Business for Profit

Appendix G

Data Request: Retailer Summary Transaction Data Files

memorandum

Date	Month day, 2019
To	Recipient
From	Robin Koralek, Project Director Chris Logan, Principal Investigator Dr. Eric Williams, Social Science Analyst/Project Officer (USDA/FNS)
Subject	SNAP Online Pilot Evaluation: Retailer Data Request

Welcome to the SNAP Online Pilot Evaluation. As you know, the United States Department of Agriculture (USDA) Food and Nutrition Service (FNS) is seeking to learn how the Electronic Benefits Transfer Online Purchasing Pilot is implemented across retailers and States. The evaluation will inform FNS’ decisions about whether and how to make SNAP online purchases more widely available, how to ensure the accessibility and quality of online purchases, and how to ensure that protections against abuse remain strong or grow stronger. This document provides information about the data we will collect from you.

Abt Associates, and its partners Altarum and Novo Dia Group, are conducting this evaluation the Online Purchasing Pilot, which is authorized under the Agricultural Act of 2014¹ (The Farm Bill). FNS is sponsoring this data collection.

Overview of the Evaluation

This evaluation will help FNS learn how the pilot operates, the implementation challenges and lessons learned, the characteristics of SNAP online customers, the risks and benefits of online purchasing for the integrity of SNAP, and the requirements for expansion. The evaluation team will examine:

1. Pilot implementation, including SNAP online transaction approaches; the process, challenges, and lessons of implementation; the characteristics of SNAP households that shop online; and the level of effort for stakeholders (e.g., retailers, states, EBT processors).

2. Pilot integrity, including delivery patterns and their relationship to customer addresses and retailer locations; customer profiles and their relationship to EBT cards and SNAP households; customer shopping patterns; and problems such as refunds and cart abandonment.

After pilot operations begin, we will be in touch about setting up conversations about the pilot design, testing, and implementation. For now, we will focus on data requests to ensure you are able to provide the data elements we need for the evaluation.

Data Requests

We will request the following types of data for the evaluation:

1. Optional transaction data including information about individual transactions including order number, store, amount charged, payment method, number of items, fulfillment method, delivery address, quantity ordered and fulfilled, and timing of fulfillment. These data will include all purchases associated with orders made by SNAP customers and any refunds associated with these orders.

2. Summary data including information about orders and payment methods, customers and customer profiles, deliveries and shipping, fees, refunds, and customer service calls.

3. Retailer store data including retailer’s store numbers, FNS retailer numbers, store names and addresses. Retailer store lists are only applicable to retailers with multiple stores in the pilot.

Attached are three documents with the data elements and specifications we will request for the transaction data and the summary data. Please be aware that these are not data templates or layouts; they are provided as a basis of information and discussion.

Attachment 1 – SNAP online transaction data elements.xlsx

Attachment 2 – SNAP online summary data elements.docx

Attachment 3 – SNAP online retailer store data elements.docx

Data Use Agreement

We will work with you to develop a data use agreement (DUA). The DUA will authorize recipients (i.e., the Abt team) to use the data you provide for a designated purpose (i.e., this evaluation). Terms of the DUA will include: whom the data may be released to, how the data is to be handled and protected, data destruction or return requirements when the work is done or period of performance is over, procedures in the event of unauthorized disclosure, name of custodian or security officer, legal terms (such as termination, audit rights, liability, indemnification, liquidated damages, civil and criminal penalties), and flow down requirements for business partners. We understand you may already have a standard DUA which we can work with.

Data Security

Protecting the data we receive from you is our highest priority. Abt Associates Inc. Abt complies with the Privacy Act of 1974, Health Insurance Portability and Accountability Act of 1996 (HIPAA), and the E-Government Act of 2002, including Title III: Federal Information Security Management Act (FISMA), which covers site security, security control documentation, access control, change management, incident response, and risk management.

Abt has a dedicated Analytic Computing Environment (ACE3) for the purpose of only storing and analyzing sensitive information such as Personally Identifiable Information (PII) and Protected Health Information (PHI). This environment allows our team to access and work with data using tools such as SAS and STATA without the data leaving the environment. ACE3 is FISMA Moderate (SP800-53 rev.4) and FIPS 140-2 compliant and built with a combination of FedRAMP and Abt-developed systems. PII will be stored and accessed only in this secure environment. Approval must be granted for access to this environment and its project folders. Authentication to ACE3 is via Active Directory and requires the use of DUO multi-factor authentication. Users connect to the system via remote desktop sessions (RDS) directly to the environment.

Transferring Data to Abt

Abt uses a secure data transfer application, Huddle, for the secure, encrypted transmission of sensitive information such as Personally Identifiable Information (PII) and Protected Health Information (PHI), to and from its external clients. Huddle is a FedRAMP collaboration cloud service tool that stores data in the cloud and meets U.S. federal security requirements for encryption. The data is encrypted in-transit through the use of TLS, and at rest with 256-bit AES, and is FIPS 140-2 compliant. Only authorized study team members have access to the project folders in Huddle, and access is restricted and controlled via folder permissions in Huddle. Details about accessing this site will be sent separately.

