Mallet and Company Inc v. Ada Lacayo

• Court: Court of Appeals for the Third Circuit
• Decided: October 15, 2021
• Docket: 20-3584
• Status: Published

Opinion

PRECEDENTIAL

UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT _____________

Nos. 20-3584, 21-1028, and 21-1029 _____________

MALLET AND COMPANY INC.

v.

ADA LACAYO; RUSSELL T. BUNDY ASSOCIATES, INC. d/b/a Bundy Baking Solutions; SYNOVA LLC; WILLIAM CHICK BOWERS

Russell T. Bundy Associates, Inc. d/b/a Bundy Baking Solutions; Synova LLC, Appellants in No. 20-3584

William Chick Bowers, Appellant in No. 21-1028

Ada Lacayo, Appellant in No. 21-1029 _______________ On Appeal from the United States District Court for the Western District of Pennsylvania (D.C. No. 2-19-cv-1409) District Judge: Hon. Cathy Bissoon _______________

Argued April 16, 2021

Before: JORDAN, GREENAWAY, JR., and SCIRICA, Circuit Judges.

(Filed: September 24, 2021) _______________

Laura C. Bunting Marla N. Presley [ARGUED] Jackson Lewis 1001 Liberty Avenue – Suite 1000 Pittsburgh, PA 15222

Allison G. Folk Jackson Lewis 6100 Oak Tree Boulevard, Suite 400 Cleveland, OH 44131 Counsel for Mallet and Company Inc.

Ada Lacayo 328 Michigan Avenue Lower Burrell, PA 15068 Pro Se

2 Ronald L. Hicks, Jr. [ARGUED] Carolyn B. McGee Porter Wright Morris & Arthur 6 PPG Place – Third Floor Pittsburgh, PA 15222 Counsel for Russell T. Bundy Associates, Inc., d/b/a Bundy Baking Solutions; Synova LLC

Nicholas J. Bell Kathleen J. Goldman Buchanan Ingersoll & Rooney 501 Grant Street – Suite 200 Pittsburgh, PA 15219 Counsel for William Chick Bowers _______________

OPINION OF THE COURT _______________

JORDAN, Circuit Judge.

Behind the breads, cakes, and other treats on our grocery store shelves, there is a ferociously competitive market for baking supplies, and that is the setting for this trade secret and unfair competition case.

In 2019, Mallet and Company Inc. (“Mallet”) learned that Russell T. Bundy Associates, Inc., doing business as Bundy Baking Solutions (“Bundy”), was becoming its newest competitor in the sale of baking release agents. Release agents are lubricants that allow baked goods to readily separate from the containers in which they are made. Bundy was already well-known for other products it offered to the commercial

3 baking industry when it decided to launch a new subsidiary, Synova LLC (“Synova”), to sell baking release agents. Synova hired two of Mallet’s employees, both of whom had substantial access to Mallet’s proprietary information. Taking some of that information with them from Mallet to Synova, they helped Synova rapidly develop, market, and sell release agents to Mallet’s customers. Mallet sued, saying such progress would have taken years to accomplish but for the misappropriation of its trade secrets. Agreeing with Mallet, the District Court issued the preliminary injunction now challenged on appeal, restraining Bundy, Synova, and those employees (collectively, “the Defendants”) from competing with Mallet.

While we appreciate the challenges inherent in disputes involving trade secrets and requests for preliminary relief, the injunction at issue is flawed and must be vacated. For the reasons that follow, we will remand for further consideration of what, if any, equitable relief is warranted and what sum Mallet should be required to post in a bond as “security … proper to pay the costs and damages sustained by any party found to have been wrongfully enjoined or restrained.” Fed. R. Civ. P. 65(c).

I. BACKGROUND

A. Factual Background

1. Mallet and the Defendant Employees

For over eighty years, Mallet has been in the business of developing, manufacturing, and selling baking release

4 agents as well as the equipment used to apply such agents.1 Release agents are applied to commercial baking pans to ensure the consistent release of baked goods over hundreds of uses. They thus play a crucial role in large-scale baking operations. While the ingredients used to create them – mineral oils, vegetable oils, and lecithin – are commonly known, developing a successful release agent is not as simple as knowing a few of its components. There are “a wide range of factors that have to be considered when formulating a release agent,” including product performance, stability, application, cost, availability, and packaging. (J.A. at 10984- 85 (Mallet2 Depo.).) And the efficacy of a release agent can greatly depend on the customer’s product, pan condition, storage conditions, and machinery used to apply the agent. As a result, there are different kinds of release agents, each with unique properties that may be further tailored to maximize performance when used in the production of certain goods. Still, competitors in the release agent market often manufacture and sell identical or similar products.

Mallet proclaims itself “a service business delivering value through the combination of high quality, consistent products and the equipment to apply them.” (J.A. at 2232 (Mallet Website).) Prior to 2018, it manufactured about fifty

1 Mallet was acquired in 2016 by Vantage Specialty Chemicals, Inc. and, though the record is not clear on this, now appears to be a subsidiary operating under Vantage’s food division. 2 We refer to Mallet’s Federal Rule of Civil Procedure 30(b)(6) Deposition as “Mallet Depo.”

5 different release agents, including its “Vegalube Super P” (“Super P”), which it calls “the premier and best-performing baking release agent product in the market.” (J.A. at 2008 (Porzio3 Decl.), 11006 (Mallet Depo.).) Mallet contends that it has “take[n] substantial time, research, and effort” to formulate and perfect its release agents, including Super P. (J.A. at 4332 (Ergun4 Decl.); see also J.A. at 2008 (Porzio Decl.).) After developing a product in the laboratory, additional work is needed to bring that product to scale and optimize its performance at a customer’s facility. Mallet says that its “competitive advantage … derive[s] from a unique ability to solve customer problems by cohesively integrating research and development, technical service, custom packaging and manufacturing, and efficient distribution.” (J.A. at 2220 (Mallet Website).) To safeguard that competitive advantage, Mallet has put in place several measures to protect its information, including nondisclosure and noncompetition agreements with its employees, restricted access to its lab and formulas, and password protection for its computer network.

Along with its release agent “formulas and [the] processes used to make them[,]” Mallet considers the

3 Robert Porzio is the Senior Vice President of Sales and Marketing for Vantage. In his role, he manages sales and marketing and he is responsible for overseeing the profits, losses, and overall performance for Vantage’s food business, including Mallet. 4 Roja Ergun is the Food Technology Director for Mallet.

6 following information to be its “confidential, proprietary, trade secret information”:

specific products sold to customers or purchased from suppliers; all information pertaining to Mallet’s business with its customers and its suppliers; Mallet’s sales data and cost data; the body of knowledge about the development, production, and application of Mallet’s release agents and equipment, including the tailoring of release agents and equipment for specific customer challenges; information about the internal business affairs of any customers, suppliers, distributors, agents and contractors doing business with Mallet; pricing information; strategies; marketing information; and exclusive relationships with certain suppliers of release agent ingredients.

(J.A. at 1638 (Mallet’s Proposed Findings of Fact), 1937-38 (Topercer5 Decl.).) According to Mallet, “the trade secret in question here is the overall body of knowledge that connects … the development, production, application and implementation of the release agent … coupled with Mallet’s proprietary equipment, which go hand in hand with [a] formulated solution.” (J.A. at 11000-01 (Mallet Depo.).)

5 Benjamin Topercer is the Chief Human Resources Officer at Vantage and, in his role, supports human resources functions for Mallet.

7 As sweeping as that statement is, Mallet does recognize some limits on what it can claim as a trade secret. For example, it does not consider its “product data sheets” to be trade secret information, since those specification sheets are “produced and provided to consumers of its products[.]”6 (J.A. at 10993 (Mallet Depo.).) It also agrees that some ingredients in baking release agents – again, mineral oils, vegetable oils, and lecithin – have been common knowledge in the industry for more than thirty years, and that the components for release agents are published in product data sheets, articles, and company websites, and are therefore public knowledge, though the precise ratios and processes for combining them are not. In addition, Mallet acknowledges that “there are numerous patents … that have been published … since at least the early 1900s that talk about the manufacturing and processes and formulations that can be used to create bakery release agents[.]” (J.A. at 10982 (Mallet Depo.).) It thus admits that the contents of patents and other information generally known in the industry about “various ingredients for use in bak[ing] release agents” cannot be considered proprietary. (J.A. at 10990, 11000-01 (Mallet Depo.).)

Mallet further recognizes that its own patents disclose “various formulas for the creation of the lubricants[,]” “examples of blends and blend ratios[,]” and a “series of different formulated release agents[.]” (J.A. at 10995-96,

6 Product data sheets are public product descriptions that identify the “ingredients as well as the origins of those ingredients for each of the baking release agents.” Mallet distributes product data sheets to customers and utilizes them as marketing materials at trade shows.

8 10999-11000 (Mallet Depo.).) Those patents publicize some properties of each formulated release agent “based on various tests that Mallet … had conducted,” including “viscosity, stability, texture and other releasing characteristics.” (J.A. at 10999-11000 (Mallet Depo.).) While seeming to concede that information in patents cannot – at least by itself – constitute trade secrets, Mallet contends that even formulas in its patents can be part of its trade secrets. It says that such formulas may “form a part of the examples of the patent” and still be “part of a trade secret.”7 (J.A. at 11001-02 (Mallet Depo.).) In addition, it distinguishes the “particular formulation[s]” that its patents cover from the “know-how” that Mallet has developed over its eighty-year presence in the marketplace and that it continues to utilize on an ongoing basis for the “formulation, application[,] and implementation of [its] release agents for customers.” (J.A. at 10974, 10999 (Mallet Depo.).) According to Mallet, that know-how is a trade secret. And two of its former employees, Ada Lacayo and William Bowers, had substantial access to it.

a. Lacayo’s Employment with Mallet

Lacayo first worked for Mallet from 1997 to 2001 as a Technical Services Manager. While in that role, she “managed Quality Control laboratory employees, created specifications

7 At oral argument, for the first time, Mallet drew our attention to a set of documents that it said are specific examples of formulas it treats as proprietary. Having reviewed those pages, and without the aid of any clarifying testimony in the record, we remain at a loss to know whether they contain trade secrets.

