculation software being used. Our primary testing was performed on the last versions of the recalculation software. Our testing of the software confirmed that the recalculation software in its later iterations performed well and returned accurate and appropriate price information for customers. Evaluation of DIA 's capabilitv and objectivitv Our evaluation of DIA's capability and objectivity was a process that continued throughout our work. We also assessed the support and commitment of PFJ's senior management and PFJ's internal audit group to the project. We found the DIA leadership and supporting teams to be capable and well trained. The PFJ senior management dedicated corporate resources and retained significant outsourced accountaQts and other resources to complete the complex audit in as short a time period as possible. There were some inconsistencies in the work product of the staff accountants, which is not unusual for a project of this size. We found that DIA's supervisor work reviews mitigated these issues. - 7- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 8 of 12 Determination that DIA properly identified Eligible Class Members that are entitled to compensation Our process to determine the completeness of the manual rebate customer population was tailored around understanding the steps performed by DIA to extract customers that had received payments from Pilot from the accounts payable system. There was no one indicator or flag in the accounts payable system that allowed DIA to readily determine a payment to be a manual rebate. DIA requested Pilot IT to extract all payments from the general ledger accounts associated with customer rebates and to review invoices with naming conventions that had been identified as invoices used to pay customer rebates. Based on our review of the process, we believe DIA has undertaken reasonable procedures to identify manual rebates paid to customers. DIA utilized a query to extract all customers from its billing system which provided the point of origination for the direct bill population. We obtained, reviewed, and re-performed the query identifying no discrepancies. This population was reduced by certain exclusions such as customer accounts with no gallons, internal Pilot accounts, and certain customers receiving daily pricing files. We gained an understanding of each exclusion type and performed sampling to ensure completeness of the direct bill population. Because of the size and nature of the hedging and point of sale accounts, these were easily identified and were fully included in the population. Inspect on a test basis the documentary evidence of the procedures performed by DIA: test the results of DIA 's procedures and the accuracy of the results of those procedures Our testing of the documentary evidence of the procedures performed by DIA and the results and accuracy of those procedures were performed during our statistical and other testing of DIA's work. Our work process, which is outlined below, is essentially -8- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 9 of 12 the same work process that was used by DIA in their audit process. Consequently, we re-performed 100% of the work done by DIA for the accounts we selected in our sample. Our team members received the same training that DIA staff received, we were able to access copies of the same databases that were tested by us and other data sources utilized by DIA staff, and we shadowed their staff members as they performed their initial testing. This and other information allowed us to perform the same audit work for our sample testing that DIA had performed for their work. For each account in our sample, we obtained an Excel spreadsheet that contained the historical pricing information from Pilot's systems. The pricing information contained in these spreadsheets was the historical information Pilot originally used to price diesel purchases for their customers. To determine whether that information was what was actually agreed to with the customer or contained incorrect information, we electronically searched millions of historical documents from Pilot's records, as discussed above in the Overview of the Audit Process section. By using various search techniques, we identified the documents that related to that customer. Those documents were then individually reviewed to identify documents that contained Discount Program information. 2 We established the time period that each agreement was in effect and the Pilot locations covered by the agreement. This was recorded in an Excel spreadsheet for each account. It was normal for agreements to change over time and those changes had to be tracked, sometimes on a store-by-store basis. This information was compared to the information in the historical spreadsheets and any differences were flagged and adjusted. For each account sampled , we compared our results to those of DIA and identified any differences. 2 If there was a conflict in the information, the pricing information that provided the best price to the customer was used. -9- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 10 of 12 Differences between our results and those obtained by DIA were not considered errors at this point. Because of the tremendous body of information and minor variances as to how it was interpreted, many small differences arose. ', Once the historical Discount Program information had been checked through the process outlined above, accounts were classified in one of two ways. Accounts that had no changes to their pricing information did not require recalculation. Accounts that had changes to their pricing information did require recalculation and were recalculated through the use of the recalculation software described previously. The use of the recalculation software was quite labor-intensive. It required significant care and review to ensure that the data being input into the recalculation software was accurate and complete. This was typically done in two-person teams to ensure accuracy. For purposes of our testing, we defined an error as a difference between the HORNE calculated amount and the DIA calculated amount that was the greater of $2,000 or 5% of the correctly calculated amount. Only underpayments were counted as errors because the overpayments that we identified were generally caused by using a previous version of the recalculation software. As the tool was updated and customer accounts were reprocessed, it was discovered that Pilot had overpaid a number of customers. Since no attempt was made to recover any overpayments from customers, we did not consider this to be an error. Our testing also included confirming the interest calculation amount and the mailing of the check to the customer. The following table shows the type of customer, the population related to that customer (expressed as the number of accounts), the sample we tested for that customer type, the errors allowed, and the errors incurred. - 10- Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 11 of 12 Population Statistical Sample Errors 6,778 179 43 Manual Rebate 783 149 3 Point of Sale 257 N/A 0 35 N/A 0 Type Direct Bill Hedging Some customers have more than one account. Our work and sample sizes and sample selections were based at the account level. Therefore, for certain customers we may have selected only one of two or more accounts that the customer had. As can be seen, the large majority of the accounts were direct bill or manual rebate accounts. For the hedging and point of sale accounts, our testing was not strictly statistical, but did include sufficient sample testing and a review and approval of the process used by DIA. In our opinion, this was adequate to ensure that the accuracy thresholds, as spelled out in the Agreement, were met as a result of our testing. Conclusion HORNE hereby certifies that it has reviewed the work performed by Defendants' Internal Auditors consistent with the requirements of Paragraphs 35 through 37 of the Agreement. Further, HORNE hereby confirms, to a reasonable degree of certainty, that the work performed by DIA (1) properly identifies the Eligible Class Members that are entitled to compensation, and (2) accurately quantifies the amount of compensation due to the Eligible Class Members under the Agreement in accordance with the required 3 The errors found in our statistical sampling did not exceed the allowable level. Therefore, statistically we determined with a 95% confidence level, with a 5% precision, that the error rate did not exceed 5%. - 11 - Case 4:13-cv-00250-JLH Document 48-1 Filed 06/13/14 Page 12 of 12 sampling method. More specifically, it is HORNE's opinion that the work of DIA is accurate according to the criteria set forth in Paragraph 37 of the Agreement. Respectfully Submitted, jk~~~~ Partner HORNE LLP 1020 Highland Colony Parkway., Suite 400 Ridgeland, MS 39157 - 12 -