MEMORANDUM OPINION AND ORDER
KAREN E. SCHREIER, UNITED STATES DISTRICT JUDGEPlaintiff, Argus Leader Media, brings this Freedom of Information Act (FOIA) suit against defendant, United States Department of Agriculture (USDA). Argus seeks data on the Supplemental Nutrition Assistance Program (SNAP), formerly known as the Food Stamp Program. The USDA opposes releasing the data based on FOIA Exemption 4, arguing that such disclosure would cause substantial competitive harm to grocery stores participating in SNAP. This court disagrees and holds that disclosure of the requested data will not cause substantial competitive harm to SNAP retailers. The data should be disclosed under FOIA.
BACKGROUND
The USDA administers SNAP through the Food and Nutrition Service, an agency within the USDA. The purpose of SNAP is to give children and needy families access to food and nutrition education. When SNAP households redeem their benefits, the transaction looks like a customer using a debit card. The SNAP household pays for its groceries by swiping a benefits card and entering a PIN number. A third-party processor verifies that the SNAP account has available benefits and then approves or denies the transaction. If the transaction is approved, the third-party processor then transfers money from the SNAP household’s account to the retailer’s bank and sends the redemption data to the Food and Nutrition Service.
Argus, in 2011, made a FOIA request to the Food and Nutrition Service. Argus sought a variety of SNAP data—including yearly spending totals at individual retail locations. About two weeks later, the Food and Nutrition Service provided some of the requested information and withheld the remainder citing FOIA Exemptions 3 and 4. Argus then filed an administrative appeal. Before the USDA formally denied the appeal, Argus filed this action.
The USDA, in 2012, filed its first motion for summary judgment. This court granted the USDA’s motion and held that Exemption 3 of FOIA applied to the undisclosed data. The Eighth Circuit Court of Appeals *830reversed and remanded the case. In 2015, the USDA filed its second motion for summary judgment, arguing FOIA Exemptions 4 and 6 applied to the requested data. This court denied the motion and scheduled the case for a bench trial. Before trial, the USDA withdrew its Exemption 6 argument, and the parties stipulated that .the only issue remaining for the court to decide was whether Exemption 4 applied to yearly SNAP revenues for individual stores.
The bench trial began on May 24, 2016. A number of Food and Nutrition Service employees testified about the collection of SNAP data. Witnesses also testified about the potential harm in disclosing the requested data, including the USDA’s witness Joey Hays. Hays is the President and Owner of Dyer Foods Incorporated, a supermarket chain that started in Dyer, Tennessee and has expanded to 13 locations. Hays has spent 35 years in the grocery business. Hays testified that individual store SNAP sales data is not public information and that release of the data would cause competitive harm to his business because competitors could use the information against him. On cross examination, however, Hays admitted that releasing the SNAP data would not give competitors a store’s total profits. Hays also admitted that much of a store’s business is already visible to the public such as product selection and price. Hays further testified that Wal-Mart has already saturated his market, even without the requested SNAP information.
Andrew Johnstone, Associate General Counsel for Sears Holdings Management Corporation, also testified for the USDA. Johnstone echoed Hays’s comments that the grocery business is especially competitive because the profit margins are low. Johnstone testified that if the requested SNAP data was disclosed it would help competitors take away business from Kmart stores. Finally, Johnstone noted the potential stigma that might result from publishing SNAP data. Specifically, John-stone was concerned that landlords would not renew their lease agreements if the data showed that KMart stores had high SNAP sales. On cross examination, John-stone testified that store data is already publically available to market researchers. Such data includes a store’s location, products, and pricing. Johnstone also testified that if the SNAP information was released, it would be released regarding all SNAP retailers.
Peter Larkin, President and CEO of the National Grocers Association, testified that profit margins in the grocery industry are roughly 1% before tax. He also testified that a 2014 study found profit margins to be $0.0091 on every dollar. Thus, grocery stores must have a high sales volume to make a profit. Larkin also testified that individual store SNAP data is not available currently to the public, and Larkin reasoned that injecting new sales information into the public domain could impact stores because competitors could target high dollar SNAP locations and build new stores in that area. On cross examination, Larkin admitted that a number of factors play a role in a customer’s decision to shop at a grocery store such as better produce, convenient location, unique products, or better customer service. Larkin also testified that disclosing SNAP data would not be the same as disclosing a store’s profits or net sales.