9 and qualified new vendors, and developed nutritional information for products, among other things.” (J.A. at 2423 (Lacayo Decl.).) After a few years away from the company – during which she did not work with baking release agents – Lacayo returned to Mallet in 2006 to work as the Director of Lab Services until 2014, when she became the Director of Technical Services in the Sales and Marketing group.

Lacayo’s job responsibilities spanned all aspects of Mallet’s release agent business, from product development and quality control to customer-specific applications and technical support. Through her director positions, she obtained extensive access to Mallet’s technical information. That information allowed her to analyze ingredient interactions, create over two dozen new product formulas and processing methods, and perform “economic justifications and case studies to substantiate improvements.” (J.A. at 4737 (Lacayo Resume).) In addition, Lacayo played a key role in quality control, running onsite tests for customers, troubleshooting issues, and recommending changes to improve product performance. Along with educating individual customers, Lacayo promoted Mallet’s products, solutions, and machinery more generally. She “wrote and designed manuals, instructional programs, marketing materials, [and] presentations,” which she “delivered … to diverse audiences in English and Spanish.” (J.A. at 4737 (Lacayo Resume).) She also participated in trade shows, “[m]anaged the Latin American machinery and product introduction program,” and “[c]onducted seminars on product lines.” (J.A. at 4737 (Lacayo Resume).) Lacayo was, as she describes herself, a “product portfolio and applications expert” for Mallet. (J.A. at 4737 (Lacayo Resume).) And as a result of her extensive exposure to all sides of Mallet’s business and the “know-how

10 [she] gleaned from Mallet over decades[,]” she was widely known by “the customer base of the baking industry[.]” (J.A. at 2004 (Porzio Decl.).)

Given the extent to which she was engaged in Mallet’s business, the company insisted that, as a condition of her employment, Lacayo execute a nondisclosure and noncompetition agreement. According to that agreement, Lacayo could “not disclose any information regarding [Mallet’s] affairs” during or after her employment.8 (J.A. at

8 The scope and validity of that agreement is not before us now. To the extent Mallet seeks relief for disclosure or use of non-trade secret information that it contends is contractually protected under the nondisclosure agreement, that is an inquiry for the District Court to resolve in the first instance. In deciding to impose the injunction now at issue, the District Court tied the reasonableness of Lacayo’s noncompetition agreement and the irreparable harm deriving from her breaches of that agreement to Mallet’s trade secrets. (See J.A. at 19 ¶ 8 (“Lacayo’s three-year restrictive covenant was reasonably tailored to protect trade secrets given her work with Mallet’s formulas, its most secret information.”); J.A. at 35 ¶ 82 (“The irreparable harm Mallet would suffer without injunctive relief is varied and evident. Lacayo … [is] actively working for a direct competitor in [a] high-level position[] in which [she] ha[s] used and disclosed Mallet’s trade secrets in violation of [her] covenant[].”).) Since consideration of Lacayo’s alleged nondisclosure agreement breaches was limited to her alleged disclosure and use of Mallet’s trade secrets, we limit our discussion to trade secrets and do not separately consider Lacayo’s contractual obligations.

11 1950.) She also agreed not to “directly or indirectly” work for, “engage in, or be connected with, any business competitive with [Mallet’s] business” in any capacity for three years after her employment with Mallet ended. (J.A. at 1950.)

b. Bowers’s Employment with Mallet

Bowers began working for Mallet in 1978. Except for two brief stints away from the company, he was employed by Mallet until January 2019. Over that forty-year period, Bowers worked in sales and ultimately became Mallet’s Director of National Accounts. Throughout his long tenure with Mallet, Bowers gained access to its trade secrets and worked with some of its most valued customers. He understood how to provide service to Mallet’s customers and was privy to information about Mallet’s sales strategies and the “types of [product application equipment] Mallet might use for some customers[.]” (J.A. at 5668-70 (Bowers Depo.).) Bowers also worked closely with Mallet’s research lab to improve product performance and with its customers to test products and resolve complaints.

As a condition of his employment, Bowers entered into a nondisclosure agreement prohibiting him from disclosing or using Mallet’s trade secrets except in furtherance of Mallet’s interests.9

9 Although Bowers did enter into a noncompetition agreement with Mallet when he initially joined the company in 1978, he did not have one in place at the time of his departure in 2018, and Mallet has not pleaded the existence of such an agreement in its complaint. As with Lacayo’s nondisclosure and noncompetition agreement, see supra note 8, Bowers’s

12 2. Bundy, Synova, and the Baking Release Agent Industry

Founded in 1964, Bundy is a privately-held company with twenty-six facilities worldwide and a large customer base. It operates under several brands and is “dedicated to offering products and services in the industrial baking business[.]” (J.A. at 2409-10 (R. Bundy10 Decl.).) Bundy manufactures and sells baking pans and coatings, offers commercial food services, provides pan cleaning and recoating services, buys and resells pre-owned bakery equipment, and offers mixers and processing equipment for sale. Touching off the present battle, Bundy has added baking release agents to its list of product offerings. According to the Defendants, several of Bundy’s customers wanted “to see some alternatives brought to the [release agent] marketplace” and requested that Bundy enter the market. (J.A. at 13331 (R. Bundy Hearing Testimony); see also J.A. at 2411 (R. Bundy Decl.).) That led to the creation of the most recent Bundy brand, Synova.

nondisclosure agreement is relevant now only to the degree that it overlaps with his obligation not to misappropriate Mallet’s trade secrets, so it is not treated separately. (See J.A. at 35 ¶ 82 (“The irreparable harm Mallet would suffer without injunctive relief is varied and evident. … Bowers [is] actively working for a direct competitor in [a] high-level position[] in which [he] ha[s] used and disclosed Mallet’s trade secrets in violation of [his] covenant[].”).) 10 Robert Bundy is the president of Synova.

13 Synova was formally “created on April 27, 2017 and launched on May 15, 2019 to manufacture and distribute external baking release agents and oils.” (J.A. at 2410 (R. Bundy Decl.).) Between its creation and its launch, Synova’s President, Robert A. Bundy, was “engaged in business development[,] … looking for as much information on as many topics as [he] could get” on the baking release agent industry. (J.A. at 3628-29 (Bundy/Synova Depo.).) Along the way, he sought information from and recruited Lacayo and Bowers.

But first, Mr. Bundy approached Shane Zhou, a former Mallet employee. He wanted Zhou to help design Synova’s production facility. He asked Zhou for “the type and general formula” for ten products – and more specifically, whether those products “are pan oils, greases, etc. and some information about the quantities of the different base ingredients” – to “help [him] size some of the bulk tanks and piping.” (J.A. at 1894 (R. Bundy Email).) On January 4, 2018, Mr. Bundy offered Zhou a position as Synova’s Lab Director and agreed to indemnify Zhou “for any non-compete and/or legal action that may result from [his] becoming an employee of Synova.” (J.A. at 4283-84 (Offer Letter).) Synova withdrew its offer when Zhou sought only to become a consultant but still wanted the same indemnification. Shortly thereafter, on January 22, 2018, Mr. Bundy interviewed and hired Lacayo as Synova’s Lab Director, while she was still employed with Mallet. Around the same time, he reached out to Bowers, who was also then with Mallet, asking him if he “would be comfortable … helping [Synova] gather some information about the oils that [its] future customer base will require.” (J.A. at 2023 (R. Bundy Email).) Specifically, Mr. Bundy sent Bowers a questionnaire that Synova had put

14 together to use as “a data collection form” that would help Synova understand “what the customer order patterns will look like[,]” since it was new to the business. (J.A. at 2023-24 (R. Bundy Email).) “That data would be very helpful” while Synova was “still designing the process[,]” and Mr. Bundy “wanted to see if [Bowers] thought [the questionnaire] was the right idea[.]” (J.A. at 2023-24 (R. Bundy Email).)

a. Lacayo’s Employment with Synova

Although Lacayo secretly interviewed and accepted a position with Synova on January 22, 2018, she remained employed with Mallet until February 12, 2018. When she did finally announce that she was leaving Mallet, Lacayo concealed her employment with Synova and informed Mallet that she was instead leaving to take care of her mother.

Just three days before her interview with Synova, on January 19, 2018, Lacayo copied 1,748 files onto a USB drive. Those “bulk copied files were stored across four main (root) folders” titled: “Mallet Lab Methods, MRO Project, Supplier Approval Program, and Supplier Information.” (J.A. at 6105 (Price11 Decl.).) She also emailed information, including screenshots of two formulas, from Mallet’s files to her private Gmail account. On February 28, 2018, when she was no longer employed with Mallet, Lacayo emailed to herself a spreadsheet with technical data from Mallet’s research.

11 Paul Price is a digital forensic expert for Mallet, who was asked to determine if any of Mallet’s data had been transferred to Lacayo’s devices, including those she used while working for Synova.

15 Around that time, Mallet discovered that Lacayo had sent emails containing Mallet’s formulas from her Mallet email account to her personal Gmail account just before she resigned. It consequently sent Lacayo a cease-and-desist letter, demanding that she honor her obligation not to work for a competitor, immediately return all of Mallet’s data, and stop using or disclosing Mallet’s confidential information. Lacayo responded that she had not shared Mallet’s information and that she would destroy all Mallet information in her possession. She continued to conceal that she was working for Synova, saying that she was “taking some time off.” (J.A. at 4571 (Lacayo Depo.); see also J.A. at 1941-42 (Topercer Decl.).)