Evaluation Contacts

If you have any questions, please contact:

Robin Koralek (Project Director) at [email protected] or (301) 347-5613

Chris Logan (Principal Investigator) at [email protected] or (617) 349-2821

Eric Williams (FNS Contracting Officer’s Representative) at [email protected] or (703) 305-2640

Next Steps

After you review this letter we would like to meet to discuss the data requests and the DUA, and to answer any questions you may have. In particular, we want to know if there are any data elements you will not be able to provide or have concerns about.

We expect it will take you 12 hours (one hour per month) to prepare and submit the summary transaction files.

Public reporting burden for this collection of information is estimated to take one hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB control number. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to: U.S. Department of Agriculture, Food and Nutrition Services, Office of Policy Support, 3101 Park Center Drive, Room 1014, Alexandria, VA 22302, ATTN: PRA (0584-xxxx*). Do not return the completed form to this address.

SNAP Online Purchasing Evaluation: Summary Transaction Data Files Request

The following list is based on the Request for Volunteers. It has been amended and clarified. All counts are intended to be provide monthly in Excel or similar format suitable for import into a database for the first 12 months of the pilot.

Counts should be broken out by State, based on the customer’s State SNAP agency.

Multiple files may be provided if this facilitates reporting.

1. Total number of orders placed online with EBT cards that are partly or completely fulfilled (do not count orders that are completely cancelled or that do not include a payment charged to an EBT card), by combination of payment methods used.

• SNAP only

• Cash EBT only

• SNAP and cash EBT combination

• SNAP and any other card* combination

• Cash EBT and any other card* combination

• SNAP, cash EBT and any other card* combination

• Total, all payment methods

* Any other card includes credit, signature or PIN debit, gift cards, PayPal, and other non-EBT payment methods. If possible, gift cards should be treated as a separate payment method in the data on individual transactions.

Each order that is fulfilled – picked up or delivered – counts once.

2. Number of orders placed online with EBT cards that were completely cancelled -- that is, nothing was dispensed for the EBT portion of the order-- broken down by the reason for cancellation.

(Note: This definition would count an EBT order as “cancelled” where all EBT purchases were cancelled but other purchases were fulfilled.)

• Orders cancelled by customer before they are picked

• Orders cancelled by customer after they are picked

• Orders that were picked and not picked up, canceled by retailer (if applicable)

• Orders that were picked and sent for delivery but not delivered, canceled by retailer (if applicable)

• Total cancelled orders

3. Number of customers making an online SNAP EBT purchase by prior online purchasing history:

• Made online grocery purchase from the retailer in the prior 12 months* using any payment method

• No online grocery purchase from the retailer in the prior 12 months* using any payment method

*Any order in last 12 months including same date makes a customer a repeat customer.

4. This item is applicable only if creating a customer profile is not required for online SNAP purchases.

• Number of EBT customer profiles created during the month

• Total number of EBT customer profiles

• Number of EBT customer profiles by ZIP code²

• Number of EBT orders placed by guests (no profile)

5. Number of online EBT orders not picked up (These counts are only needed if EBT customers can pick up orders at locations other than a store operated by the retailer.)

• Online EBT orders for pickup in store operated by the retailer that were not picked up

• Online EBT orders for pickup in alternate location (locker etc., if applicable) that were not picked up

6. Number of SNAP orders fulfilled by each store in pilot (if that model used)
   This item requires a list of stores (including store name, retailer’s store number, and State) with the count of SNAP orders fulfilled. If orders may be fulfilled by a combination of stores, then three counts should be provided:

• Orders fulfilled entirely by specified store

• Order mostly fulfilled by specified store, partly by another store

• Order partly fulfilled by specified store, mostly by another store

7. Number of SNAP orders fulfilled by each warehouse location in pilot (if retailer fulfills orders from more than one warehouse)

• Orders fulfilled entirely by warehouse

• Order mostly fulfilled by warehouse, partly by another warehouse

• Order partly fulfilled by warehouse, mostly by another warehouse

8. Delivery fees

• Number of SNAP orders [i.e., orders paid for in part or entirely with SNAP] with delivery and/or pickup fees, by payment method (EBT cash, debit, credit, gift or store card, other).

• Total dollar amount of fees charged for SNAP orders by payment method.

9. Number of refunds for SNAP purchases by type of refund

• Number of full SNAP refunds (full SNAP amount refunded)

• Number of partial SNAP refunds (partial SNAP amount refunded)

• Number of SNAP orders with fees refunded

• Number of SNAP orders with refund for which fees were charged but not refunded

We also need to get clarification of the retailer’s policy regarding when fees are charged for SNAP orders and when they are or are not refunded.

10. Number of refunds for orders placed with EBT cards – counts of SNAP refunds and total orders with refunds by reason

• Overestimate of weight

• Out of stock item

• Order cancellation

• Dissatisfaction with product(s)

• Excessive time to complete order

• Non-receipt of order

• Other

11. Number of customer service calls about SNAP EBT orders by reason for call and total unique customer calls about SNAP EBT orders. If there is more than one reason of the call, it would be counted for each of the reasons.

Example: Order Late and Damage Goods counts as 2 calls, 1 for each reason.

Note: additional summary data may be requested if one or more requested transaction data elements are not available (times for orders, fulfillment and delivery/pickup, incidence of product returns, etc.)

1 Agricultural Act of 2014, Conference Report to Accompany H.R. 2642. P. 144-147.

2 Number of EBT customer profiles by ZIP code is not required if EBT customer billing addresses are provided in transaction data or in separate monthly data file.