Gwen Forman, Senior Vice President of Marketing at Cumberland Farms, theorized that SNAP data is valuable because it confirms whether or not a store’s practices are successful. Although the public sees a store’s advertising and customer outreach, a competitor cannot confirm whether the store’s strategy is effective. *831Releasing SNAP data—Forman argues— allows competitors to gauge whether a store’s marketing strategy is effective. On cross examination, Forman admitted that a number of other factors also determine whether a store is ultimately successful at a given location. SNAP data alone will not determine a store’s future plans or business strategy.
Argus’s first witness was Dr. Richard Volpe, an Assistant Professor in the Agribusiness Department at California Polytechnic State University. Dr. Volpe testified that a variety of store data is already public. For example, a store’s prices, level of activity, layout, and assortment of products are visible to anyone visiting the store. This data is already being collected and is available to retailers. Dr. Volpe also addressed concerns about benchmarking, the practice of analyzing a store’s sales over a period of years. Dr. Volpe explained a benchmarking analysis of SNAP data has limited value because a store’s increased SNAP revenue may be attributable to a number of factors such as increased prices, change in customer demographics, or increased number of SNAP customers. Because additional data is necessary to determine the reason for increased SNAP sales, the release of individual store SNAP data would not cause competitive harm.
Argus’s next witness was Dr. Ryan Sougstad, Associate Professor of Business Administration at Augustana University. Dr. Sougstad testified about how companies use data to come to decisions. Dr. Sougstad testified that individual retailer SNAP data would not likely play a significant role in helping businesses decide where to locate stores. On cross examination, Dr. Sougstad admitted that he was unable to determine to what degree stores would be less profitable if the requested SNAP data was released, but he believed any economic harm would be marginal.
The USDA’s one rebuttal witness, Bruce Kondracki, Vice President of Market Insights and Consumer Research at Dakota Worldwide Corporation, testified about market analysis in the food industry. Kon-dracki explained that grocery stores use model forecasts to determine where to add locations and that releasing SNAP data could improve the accuracy of these models. Kondracki also testified that the addition of new stores does not necessarily mean customers quit frequenting their current store, but customers may spend less money at their current store.
LEGAL STANDARD
“‘Congress intended FOIA to permit access to official information long shielded unnecessarily from public view.’ ” Hulstein v. Drug Enf't Admin., 671 F.3d 690, 694 (8th Cir. 2012) (quoting Milner v. Dep’t of Navy, 562 U.S. 562, 565, 131 S.Ct. 1259, 179 L.Ed.2d 268 (2011)). “FOIA generally mandates broad disclosure of government records.” Cent. Platte Nat. Res. Dist. v. U.S. Dep’t of Agric., 643 F.3d 1142, 1146 (8th Cir. 2011) (citations omitted). FOIA requires that an agency offer records upon request unless they are the sort of records protected by one of the nine exemptions under the Act. Milner, 562 U.S. at 565, 131 S.Ct. 1259. The exemptions “are to be narrowly construed to ensure that disclosure, rather than secrecy, remains the primary objective of the Act.” Mo. Coal. for Env’t Found. v. U.S. Army Corps of Eng’rs, 542 F.3d 1204, 1208 (8th Cir. 2008) (citations omitted). The district court engages in a de novo review of an agency’s decision to deny a request for information under FOIA, and the burden is upon the agency to show that the specific exemption applies. 5 U.S.C. § 552(a)(4)(B); In re Dep’t of Justice, 999 *832F.2d 1302, 1305 (8th Cir. 1993). “The government bears the burden of proving by a preponderance of the evidence that a withheld document falls within one of the exemptions.” Enviro Tech Int’l, Inc. v. U.S. EPA, 371 F.3d 370, 374 (7th Cir. 2004) (citing § 552(a)(4)(B)). This court has made all of its factual determinations by a preponderance of the evidence.
DISCUSSION
The exemption at issue here is FOIA Exemption 4. The Eighth Circuit Court of Appeals has explained that “[t]he plain language of [Exemption 4] exempts only (1) trade secrets and (2) information which is (a) commercial or financial, (b) obtained from a person, and (c) privileged or confidential.” Brockway v. Dep’t of Air Force, 518 F.2d 1184, 1188 (8th Cir. 1975) (internal quotation omitted). Neither party has argued that the requested SNAP data is a trade secret, so the issue before the court is whether the requested FOIA information is (1) commercial or financial; (2) obtained from a person; and (3) privileged or confidential. The parties have stipulated that the information is commercial or financial, so only the remaining two elements are discussed below.