During discovery in this case, “over 1,000 documents” containing “metadata associated with Mallet” were found on Lacayo’s Synova computer, with 649 of those documents having “a Mallet logo … [branded] on the face of the document.” (J.A. at 6109-10 (Price Decl.).) Another “108 files that are an exact match for documents on” Lacayo’s Mallet computer were found on her Synova computer. (J.A. at 6106 (Price Decl.).) Digital forensic evidence indicates that Lacayo not only copied those documents but also used them,12

12 Price explained that “use” of a Mallet document involved copying the document from Lacayo’s USB drive – on which she had originally copied documents from her Mallet computer – to her Synova computer and then editing that document “some time later. What that means is the document was opened, changes were made, and those changes were saved on the date shown as ‘File System Last Modified Date.’” (J.A. at 6110 (Price Decl.).)

16 including a Mallet release agent formula and associated pricing information, while working for Synova.

The purloined documents, however, are not the whole of the problem Mallet has with Lacayo. It says that “the value she brings [to a competitor] goes far beyond any particular formula she may have provided” or any documents she may have stolen. (J.A. at 10988 (Mallet Depo.).) “It’s really the know-how that she brings” to Synova that Mallet says it is worried about. (J.A. at 10988 (Mallet Depo.).) After “work[ing] for Mallet for” so long, Lacayo “has quite a lot of know-how that went with her to the Bundy organization[,]” including information about “the formulation, application and implementation of release agent” products that Mallet had “developed over the course of its 80 years.” (J.A. at 10972-74 (Mallet Depo.).) And that know-how, it says, “would be impossible to erase from her mind.” (J.A. at 10973 (Mallet Depo.).)

In January 2018, Synova was in the earliest stages of its existence, and while “[t]he development of the release agents had already begun,” it had not completed a final product. (J.A. at 3578, 3626 (Bundy/Synova Depo.).) Synova had “identified the archetypes of ingredients that would be required and broadly [knew] the ratios of those ingredients.” (J.A. at 3581 (Bundy/Synova Depo.).) But it was still in the research and development process and had not yet conducted “any internal product testing on a release agent[.]” (J.A. at 3578 (Bundy/Synova Depo.).) As Mr. Bundy explained, “that was part of the reason to hire someone with a good science background[,]” like Lacayo. (J.A. at 3578 (Bundy/Synova Depo.).) Less than ten months after joining Synova, Lacayo had formulated a lineup of release agents, which Synova

17 marketed as direct replacements for Mallet’s release agents. Indeed, in internal correspondence it explicitly described its new formulas as “Synova=Mallet.” (J.A. at 6032-34 (Lacayo Email).) Lacayo provided oil blend recipes to Synova, built Synova’s processes and programs, and touted her ability to match a Mallet product for a customer.

b. Bowers’s Employment with Synova

Bowers was not burdened with a sense of loyalty either. After learning that Bundy was considering entry into the baking release agent marketplace, and while he was still employed with Mallet, Bowers began sharing information with Mr. Bundy about the release agent business generally and about Mallet’s business specifically. He forwarded to Mr. Bundy internal emails about Mallet’s customers, its pricing, its overall performance, and problems that customers were experiencing with Mallet. He later said he did so to “save [Mr. Bundy and Synova] some legwork.” (J.A. at 5714 (Bowers Depo.).)

When he resigned from Mallet, Bowers forwarded Mallet’s customer and product information to his wife’s email account and wiped clean all of his Mallet electronic devices. He admitted that if Mallet had the opportunity to search his personal email account, it could find emails about Mallet’s business dealings with customers, tech sheets, and pricing.

On January 23, 2019, Synova hired Bowers as its Business Development Manager. His position with Synova, similar to his previous position with Mallet, centered on selling release agents. But he joined Synova in a more limited

18 capacity. His role was to “com[e] in, mak[e] the introductions, and [the Bundy reps] took it from there.” (J.A. at 5715-16 (Bowers Depo.).) Bowers “was just [t]here to help things get off the ground.” (J.A. at 5716 (Bowers Depo.).)

At that point, Synova’s facility was still under construction but it was already testing with a prospective customer an early version of its new Supra 130 product, a release agent Synova marketed as a direct competitor of Mallet’s Super P. Over the next couple of months, Synova completed five successful product test runs with several of Mallet’s top customers, placing Synova in a position to gain immediate market penetration. And that gain was realized when those test runs in fact led to business for Synova. As a result of Lacayo’s success in bringing several products to market, along with Bowers’s concentrated efforts to sell to companies whose accounts he had serviced at Mallet, Synova was able to make its competitive debut before the construction of its baking release agent production facility was fully completed.

B. Procedural Background

After discovering that Lacayo was working for Synova in violation of her noncompetition agreement, Mallet filed this lawsuit. It brought claims for trade secret misappropriation under both federal and state law, inevitable disclosure, conversion, and unfair competition against the Defendants; breach of contract and breach of fiduciary duty against Lacayo and Bowers; and tortious interference with contractual relations and aiding and abetting breach of fiduciary duty against Bundy and Synova. Based on those claims, Mallet

19 sought to preliminarily enjoin the Defendants from engaging in any competition against it.

The District Court promptly acted upon Mallet’s application for emergency relief, denying a motion for a temporary restraining order, granting limited discovery, and entering an order governing the preliminary injunction proceedings. It twice granted extensions of the expedited proceedings, first in response to a joint motion and then for the benefit of Bowers, who was only informed nine days before discovery ended, during his deposition, that Mallet intended to amend its motion for preliminary injunction to seek injunctive relief against him.13 After dealing with multiple discovery disputes, the District Court presided over a preliminary injunction hearing where it took testimony, admitted 181 exhibits, and ultimately considered more than 10,000 pages of evidence. It then decided that Mallet was entitled to injunctive relief on most of its claims.14

Adopting many of Mallet’s proposed findings of fact, with the injunction order in turn incorporating certain factual findings by broad reference to the ranges of paragraph numbers listed in Mallet’s proposed order, the Court determined that Mallet had demonstrated a likelihood of success on the merits

13 Mallet initially sought to preliminary enjoin Bundy, Synova, and Lacayo. Bowers was not added as a defendant until January 29, 2020. 14 The District Court concluded that it was unnecessary to address Mallet’s claims for inevitable disclosure and conversion.

20 for several of its claims, including its trade secret misappropriation claims. Specifically, the District Court found that “[a]t least some of the Mallet information in question, possessed by Defendants, satisfies the trade secret definition(s),” including, “among other things, highly sensitive details about how Mallet produces, markets and sells its release agents[.]” (J.A. at 25.) The Court listed thirteen categories of Mallet information it deemed “protected materials,” as follows:

Mallet’s formulas; customer purchase orders demonstrating Mallet’s pricing; identification of customers experiencing difficulty with Mallet’s products; internal discussions of “actual major problems” at customer locations; internal discussions of how Mallet would address issues with its products; internal discussions of customers’ preferences and complaints; Mallet’s completed organic certifications; identification of Mallet’s supply source for product ingredients; Mallet’s internal manuals and procedures showing how Mallet’s lab is operated; pricing and volume data; information about Mallet’s equipment; Mallet’s training materials showing how Mallet markets and sells its products; and a compilation of Mallet’s product specification sheets.

(J.A. at 25-26.) The Court then held there had been “a misappropriation of Mallet’s trade secrets[,]” based on “[t]he timing of Bowers’s and Lacayo’s sending Mallet information to themselves – around the time they agreed to work for Bundy/Synova[.]” (J.A. at 28.) It also concluded that Bundy

21 and Synova had “acquired a substantial volume of Mallet’s confidential and trade secret information and [had] done nothing to stop use of this data, even after litigation commenced.” (J.A. at 29.) And it found that “[a]ctively concealing plans to form a competing company; using employee status to copy documents onto external storage drives; the existence of information taken close to or immediately following resignations on the new company’s computers; and failure to return devices used during former employment are all factors showing the substantial threat of a defendant disclosing trade secrets.” (J.A. at 30.)

The District Court determined that such actual and threatened misappropriation would irreparably harm Mallet, absent an injunction. The Court reasoned that “Lacayo and Bowers are actively working for a direct competitor in high- level positions in which they have used and disclosed Mallet’s trade secrets in violation of their covenants.” (J.A. at 35.) And “[e]ven without Lacayo and Bowers, Bundy/Synova would continue to have access to and use Mallet’s information.” (J.A. at 35.) So, according to the District Court, “[t]he harm Bundy/Synova will visit on Mallet if it is not enjoined will be long-term given that Bundy/Synova is targeting Mallet’s customers.” (J.A. at 35.) And that “harm to Mallet is magnified” since Synova’s “portfolio of release agents to replace Mallet products” is “expanding[.]” (J.A. at 35.) Additionally, the District Court concluded that, “[w]ithout injunctive relief, [the] Defendants’ actions will disrupt Mallet’s business, harm its relationship with customers and cause loss of market position, reputation and customer goodwill.” (J.A. at 35.)

22 Balancing Mallet’s irreparable harm against the harm that the Defendants would suffer from an injunction, the District Court found that “[t]he nature of the relief Mallet requests is proportionate to the severity of [the] Defendants’ wrongful acts and the potential harm[,]” particularly where “[i]t took Mallet years to develop some of its key products,” “Synova has made comparable products within a few months of Lacayo joining the Company[,]” and Mallet has demonstrated “that th[o]se comparable products benefited from Mallet’s trade secrets” that “Bundy/Synova purposefully sought out from Mallet’s current and former employees.” (J.A. at 36.)

After concluding that a preliminary injunction was warranted, the District Court ordered the Defendants enjoined from the following activities (among others):

• “using Mallet’s confidential, proprietary and/or trade secret information in any respect” (J.A. at 42);

• “directly or indirectly formulating, manufacturing, distributing or selling products competitive to Mallet products, including any release agents and related equipment[,]” “working in the industry of release agents and associated equipment (in any capacity including, but not limited to, working as a marketer, agent, consultant, contractor or distributor) using Mallet’s protectable information[,]” and “directly or indirectly soliciting any Mallet customers for marketing, testing or the sale of release agents, oils or equipment in any regard” (collectively, a “production ban”) (J.A. at 44);

23 • “soliciting and/or unlawfully interfering with the business, employment or contractual relationship between Mallet and its agents, employees and/or independent contractors who have access to Mallet’s confidential, proprietary and/or trade secret information” (J.A. at 44);

• “directly or indirectly disseminating any marketing materials, client communications (written or verbal), or other documents comparing Bundy and/or Synova products to any Mallet products.” (J.A. at 45.)

It also prohibited Lacayo and Bowers from engaging in the following activities:

• “working, directly or indirectly, in any capacity (including, but not limited to, working as an employee, independent contractor, marketer, agent, or consultant), for Bundy, Synova or any person or entity that is competitive with Mallet” (J.A. at 40);

• “contacting Bundy’s and/or Synova’s employees, officers, directors, leadership, consultants or any other persons in a fiduciary relationship with Bundy and/or Synova or any parent, subsidiary or affiliate of Bundy and/or Synova” (J.A. at 40-41);

• “directly or indirectly contacting or soliciting Mallet’s customers.” (J.A. at 45.)

24 The injunction order did not elaborate on what information constitutes trade secrets.15

In setting the amount of the bond that Mallet would have to post to secure the preliminary injunction, the District Court found the “Defendants’ proposal, in excess of $20 million, [to be] astronomical by comparison” to Mallet’s proposal of providing a $500,000 bond. (J.A. at 45 n.2.) And it reminded the parties that it had “expressly advise[d] that it would ‘summarily … adopt the [bond proposal] it believe[d to be] most reasonable, appropriate and consistent with’” its ruling. (J.A. at 45 n.2 (third alteration in original).) The District Court subsequently set the bond amount at $500,000. The injunction order went into effect immediately after Mallet posted the bond and was to “remain in full force and effect until the [District] Court’s ruling on permanent injunctive relief.” (J.A. at 45.)

Bundy and Synova asked the District Court to stay the injunction, pending an appeal. The following day, they also moved for clarification on the scope of the injunction order’s production ban since its terms “do not specify” whether the production ban “appl[ies] only to the commercial baking industry[.]” (J.A. at 1856.) Although the commercial baking industry is the sole industry that Mallet services, Bundy and

15 The District Court issued two orders associated with the preliminary injunction, one being the decision to grant Mallet’s request for injunction relief and the other consisting of the terms of the injunction and the bond amount. For purposes of simplicity, we refer to those orders in the singular as one injunction order. December 15, 2020 is when the second order issued.

25 Synova sought assurance “that they [were] permitted to operate outside of” that business. (J.A. at 1858.) They did so, they said, “out of an abundance of caution to avoid” contempt violations and “in an effort to maintain their compliance with” the injunction order. (J.A. at 1857.)

That same day, the District Court denied the motion for clarification. (J.A. at 56.) But, in the order of denial, the Court did in fact clarify that producing release agents for “other industries, such as processed meats, pharmaceuticals, construction and other industrial applications” would violate the injunction. (J.A. at 56.) It explained that the “Defendants’ potential use of [Mallet’s] trade-secret and/or confidential information to pursue other industrial applications would be inconsistent with the terms, spirit and intent of the detailed factual findings and legal conclusions reflected in the merits- ruling.” (J.A. at 56.) The District Court also denied the motion to stay, but we stayed the injunction in part,16 while considering this appeal.

16 Specifically, we immediately stayed paragraphs one through four, ten through eleven, and fourteen of the preliminary injunction, which enjoined Lacayo and Bowers “from working, directly or indirectly, in any capacity … for Bundy, Synova or any person or entity that is competitive with Mallet”; contacting anyone in a fiduciary relationship with Bundy, Synova, or any parent, subsidiary, or affiliate of them; and accessing any Bundy or Synova electronic systems; and which enforced a production ban against the Defendants; and prohibited the Defendants from “directly or indirectly disseminating any marketing materials, client communications … , or other documents comparing” Bundy

26 II. DISCUSSION17

Federal Rule of Civil Procedure 65(d) mandates that “[e]very order granting an injunction” set forth the reasons why an injunction is warranted, “state its terms specifically[,]” and

and Synova products with Mallet products. (J.A. at 40-41 ¶¶ 1- 4, 44 ¶¶ 10-11, 45 ¶ 14.) 17 The District Court had original jurisdiction over Mallet’s federal trade secret misappropriation claim pursuant to 18 U.S.C. § 1836(c) and 28 U.S.C. § 1331. The District Court had supplemental jurisdiction over Mallet’s state law claims under 28 U.S.C. § 1367. The Defendants have contended that the District Court lacked jurisdiction because Mallet failed to prove that the Defendants misappropriated a protectable trade secret. But that is a merits argument, not a jurisdictional one, and we reject it. We have jurisdiction over the Defendants’ interlocutory appeal under 28 U.S.C. § 1292(a)(1). “When reviewing a district court’s grant of a preliminary injunction, we review the court’s findings of fact for clear error, its conclusions of law de novo, and the ultimate decision granting the preliminary injunction for an abuse of discretion.” Bimbo Bakeries USA, Inc. v. Botticella, 613 F.3d 102, 109 (3d Cir. 2010). “Despite oft repeated statements that the issuance of a preliminary injunction rests in the discretion of the trial judge[,] whose decisions will be reversed only for ‘abuse,’ a court of appeals must reverse if the district court has proceeded on the basis of an erroneous view of the applicable law.” Kos Pharms., Inc. v. Andrx Corp., 369 F.3d 700, 708 (3d Cir. 2004) (alteration in original) (quoting Apple Comput., Inc. v. Franklin Comput. Corp., 714 F.2d 1240 (3d Cir. 1983)).

27 articulate “in reasonable detail” the conduct it enjoins. While “[t]he degree of particularity required” to satisfy the Rule’s specificity provisions will “depend[] on the nature of the subject matter[,]” Ideal Toy Corp. v. Plawner Toy Mfg. Corp., 685 F.2d 78, 83 (3d Cir. 1982), those provisions are not mere technicalities. Schmidt v. Lessard, 414 U.S. 473, 476 (1974). They are designed to serve two vital functions: first, “to prevent uncertainty and confusion on the part of those faced with injunctive orders, and [thus] to avoid the possible founding of a contempt citation on a decree too vague to be understood”; and second, to ensure that sufficient information is placed on the record so that “an appellate tribunal [will] know precisely what it is reviewing.” Id. at 476-77. Failure to “satisfy the important requirements of Rule 65(d)” will typically result in an injunction’s vacatur. Id. at 477; see also Kos Pharms., Inc. v. Andrx Corp., 369 F.3d 700, 712 (3d Cir. 2004) (“When reviewing an order that does not adequately support the resolution of a motion for preliminary injunction, we may vacate and remand for additional findings[.]”).

That is the result required here. We fully appreciate the challenges inherent in expedited proceedings. Nevertheless, when an injunction lacks sufficient specificity to permit meaningful appellate review, there needs to be another effort at crafting the contours of the order. Because the District Court did not identify with specificity the information it found to be Mallet’s trade secrets, we are not in a position to make an informed decision as to whether Mallet is likely to prevail on its trade secret misappropriation claims. And, notwithstanding the existence of other claims that the District Court concluded Mallet was likely to prevail on, there is no remaining basis to uphold the Court’s decision to grant injunctive relief since its

28 irreparable harm determination appears to depend entirely on the Defendants’ misappropriation of Mallet’s trade secrets.18

We will accordingly vacate the injunction order and remand for reconsideration. In doing so, we outline a few matters to be considered when identifying allegedly misappropriated trade secrets. We also discuss the permissible scope of an injunction and the limits of a district court’s discretion when determining the associated amount of a bond.

A. A preliminary injunction predicated on trade secret misappropriation must adequately identify the allegedly misappropriated trade secrets.

To prove eligibility for a preliminary injunction, there is a well-established four-part test. The moving party must demonstrate, first, a likelihood of success on the merits, and second, that “it is more likely than not to suffer irreparable harm in the absence of preliminary relief.” Reilly v. City of Harrisburg, 858 F.3d 173, 179 (3d Cir. 2017) (footnote omitted). Absent either of those threshold factors, “[w]e cannot sustain a preliminary injunction ordered by the district court[.]” Hoxworth v. Blinder, Robinson & Co., Inc., 903 F.2d

18 The facts preliminarily found concerning Lacayo’s extensive knowledge of Mallet’s business and her work for Bundy and Synova, see supra Sections I.A.1.a & 2.a, may be enough to demonstrate irreparable harm for breaching her noncompetition agreement with Mallet – even without trade secret misappropriation. But we leave that to the District Court to decide in the first instance.

29 186, 197 (3d Cir. 1990) (quoting In re Arthur Treacher’s Franchisee Litig., 689 F.2d 1137, 1143 (3d Cir. 1982)). If both factors are established, however, the district court considers the two remaining factors – whether granting relief will result in even greater harm to the nonmoving party or other interested persons and whether the public interest favors such relief. The court then “determines in its sound discretion” whether the balance of all four factors warrants granting preliminary relief. Reilly, 858 F.3d at 179.

To establish a likelihood of success, a plaintiff must show that “there is ‘a reasonable chance, or probability, of winning.’” In re Revel AC, 802 F.3d 558, 568 (3d Cir. 2015) (quoting Singer, 650 F.3d at 229)). That does not require a “more-likely-than-not showing of success on the merits.” Id. at 179 & n.3. But it does require the plaintiff to “demonstrate that it can win on the merits,” which involves a showing that its chances of establishing each of the elements of the claim are “significantly better than negligible.” Id.

For a federal trade secret misappropriation claim, those elements are: “(1) the existence of a trade secret … (2) that ‘is related to a product or service used in, or intended for use in, interstate or foreign commerce[,]’ and (3) the misappropriation of that trade secret[.]” Oakwood Labs. LLC v. Thanoo, 999 F.3d 892, 905 (3d Cir. 2021) (second alteration in original) (citations omitted). And, of course, each of those elements is predicated on an adequate identification of what the plaintiff contends to be its trade secret. See id. (“To plead the existence of a trade secret in a misappropriation claim …, [a plaintiff] must sufficiently identify the information it claims as a trade secret[.]”).