A. The SNAP information is obtained from a person.
The Supreme Court has explained that information is “obtained from a person” if the “information [is] obtained outside the Government.” Fed. Open Market Comm. of Fed. Reserve Sys. v. Merrill, 443 U.S. 340, 360, 99 S.Ct. 2800, 61 L.Ed.2d 587 (1979). In FCC v. AT & T, 562 U.S. 397, 408-09, 131 S.Ct. 1177, 179 L.Ed.2d 132 (2011), the Supreme Court held a corporation was a person under Exemption 4 analysis. Argus asserts that because SNAP is a government program, the requested SNAP data is obtained from inside the government. The government is giving SNAP benefits to qualifying households, and the government then tracks where the SNAP households are spending their benefits. Argus contends that the government is essentially keeping track of its own spending. Thus, Argus’s position is that all of the redemption data is generated and collected by the government and that the SNAP data is obtained from the government.
The Eighth Circuit Court of Appeals has already held that Exemption 3 does not apply to this case. In that opinion, the Eighth Circuit held that the requested information is “‘obtained’ from third-party payment processors, not from individual retailers.” Docket 44 at 6 (citations omitted). This conclusion is supported by the testimony from the Food and Nutrition Service employees. Neither party disputes that it is the third-party processor who verifies whether the SNAP household has available SNAP benefits and that the third-party processor submits the redemption data to the Food and Nutrition Service. Based on the Eighth Circuit’s ruling and the testimony at trial, this court finds that the requested information is obtained from a person, namely the third-party processors who facilitate the SNAP transactions.
B. The SNAP information is not privileged or confidential.
Information is confidential if “disclosure of the information is likely to have either of the following effects: (1) to impair the Government’s ability to obtain necessary information in the future; or (2) to cause substantial harm to the competitive position of the person from whom the information was obtained.” Contract Freighters, Inc. v. Sec’y of U.S. Dep’t of Transp., 260 F.3d 858, 861 (8th Cir. 2001) (quoting Nat’l Parks & Conservation *833Ass’n v. Morton, 498 F.2d 765, 770 (D.C. Cir. 1974)).1 This test, which the Eighth Circuit Court of Appeals adopted from the District of Columbia Court of Appeals, is commonly known as the National Parks test and “has been widely recognized and applied by the circuit courts when construing Exemption 4.” Id. Because the parties agree that prong 1 of the National Parks test is inapplicable, only prong 2 is addressed. Docket 61 at 19-20.
The Eighth Circuit Court of Appeals has not articulated the showing necessary for a party to prove that release of information would cause substantial harm under prong 2. But the District of Columbia Circuit has found that prong 2 competitive harm may be established if there is evidence of “actual competition and the likelihood of substantial competitive injury .... ” Pub. Citizen Health Research Grp. v. Food & Drug Admin., 704 F.2d 1280, 1291 (D.C. Cir. 1983) (internal quotation omitted).2 Competitive harm is limited to “harm flowing from the use of proprietary information by competitors. Competitive harm should not be taken to mean simply any injury to competitive position, as might flow from customer or employee disgruntlement or from the embarrassing publicity attendant upon public revelations ....” Id. n. 30. When assessing the potential for competitive harm, a court may consider the nature of the material sought, the competitive circumstances surrounding the disclosure, and credible opinion testimony. Nat’l Parks & Conservation Ass’n v. Kleppe, 547 F.2d 673, 683 (D.C. Cir. 1976). Although a party opposing disclosure “need not ‘show actual competitive harm,’ ” conclusory and generalized allegations—standing alone—are not sufficient. Food & Drug Admin., 704 F.2d at 1291.
1. Actual competition
Competition in the grocery business is fierce. This conclusion is supported by the testimony of Joey Hays, Andrew Johnstone, Peter Larkin, Gwen Forman, and Bruce Kondracki. Johnstone noted that competition in the grocery business has increased with the entrance of new competitors, and Larkin provided testimony that the profit margins in the grocery industry are at $0.0091 on every dollar spent. Kondracki also explained competition is not measured solely in terms of lost customers, but in lost dollars to other stores. Kondracki testified the entrance of a new store into an existing market can cause a significant loss in business. Based on this testimony, the court finds that the grocery industry has actual competition.
2. Likelihood of substantial competitive harm
The competitive harms alleged by the USDA fall into two main categories: (1) harms arising from competitors using SNAP data to lure away customers from other businesses and (2) harms arising from the potential stigma associated with being a high volume SNAP retailer. For-man’s testimony encapsulated the retailers’ overarching concerns of the former. Forman explained how competitors could *834use SNAP data to choose the locations of new stores, evaluate an existing store’s overall success, and ultimately cut into another store’s profits. Hays, Johnstone, Larkin, and Kondracki all gave testimony supporting this conclusion that disclosure of individual store redemption data could cause competitive harm because competitors in the grocery business could use the information to target an existing store’s customers.