30 The Defend Trade Secrets Act (“DTSA”)19 defines a trade secret as information that “the owner thereof has taken reasonable measures to keep … secret” and that “derives independent economic value, actual or potential, from not being generally known to, and not being readily ascertainable through proper means by, another person who can obtain economic value from the disclosure or use of the

19 Mallet has asserted claims for misappropriation under the DTSA and the Pennsylvania Uniform Trade Secrets Act (“PUTSA”). (J.A. at 409-10 (claiming “actual and continuing misappropriation in violation of the DTSA”), 411-12 (claiming “actual or threatened misappropriation of trade secrets … in violation of the [PUTSA]”).) The DTSA and the PUTSA are substantially similar, as both are closely related to the Uniform Trade Secrets Act (“UTSA”). Heraeus Med. GmbH v. Esschem, Inc., 927 F.3d 727, 736 (3d Cir. 2019) (“The Pennsylvania General Assembly based the PUTSA on the provisions of the [UTSA].”); Oakwood Labs. LLC v. Thanoo, 999 F.3d 892, 910 (3d Cir. 2021) (“[Congress] recognized the DTSA and the UTSA as similar.” (quotation marks and citation omitted)). Indeed, the definitions of “trade secret” under the DTSA and the PUTSA are almost identical. Compare 18 U.S.C. § 1839(3), with 12 Pa. Cons. Stat. § 5302. Thus, although our discussion focuses on the DTSA, we conclude that the same analysis applies to Mallet’s claims under the PUTSA and the same outcome results. See, e.g., Teva Pharms. USA, Inc. v. Sandhu, 291 F. Supp. 3d 659, 675 (E.D. Pa. 2018) (consolidating the analysis of claims under the DTSA and the PUTSA); Jazz Pharms., Inc. v. Synchrony Grp., LLC, 343 F. Supp. 3d 434, 445 (E.D. Pa. 2018) (same).

31 information[.]”20 18 U.S.C. § 1839(3). We cannot evaluate whether a plaintiff is likely to succeed on any element of a trade secret misappropriation claim until the plaintiff has sufficiently described those trade secrets. See Porous Media Corp. v. Midland Brake Inc., 187 F.R.D. 598, 600 (D. Minn. 1999) (“Failure to identify the trade secrets with sufficient specificity renders the Court powerless to enforce any trade secret claim.”). It follows that a district court’s injunction order must first adequately identify the information to which it accords trade secret status. Otherwise, the injunction order lacks the foundation necessary for holding a plaintiff likely to prevail on its misappropriation claim. Without that information, the injunction order fails to comply with Rule 65(d), and it must be vacated. Cf. PMC, Inc. v. Sherwin-Williams Co., 151 F.3d 610, 620 (7th Cir. 1998) (“Legal rules committing decisions to judicial discretion suppose that the court will have, and give, sound reasons for proceeding one way rather than the other.” (citation omitted)).

20 That information includes “all forms and types of financial, business, scientific, technical, economic, or engineering information, including patterns, plans, compilations, program devices, formulas, designs, prototypes, methods, techniques, processes, procedures, programs, or codes, whether tangible or intangible, and whether or how stored, compiled, or memorialized physically, electronically, graphically, photographically, or in writing[,]” 18 U.S.C. § 1839(3), so long as it is “information with independent economic value that the owner has taken reasonable measures to keep secret,” consistent with the DTSA. Oakwood, 999 F.3d at 905.

32 The District Court determined that “[a]t least some of the Mallet information in question” constitutes protectable trade secrets, “includ[ing], among other things, highly sensitive details about how Mallet produces, markets and sells its release agents[.]” (J.A. at 25.) Absent from the District Court’s high-level description, however, are any specifics of what those “highly sensitive details” are. Rather, we are left with a list of thirteen broad categories of Mallet information which the District Court deemed “protected materials”:21

Mallet’s formulas; customer purchase orders demonstrating Mallet’s pricing; identification of customers experiencing difficulty with Mallet’s products; internal discussions of “actual major problems” at customer locations; internal discussions of how Mallet would address issues with its products; internal discussions of customers’ preferences and complaints; Mallet’s completed organic certifications; identification of Mallet’s supply source for product ingredients; Mallet’s internal manuals and procedures showing how Mallet’s lab is operated; pricing and volume data; information about Mallet’s equipment; Mallet’s training materials showing how Mallet markets and sells

21 It is not clear whether the term “protected materials” was intended to be synonymous with trade secrets, but we assume it was. (J.A. at 25.) Information can be contractually protected from use or disclosure and not be a trade secret.

33 its products; and a compilation of Mallet’s product specification sheets.

(J.A. at 25-26.) While some information falling within those categories may very well include trade secrets, there is a fair probability that many of the categories – and perhaps all of them – also include information that does not qualify for trade secret protection.

The injunction order’s statement of protected material is better characterized as a list of general categories of business and technical information, a list that could be used to describe documents found in any number of corporations. See A&P Tech., Inc. v. Lariviere, No. 1:17-cv-534, 2017 WL 6606961, at *10 (S.D. Ohio Dec. 27, 2017) (“Terms such as ‘engineering,’ ‘research and development procedures and materials,’ and ‘marketing materials’ could be applied to almost any corporation in existence, and do not in any way allow Defendants to properly craft a defense around the alleged misappropriation of trade secrets.”). The generic list thus falters against the standard for specifying a trade secret. See id. At a minimum, “the subject matter of the trade secret must be described ‘with sufficient particularity to separate it from matters of general knowledge in the trade or of special knowledge of those persons who are skilled in the trade, and to permit the defendant to ascertain at least the boundaries within which the secret lies.’”22 Oakwood, 999 F.3d at 906 (quoting

22 In Oakwood, we noted that “deciding whether a plaintiff has sufficiently disclosed its trade secrets is a fact- specific question to be decided on a case-by-case basis.” See 999 F.3d at 906 (internal quotation marks omitted) (quoting Vesta Corp. v. Amdocs Mgmt. Ltd., 147 F. Supp. 3d 1147, 1155

34 Diodes, Inc. v. Franzen, 67 Cal. Rptr. 19, 24 (Cal. Ct. App. 1968) (describing the minimum specificity threshold to survive a motion to dismiss). That is especially the case where, as here, the record suggests that those boundaries may not be particularly clear.23

(D. Or. 2015). District courts must therefore engage with the specific facts of the case and consider the degree of specificity necessary in light of the particular industry-based context and the stage of litigation – whether that be a motion to dismiss, a discovery dispute, a motion to preliminarily enjoin a defendant from competing with the trade secret plaintiff, or a summary judgment motion. See id. at 1153 (“[T]he Court recognizes the ‘growing consensus’ of courts from around the country who have applied the ‘reasonable particularity’ standard to determine whether a party alleging a claim for misappropriation of trade secrets has sufficiently identified its trade secrets before it may compel discovery of its adversary’s trade secrets. … [T]he ‘reasonable particularity’ standard reflects the Court’s authority, pursuant to the Federal Rule of Civil Procedure 26 requirements of early disclosure of evidence, and the Court’s authority to control the timing and sequencing of discovery in the interests of justice.” (citation omitted)); DeRubeis v. Witten Techs., Inc., 244 F.R.D. 676, 680-82 (N.D. Ga. 2007) (“Unfortunately, there is no talismanic procedure the Court may apply in order to obtain the best result in any given case. The approach taken in [one dispute] may have been perfectly appropriate in that case but may be inappropriate in this one. In other words, ‘the “proper” approach is clearly fact-dependent.’” (citation omitted)). 23 Mallet contends that its trade secrets are adequately identified because of the following series of steps: (1) the

35 District Court’s order granting an injunction was based on the Court’s Findings of Fact, Conclusions of Law & Order deciding a preliminary injunction was warranted; (2) the Court’s Finding of Fact ¶ 8 – stating that “Lacayo had access to, and was intimately familiar with, Mallet’s protectable information” – incorporated by reference the contents and record citations in Mallet’s Proposed Findings of Fact ¶¶ 44- 60 (J.A. at 14); (3) those seventeen paragraphs in Mallet’s Proposed Findings of Fact – the contents of which summarized broadly that Lacayo had access to and knowledge of trade secret information – incorporated by reference the contents and record citations of numerous exhibits from the hearing; (4) all of those exhibits are declarations and excerpts of deposition testimony that themselves attach exhibits, cited by another exhibit number or bates number, and that further refer to the content therein; and (5) somewhere in there are trade secrets. At oral argument, Mallet cited a range of bates-numbered documents as proof that it had specified its trade secrets. Circumstances such as access to trade secrets, unusually accelerated or low-cost development of a competing product, and relative lack of prior expertise, may, in combination, establish a likelihood of success on the element of misappropriation, see Oakwood, 999 F.3d at 908, 911-12. But they do not establish that likelihood for the element on which we focus here: the existence of a trade secret that has been identified “with sufficient particularity.” Id. at 906. And whatever else the foregoing order of operations might be, it is not consistent with Rule 65(d)(1), which requires that every injunction “(A) state the reasons why it issued; (B) state its terms specifically; and (C) describe in reasonable detail – and not by referring to the complaint or other document – the act

36 For example, Mallet recognizes that its own patents publicly disclose some of its formulas, but it appears to contend that even formulas thus publicly disclosed are part of its trade secrets. (See J.A. at 11001-02 (Mallet Depo.) (“What I am saying is while these formulations were developed here, it doesn’t necessarily mean that the exact formula in one of these [patent] tables is at question as a trade secret. These are examples only. They form a part of the examples of the patent. … They are part of a trade secret.”) (emphasis added).) If that is really its position – and it is hard for us to tell – then it is hard to take entirely seriously. A formula disclosed in a patent is, by definition, not a secret. Nevertheless, “[w]hile the precise information provided within or directly ascertainable from a patent cannot constitute a trade secret, patent holders are not necessarily precluded from cultivating trade secrets that go beyond the corpus of the patent or that refine the patent’s process in some proprietary way.” AutoTrakk, LLC v. Auto. Leasing Specialists, Inc., No. 4:16-cv-1981, 2017 WL 2936730, at *5 (M.D. Pa. July 10, 2017). Problematically, though, Mallet fails to explain how we or anyone else is to distinguish between what is generally known or available information and what it contends to be protectable trade secrets. With a wave of the hand, it declares everything to be secret know-how. (See J.A. at 10986 (Mallet Depo.) (“The issue at hand is not so much that the formula might be the same or different. What is at hand is that the know-how that Ms.

or acts restrained or required.” Fed. R. Civ. P. 65(d)(1) (emphasis added).

37 Lacayo took from Mallet, and is applying within Bundy, is all about delivering the performance, the quality, and those other factors I mentioned to the customer to provide them the solution. That know-how was developed by Mallet over 80 years.”).)

When the breadth of a trade secret description is so far- reaching that it includes publicly available information (like patent disclosures) and admitted industry knowledge, that information is not specific enough to be accorded trade secret status.24 DeRubeis v. Witten Techs., Inc., 244 F.R.D. 676, 689

24 See, e.g., Radiant Glob. Logistics, Inc. v. Furstenau, 368 F. Supp. 3d 1112, 1125 (E.D. Mich. 2019) (“The identification of alleged trade secrets is important because ‘the general knowledge of an employee does not constitute a trade secret[.]’ … [T]o establish a trade secrets claim, the party ‘must establish more than the existence of generalized trade secrets and a competitor’s employment of the party’s former employee who has knowledge of trade secrets.’” (citations omitted)); Broker Genius, Inc. v. Zalta, 280 F. Supp. 3d 495, 515 (S.D.N.Y. 2017) (“While neither the New York Court of Appeals nor the United States Court of Appeals for the Second Circuit has expressly required trade secrets to be identified with any particular degree of specificity, it is evident that a ‘vague and indefinite’ piece of information cannot be protected as a trade secret.” (quoting Big Vision Private Ltd. v. E.I. DuPont De Nemours & Co., 1 F. Supp. 3d 224, 258 (S.D.N.Y. 2014))); L-3 Commc’ns Corp. v. Jaxon Eng’g & Maint., Inc., No. 10-cv-2868, 2011 WL 10858409, at *2 (D. Colo. Oct. 12, 2011 (“[W]hat is clear is that generic allegations and general references to products or information are insufficient to satisfy the reasonable particularity standard.” (citation omitted));

38 (N.D. Ga. 2007) (“If the list is too general, it will encompass material that the defendant will be able to show cannot be trade secret.” (citation omitted)). While we recognize the difficulty inherent in articulating what trade secrets Lacayo and Bowers may have misappropriated – and it certainly appears they took things that may qualify as trade secrets – “care must [still] be taken to not allow a plaintiff in a trade secret misappropriation case to make generalized claims that leave a defendant wondering what the secrets at issue might be[.]” Oakwood, 999 F.3d at 907; see also Patriot Homes, Inc. v. Forest River Housing, Inc., 512 F.3d 412, 415 (7th Cir. 2008).

The District Court, in effect, recapitulated Mallet’s own broadly stated categories of information. For example, the Court said that “Mallet’s formulas” were trade secrets. (J.A. at 25.) But, like Mallet, it did not identify which formulas it

UTStarcom, Inc. v. Starent Networks, Corp., 675 F. Supp. 2d 854, 875 (N.D. Ill. 2009) (“It is simply not enough for a plaintiff to ‘point to broad areas of technology and assert that something there must have been secret and misappropriated. The plaintiff must show concrete secrets.’” (quoting Composite Marine Propellers, Inc. v. Van Der Woude, 962 F.2d 1263, 1266 (7th Cir. 1992)); Dura Glob. Techs., Inc. v. Magna Donnelly, Corp., No. 7-cv-10945, 2007 WL 4303294, at *4 (E.D. Mich. Dec. 6, 2007) (“[A] list of general categories and types of information they allege comprise their trade secrets” is not enough to “identify the trade secrets at issue with the particularity necessary for Defendant to identify the information which Plaintiffs claim was misappropriated” because they “are too general to specify the trade secrets at issue.”).

39 referred to, nor did it describe any characteristics or properties contributing specific competitive value to Mallet that could serve as a marker for separating Mallet’s formulas from publicly available information or generally known formulas in the industry.25 The District Court also concluded that “pricing

25 Though our decision relies on the District Court’s findings of fact, we note that on remand the Court is not bound by its initial findings and should again carefully assess the evidentiary record, weighing all conflicting evidence. We offer two examples of findings that give us pause, at least as presently explained. First, extensive evidence was introduced showing that many of both Mallet’s and Synova’s products were single ingredient oils, some of which were simply repackaged for sale. Given that, we hesitate to agree with the District Court that the testimony relating to unformulated pure oils is a “red herring” and immaterial to Mallet’s right to relief. (J.A. at 14.) As of December 7, 2020, repackaged mineral oils constituted approximately half of Synova’s Supra and Primo series. Further, we wonder whether a single naturally occurring ingredient can be repackaged as a product and then be considered a formula warranting trade secret protection, but we leave that for consideration in the first instance on remand. Second, the Court found that “Mallet’s evidence establishes that the formulation of its relevant products takes substantial time, sometime years” and that “Lacayo’s testimony, to the effect that formulating such products is ‘very easy,’ was not credible.” (J.A. at 14.) But those findings make no mention of other evidence that supported Lacayo’s claim about the ease of developing some formulas. Two other witnesses – including Mallet’s own witness, Roja Ergun – acknowledged that some basic release agents were in fact easy to make. Ronald Wilson, the senior director of engineering at

40 and volume data” and “Mallet’s training materials showing how [it] markets and sells its products” are trade secrets. (J.A. at 26.) But those trade secret descriptions fare no better than Mallet’s assertion that “pricing information[,] strategies[, and] marketing information” are trade secrets. (J.A. at 1638 (Mallet’s Proposed Findings of Fact.) Specific examples are needed and, if provided, could very well suffice to support injunctive relief.

Because the District Court did not articulate with particularity the information to which it accorded trade secret status, we are unable to conduct an informed appellate review, assessing the alleged trade secret information against other information that the record may reveal is publicly available, easily generated, or widely applicable to or learned in the industry. SI Handling Sys., Inc. v. Heisley, 753 F.2d 1244, 1258 (3d Cir. 1985). In other words, meaningful review requires enough factual detail to permit us to draw a connection between the alleged trade secret and its value as a “particular secret[] of the complaining employer” and not general know- how of the trade. Id. at 1256 (quoting Capital Bakers v. Townsend, 231 A.2d 292, 294 (Pa. 1967)).

We appreciate the burden this places on district courts and “that with the advantage of hindsight it is much easier for [us] to find a deficiency in a decree than it would be to write a

Hostess Brands, stated that Hostess Brands has created its own baking pan oils in-house at numerous points in the company’s history. According to Wilson, the process was “[v]ery, very simple,” with “nothing to it.” (J.A. at 4117.) Further insight into the District Court’s weighing of such conflicting evidence will assist us in providing meaningful appellate review.

41 specific decree that does not offend the law when dealing with the somewhat nebulous field of trade secrets.” Id. at 1266. Importantly, however, while it is a district court’s responsibility to adequately describe the trade secrets at issue in a case like this, it is first and foremost the plaintiff’s burden to specifically identify what it contends to be its trade secrets and to demonstrate with record evidence a “significantly better than negligible” chance, Reilly, 858 F.3d at 179, of establishing the existence of those trade secrets. If a plaintiff fails to meet that burden, the district court faces the same problem we now have on appeal, and a preliminary injunction for trade secret misappropriation ought not issue.26

While we are persuaded that some of Mallet’s information – such as that contained in its patents – cannot legitimately have the protected status that it may have been afforded by the District Court,27 we lack the information

26 That does not mean there may not be other bases for injunctive relief, only that a failure to adequately make a record identifying trade secrets with sufficient specificity means no injunction on the claim of trade secret misappropriation can issue. 27 Compare BIEC Int’l, Inc. v. Glob. Steel Servs., Ltd., 791 F. Supp. 489, 541 (E.D. Pa. 1992) (“While it is quite possible that BIEC may have an extraordinary and massive body of technical trade secrets and confidential information, … the law requires that BIEC separate those interrelated processes before it has satisfied its heavy burden of proving specific trade secrets.” (internal quotation marks and citation omitted)), with Smith v. BIC Corp., 869 F.2d 194, 201 (3d Cir. 1989) (“To extrapolate from this [i.e., the

42 necessary to decide anything more about what allegedly does have that protected status and Mallet’s likelihood of success in establishing misappropriation of that specific information. So, instead, we share two observations for consideration on remand.

First, information will not necessarily be deprived of protection as a trade secret because parts of it are publicly available. A confidential compilation and organization of public information can amount to a trade secret. “Courts have long recognized that ‘a trade secret can exist in a combination of characteristics and components, each of which, by itself, is in the public domain, but the unified process, design and operation of which, in unique combination, affords a competitive advantage and is a protectable secret.’” AirFacts, Inc. v. de Amezaga, 909 F.3d 84, 96 (4th Cir. 2018) (quoting Imperial Chem. Indus. v. Nat’l Distillers & Chem. Corp., 342 F.2d 737, 742 (2d Cir. 1965)).28 Similarly, “[w]hile the precise

plaintiff’s concession that some design information was within the public domain] to say that all design information is not a trade secret because it is in the public domain is erroneous indeed.”). 28 See also Compulife Software Inc. v. Newman, 959 F.3d 1288, 1314 (11th Cir. 2020) (While “scraped quotes [are] not individually protectable trade secrets because each is readily available to the public … that doesn’t in and of itself resolve the question whether, in effect, the database as a whole was misappropriated. Even if quotes aren’t trade secrets, taking enough of them must amount to misappropriation of the underlying secret at some point. Otherwise, there would be no substance to trade-secret protections for ‘compilations,’ which

43 information provided within or directly ascertainable from a patent [or other published document] cannot constitute a trade secret,” that does not, as noted earlier, mean that a patentee is “precluded from cultivating trade secrets that go beyond the corpus of the patent or that refine the patent’s process in some proprietary way.” AutoTrakk, LLC, 2017 WL 2936730, at *5.