This analysis, however, is incomplete. Competitors in the grocery industry already use a variety of publicly available information to make decisions. This information includes a store’s location, layout, pricing, product selection, and customer traffic. Dr. Volpe and Dr. Sougstad both noted that while SNAP information may provide some insight into a store’s overall financial health, the data is a small piece in a much larger picture—disclosure would have a nominal effect on competition in the grocery industry. Kondracki’s models of consumer behavior appear to support this point. Kondracki testified that the current market models can reach correlations of .9 or .99. This appears to indicate that while SNAP data may be beneficial, it would not add significant insights into the grocery industry. This conclusion is further supported by the testimony of Hays, Hays testified that competitors such as Wal Mart have already saturated the market where he competes. Wal Mart took these actions without the requested SNAP data. This court concludes that any potential competitive harm from the release of the requested SNAP data is speculative at best.
The second concern the USDA voiced to releasing SNAP data was the potential stigma SNAP households and SNAP retailers might face. As noted above, this type of harm is not relevant in an Exemption 4 analysis because it is not a harm caused by a competitor. Even if stigma was relevant, the USDA’s evidence on potential stigma was not sufficient to meet its burden. Although Johnstone testified that high SNAP sales revenue might affect a landlord’s decision to rent its commercial space to a retailer, it seems unlikely that a landlord would be unaware of its tenant’s customer base. Johnstone also did not provide any evidence of the likelihood of this contingency occurring. At best, John-stone’s claims are speculative. Furthermore, the remaining witnesses did not explain how high or low SNAP sales would harm them stores. For example, although a high volume of SNAP sales might encourage a competitor to enter that geographical market, an equally compelling conclusion is that the competitor may decide to stay away from that market. Another equally compelling conclusion is that SNAP sales will have no or little effect on a store’s decision to expand into new sites. This is because a variety of factors influence a store’s decision to open a new location including: cost of real estate, location of real estate, the business’s long-term financial plan and goals, and other factors. This court finds that the competitive harms associated with stigma are also speculative.
The USDA, in its post-trial reply brief, cited three cases to support its claim that releasing the requested SNAP data would cause competitive harm. Each case, however, is distinguishable from the present litigation because the plaintiffs in the other cases asked for data that would give greater insights into the company’s workings. In Kleppe, 547 F.2d at 676-77, plaintiffs sought information about company assets, liabilities, net worth, balance sheet information, future and existing projects, and operating capacity. In Public Citizen Health Research Group v. National Institutes of Health, 209 F.Supp.2d 37, 40-41 (D.D.C.2002), plaintiffs requested the ne*835gotiated royalty rates between private researchers and the government. Finally, in Sharkey v. Food & Drug Administration, 250 Fed.Appx. 284, 288-290 (11th Cir. 2007), plaintiffs sought information that would result in the disclosure of domestic market share and sales volume. Here, the requested SNAP data does not provide the same insights into store profitability. SNAP sales are merely a part of the store’s total revenue. SNAP data does not disclose a store’s profit margins, net income, or net worth. SNAP data also does not disclose how a company bids on government contracts or negotiates with the federal government. In essence, SNAP data is merely a bill from the retailer to the government. As the USDA acknowledges, this type of data is regularly disclosed, and disclosure is consistent with FOIA’s underlying purpose. Docket 125 at 5-6. Because of the speculative nature of the USDA’s claims and FOIA’s preference for transparency and disclosure, this court finds that the release of SNAP data will not likely cause substantial competitive harm to SNAP stores. The data should be disclosed.
CONCLUSION
The USDA has failed to meet its burden to show that Argus’s FOIA request falls within Exemption 4 because the USDA did not prove that release of the requested data was confidential. Specifically, USDA did not show that release of the requested information would cause substantial competitive harm if it was disclosed. Thus, it is
ORDERED that judgment will be entered in favor of Argus and against USDA in accordance with this memorandum opinion and order.
. Another test is used if the person or entity submitting information is acting voluntarily. That test is inapplicable here, however, because SNAP retailers are required to disclose EBT data if they want to be compensated. See Martin Marietta Corp. v. Dalton, 974 F.Supp. 37 (D.D.C. 1997) (holding a bid to do government work is not voluntary under Exemption 4 because the bid must be submitted in order to win the contract); see also Defendant’s Memorandum in Support of Motion for Summary Judgment, Docket 61 at 19 (stating *‘[T]he agency’s position is that the information here is required to be submitted.”)
. See also Sharkey v. Food & Drug Admin., 250 Fed.Appx. 284, 288 (11th Cir. 2007).