Second, an employee’s general know-how should be distinguished from the particular secrets held by an employer. SI Handling Sys., 753 F.2d at 1256.29 In other words, while an

the law clearly provides.”); Radiant Glob. Logistics, Inc., 368 F. Supp. 3d at 1127 (“[I]t is widely accepted that a trade secret can exist in a combination of characteristics each of which, by itself, is in the public domain.” (alteration in original) (quoting Mike’s Train House, Inc. v. Lionel, L.L.C., 472 F.3d 398, 411 (6th Cir. 2006))); Fishkin v. Susquehanna Partners, G.P., 563 F. Supp. 2d 547, 582 (E.D. Pa. 2008) (“A trade secret may be based on publicly available information if it consists of a secret advance over common knowledge and practice or if it combines publicly available information in a new and secret way.” ). 29 See also Cap. Bakers, Inc. v. Townsend, 231 A.2d 292, 294 (Pa. 1967) (“It is true that true ‘trade secrets’ of an employer are a valuable asset, the disclosure and use of which by a former employee will be enjoined[.] … However, ‘trade secrets’ which will be so protected must be [p]articular secrets of the complaining employer and not general secrets of the trade in which he is engaged. The ‘trade secrets’ asserted herein are nothing more than those of the general bakery business.” (citation omitted)).

44 employee’s general know-how does not constitute trade secret information, employers remain free to identify and protect their particular proprietary information. Admittedly, the line distinguishing between the two – an employee’s general knowledge or skill and an employer’s protectable trade secrets – may often be difficult to draw. Thus, in exercising its equitable discretion, a district court need not draw the line with precision, but the plaintiff has to provide something better than sweeping generalities for the court to work with. See Van Prods. Co. v. Gen. Welding & Fabricating Co., 213 A.2d 769, 777 (Pa. 1965) (“[T]he concept of ‘know-how’ is … a very fuzzily defined area, used primarily as a short-hand device for stating the conclusion that a process is protect[a]ble. It covers a multitude of matters, however, which in the broad sense are not protect[a]ble, e.g., an employee’s general knowledge and skill.”). It is the trade secret owner that bears the burden of demonstrating its claimed secrets are protectable and are not general industry knowledge. Just how much specificity a court should require of the plaintiff-owner is again a context-specific matter. We cannot provide a bright-line rule. The best we can do is say that Mallet’s very general description of categories does not “sufficiently identify the information it claims as a trade secret,” Oakwood, 999 F.3d at 905, and thus does not suffice to justify the sweeping injunction the District Court issued.

The bottom line is this: without knowing what particular information Mallett claims as trade secrets, we cannot assess its likelihood of success in establishing that the information the Defendants acquired, disclosed, or used is trade secret information or that misappropriation of a trade secret has

45 occurred.30 By its very definition under the DTSA, the term “misappropriation” presumes that the Defendants’ misconduct

30 For example, the District Court found that Bundy and Synova were “solicit[ing] confidential information regarding [Mallet’s] product formulations from former Mallet employee, Shane Zhou.” (J.A. at 15.) Those findings derived from an email from Mr. Bundy to Zhou asking for “the type and general formula” for ten products to “help [Mr. Bundy] size some of the bulk tanks and piping” for Synova’s production facility. (J.A. at 1894.) Those products were identified by item numbers, all of which began with “FG[,]” and the email makes no mention of Mallet. (J.A. at 1894.) Thus, there is a disconnect between Mr. Bundy’s request for information on those ten products and the District Court’s finding that, more likely than not, those ten products are linked to Mallet and the information requested on those products was confidential. There could be evidence in the record to support the Court’s finding, such as a declaration from Mallet’s Food Technology Director saying something like, “Mallet’s products are identified by FG (‘finished good[’]) and a numerical code” (J.A. at 4329), as well as record evidence suggesting that Mallet considers the requested information confidential or a trade secret. We are not insisting that the record evidence be tied tightly to every factual finding, but we cannot be left to guess at the ties. They should be made explicit. See Proofpoint, Inc. v. Vade Secure, Inc., No. 19-cv-04238, 2020 WL 836724, at *2 (N.D. Cal. Feb. 20, 2020) (“To succeed on a claim for misappropriation of trade secrets, the plaintiff must, inter alia, offer evidence that ‘specifically identif[ies] [its] trade secrets’ and ‘show[] that they exist.’” (alterations in

46 – its acquisition, disclosure, or use of information – involves a trade secret.31 18 U.S.C. § 1839(5) (defining the term

original) (quoting MAI Sys. Corp. v. Peak Comput., Inc., 991 F.2d 511, 522 (9th Cir. 1993))). 31 The Defendants would also have us insulate trade secret thieves from liability as long as reverse engineering of a secret was hypothetically an available alternative to access the trade secret information. We decline to do so. The DTSA expressly addresses the relationship between reverse engineering and trade secret misappropriation. And it excludes reverse engineering from the type of conduct it defines as misappropriation. 18 U.S.C. § 1839(6). But while “reverse engineering is a defense to misappropriation of [a] trade secrets claim, the possibility that a trade secret might be reverse engineered is not a defense.” Bal Seal Eng’g, Inc. v. Nelson Prods., Inc., No. 8:13-cv-1880, 2018 WL 4697255, at *4 (C.D. Cal. Aug. 3, 2018) (internal quotation marks and citation omitted); see also Imperial Chem. Indus. Ltd. v. Nat’l Distillers & Chem. Corp., 342 F.2d 737, 743 (2d Cir. 1965) (“It is no defense in an action of this kind that the process in question could have been developed independently, without resort to information gleaned from a confidential relationship.”). To hold otherwise would fly in the face of commonsense and allow a defendant to escape liability for unlawfully stealing trade secrets as long as someone might – hypothetically and at unknown cost in time, effort, and money – figure out some means to discover them through reverse engineering. There may be situations in which reverse engineering is so straightforward that the distribution of a product is itself akin to a disclosure. That kind of situation is, we believe, addressed in our comment in SI Handling Sys., Inc. v. Heisley, that

47 “misappropriation” as “acquisition of a trade secret of another by a person who knows or has reason to know that the trade secret was acquired by improper means” or “disclosure or use of a trade secret of another without express or implied consent” (emphases added)). So, while the Defendants’ conduct appears deceitful, it will not support preliminary injunctive relief on a misappropriation claim under the DTSA unless the supposed trade secrets are adequately identified and there is some evidence tying the Defendants’ conduct to the taking of those trade secrets. A plaintiff need not have direct evidence tying each trade secret to a defendant’s acquisitive conduct. The Defendants’ actions here, plus their access to what may be trade secret information, plus the accelerated launch of Synova products may easily be sufficient circumstantial evidence to support a likelihood of success on the merits of Mallet’s misappropriation claim. See Oakwood, 999 F.3d at 912-13. The record, even as it stands now, may establish the requisite likelihood that the Defendants engaged in conduct that would be considered misappropriation of trade secrets, if the relevant information is identified with sufficient specificity. See supra Section I.A.2. But that identification has not yet occurred.

“[m]atters which are fully disclosed by a marketed product and are susceptible to ‘reverse engineering’—i.e., ‘starting with the known product and working backward to divine the process which aided in its manufacture,’—cannot be protected as trade secrets.” 753 F.2d 1244, 1255 (3d Cir. 1985) (quoting Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470, 476 (1974)). But short of that factual scenario, the mere potential for reverse engineering with unlimited resources does not foreclose the existence of a trade secret.

48 B. Should the District Court decide on remand that preliminary injunctive relief is warranted, the injunction must be sufficiently specific in its terms and narrowly tailored in its scope.

Just as the lack of specificity regarding the claimed trade secrets deprives us of a meaningful opportunity to review the decision to grant injunctive relief, see supra Section II.A, the injunction order’s absence of specific terms likewise inhibits our review of the order itself. “Unless the trial court carefully frames its orders of injunctive relief, it is impossible for an appellate tribunal to know precisely what it is reviewing.” Schmidt, 414 U.S. at 477.

“Since an injunctive order prohibits conduct under threat of judicial punishment, basic fairness requires that those enjoined receive explicit notice of precisely what conduct is outlawed.” Id. at 476. A complication here is, of course, the need to maintain the secrecy of the plaintiff’s trade secrets.32

32 That complication, however, is not unique to injunction proceedings. It is one courts commonly resolve in cases involving trade secrets and other highly sensitive information, utilizing confidentiality orders that restrict access to such information to trial counsel and other persons on a need-to-know basis. See Fed. R. Civ. P. 26(c); DeRubeis, 244 F.R.D. at 682 (“[A] protective order is in effect in this case which allows the parties to limit confidential information to attorneys’ eyes only. Therefore, [the party’s] identification of its trade secrets should not risk the information’s confidentiality.”); Bailey v. Dart Container Corp. of Mich., 980 F. Supp. 560, 583 (D. Mass. 1997) (weighing a defendant’s

49 Nevertheless, an injunction ought not leave too much “guesswork” for the Defendants “to determine if [they are] engaging in activities that violate the injunction”; it cannot be “little more than a recitation of the law.” Patriot Homes, 512 F.3d at 415; see also Corning Inc. v. PicVue Elecs., Ltd., 365 F.3d 156, 157-58 (2d Cir. 2004). Such vagueness “places an unduly harsh burden on the defendants … and is not in accord with the mandate of Rule 65 that a preliminary injunction shall be in specific terms.” E.W. Bliss Co. v. Struthers-Dunn, Inc., 408 F.2d 1108, 1114 (8th Cir. 1969) (internal quotation marks and citations omitted). “Rule 65(d) reflects Congress’[s] concern with the dangers inherent in the threat of a contempt citation for violation of an order so vague that an enjoined party may unwittingly and unintentionally transcend its bounds.” Sanders v. Air Line Pilots Ass’n, Int’l, 473 F.2d 244, 247 (2d Cir. 1972) (citing Int’l Longshoremen’s Ass’n, Local 1291 v. Phila. Marine Trade Ass’n, 389 U.S. 64, 76 (1967)).

The description of the conduct enjoined should be narrowly tailored to reach only those acts that closely relate to

right “to examine relevant evidence against the right of [the plaintiff] to protect his confidential information” and concluding that “[a] protective order with [a] provision requiring in house counsel and experts to sign an affidavit agreeing to be bound by the terms of the order sufficiently protects [the plaintiff’s] interest while allowing [the defendant] access to relevant information”). We have confidence that the District Court and the parties can craft an order that strikes an acceptable balance between specifying forbidden action and protecting trade secrets.

50 the unlawful conduct giving rise to an entitlement to injunctive relief.

We must protect that which is protectable, but, in doing so, we must limit the use of injunctive relief to situations where it is necessary to prevent immediate and irreparable injury. The dramatic and drastic power of injunctive force may be unleashed only against conditions generating a presently existing actual threat[.]

Holiday Inns of Am., Inc. v. B & B Corp., 409 F.2d 614, 618 (3d Cir. 1969).

The injunction order here is problematic because it extends the scope of the injunction to reach what appears to be lawful conduct. For example, a total production ban against the Defendants is not supported on the present record. “A ‘production injunction’ … completely bars [a] defendant from manufacturing the type of product in which the trade secret is utilized” and should only be imposed “when a trade secret is ‘inextricably connected’ to the defendant’s manufacture of the product[.]” Christopher M’s Hand Poured Fudge, Inc. v. Hennon, 699 A.2d 1272, 1277 (Pa. Super. Ct. 1997) (citations omitted). The record here shows that several formulas for baking release agents are the subject of patents and other publications, so it is unclear how the Defendants’ manufacture of a baking release agent will, of necessity, be inextricably entwined with Mallet’s trade secrets, since there are publicly available alternatives that the Defendant can utilize to manufacture such a product.

51 In addition, the production ban was not limited to just manufacturing but also restrained the Defendants from “directly or indirectly formulating, manufacturing, distributing or selling products competitive to Mallet products[.]” (J.A. at 44.) It would take a truly extraordinary showing – one not made here – to justify an order ejecting a competitor from the marketplace altogether.33 Injunction orders should not restrain competitors from engaging in lawful business activities.34 See

33 And here, the breadth of the injunction restrains Defendants from producing or selling release agents not only in the commercial baking industry, the only industry in which Mallet competes, but in any industry involving release agents. (See J.A. at 56.) 34 See Bhd. of R.R. Carmen of Am., Local No. 429 v. Chi. & N.W. Ry. Co., 354 F.2d 786, 800 (8th Cir. 1965) (“By its terms the injunction goes far beyond enjoining just the conduct of the minor dispute giving rise to it and thus is in violation of traditional concepts of judicial restraint in equity matters.”); Feathercombs, Inc. v. Solo Prods. Corp., 306 F.2d 251, 258 (2d Cir. 1962) (“We believe that although the use of such material by Solo was an indication of its intention to engage in unfair competition, Feathercombs’ grievance could be remedied by an injunctive provision which is less drastic and which would not deprive Solo of the right to legitimately market its own hair appliance. Since Solo infringes no patent by manufacturing its expandable comb, it is essential that it not be unduly curtailed from marketing it by being denied the usual promotional accoutrements of the modern market place.”); Corica v. Ragen, 140 F.2d 496, 499 (7th Cir. 1944) (“Without indicating what defendants may do in carrying on their normal business operations, the injunction prohibits them from doing

52 N.L.R.B. v. Express Publ’g Co., 312 U.S. 426, 435-36 (1941) (“This Court will strike from an injunction decree restraints upon the commission of unlawful acts which are thus [disassociated] from those [acts] which a defendant has committed.”); E.W. Bliss Co., 408 F.2d at 1114 (noting that we will find an injunction order “excessively broad and consequently invalid [if] it goes far beyond restraining unlawful conduct on the part of the defendants”).

Another example of problematic language in the injunction is the prohibition on Bowers “working, directly or indirectly, in any capacity … for … any person or entity that is competitive with Mallet.” (J.A. at 40.) Bowers was not under a noncompetition agreement, nor can we discern any other basis for saying he cannot work anywhere for anyone who might compete with Mallet in some way.

C. The bond amount must be tied to the scope of the preliminary injunction and account for the factual circumstances of the case.35

The Defendants also contend that the District Court abused its discretion when it set a $500,000 bond – less than three percent of the bond amount that they say is needed. They assert that the District Court improperly relied upon bond

acts which are usual and necessary in the business of gathering, editing, and distributing news.”). 35 “We … review district court judgments fixing the amount of an injunction bond for abuse of discretion.” Sprint Commc’ns Co. L.P. v. CAT Commc’ns Int’l, Inc., 335 F.3d 235, 239 (3d Cir. 2003).

53 amounts set in other cases without considering or evaluating the evidence the Defendants set forth in support of their bond assessment. We agree that the amount of the bond is not adequately supported by the Court’s explanation.

Before a preliminary injunction can issue, the moving party must provide “security in an amount that the court considers proper to pay the costs and damages sustained by any party found to have been wrongfully enjoined or restrained.” Fed. R. Civ. P. 65(c). Rule 65(c)’s bond requirement balances the competing interests of the adverse parties: preliminarily enjoining a defendant’s conduct prevents a plaintiff from incurring further irreparable harm, while the bond ensures at least some protection for the defendant in the event its conduct was wrongfully enjoined. The injunction bond serves as a deterrent to “rash applications for interlocutory orders; the bond premium and the chance of liability on it causes plaintiff to think carefully beforehand.” Instant Air Freight Co. v. C.F. Air Freight, Inc., 882 F.2d 797, 804 (3d Cir. 1989) (citation omitted). It is also the only recourse for a wrongfully enjoined party. Sprint Commc’ns Co. L.P. v. CAT Commc’ns Int’l, Inc., 335 F.3d 235, 240 (3d Cir. 2003); see also W.R. Grace & Co. v. Local Union 759, Int’l Union of United Rubber Cork, Linoleum & Plastic Workers of Am., 461 U.S. 757, 770 n.14 (1983) (“A party injured by the issuance of an injunction later determined to be erroneous has no action for damages in the absence of a bond.”). The bond “limits the liability of the applicant” and sets “the price [it] can expect to pay if the injunction was wrongfully issued[,]” Sprint Commc’ns, 335 F.3d at 240 (quoting Instant Air Freight, 882 F.2d at 805), since there is no other “legal or equitable means to recover against the applicant for a wrongful grant of the injunction (i.e., where the applicant does not prevail in the main action), other

54 than the bond.” Frank’s GMC Truck Ctr., Inc. v. Gen. Motors Corp., 847 F.2d 100, 103 n.5 (3d Cir. 1988).

Realistically, that means the consequences of wrongfully enjoining a defendant could be dire if a district court were to significantly underestimate the economic impact of an injunction it issues. While that risk is offset to a degree by the high burden placed on the moving party to establish that an injunction is warranted, id. at 102 (recognizing that the grant of injunctive relief is an “extraordinary remedy, which should be granted only in limited circumstances” (citation omitted)), the risk remains, especially if the scope of the injunction is far- reaching.36 District courts are therefore tasked with the

36 Mallet contends that the risk of an errant ruling is a factor courts may use in setting the bond amount, explaining that “[t]he lower the risk of an injunction being wrongfully enjoined, the more appropriate a lower bond amount.” (Answering Br. at 73-74 (citing Ark. Best Corp. v. Carolina Freight Corp., 60 F. Supp. 2d 517, 518 (W.D.N.C. 1999)). Mallet then points to the time spent litigating this case, as well as the volume of briefing and testimony leading up to the issuance of the injunction to support its assertion that there was a “limited risk” that the injunction was wrongfully issued and, therefore, that the $500,000 bond correctly accounted for that limited risk. (Answering. Br. at. 74.) We disagree. While it is true that the preliminary injunction standard’s heavy burden reduces the risk of wrongful enjoinment, the purpose of the bond is still to build in protections for the enjoined, given the risks that inherently arise from the expedited and preliminary nature of the proceedings. See Kos Pharms., Inc., 369 F.3d at 710 (recognizing that “district courts must exercise their discretion on an expedited basis in deciding whether to grant

55 responsibility of accounting for the factual circumstances of the parties and tying the scope of the injunction to the bond amount it decides to set. Because setting the bond amount – like the preceding decision to order injunctive relief – “is almost always based on an abbreviated set of facts” that “requir[e] a delicate balancing” of the adverse parties’ competing interests, the decision rests within the district court’s sound discretion. See U.S. Steel Corp. v. Fraternal Ass’n of Steelhaulers, 431 F.2d 1046, 1048 (3d Cir. 1970).

The deference we accord a district court to “mold its decree to meet the exigencies of the particular case” is not, however, unlimited. Trump v. Int’l Refugee Assistance

preliminary relief”); U.S. Steel Corp. v. Fraternal Ass’n of Steelhaulers, 431 F.2d 1046, 1048 (3d Cir. 1970) (“[T]he grant or denial of a preliminary injunction is almost always based on an abbreviated set of facts[.]”). To strip those built-in protections from the Defendants at this early stage of the litigation negates the very reason for requiring a bond and stands in tension with the standard for setting the bond amount – which measures compensable harm based on the presumption of wrongful enjoinment, not the likelihood of wrongful enjoinment. See Kos Pharm., Inc., 369 F.3d at 732 n.28. Moreover, the size of the preliminary injunction record does not negate the preliminary nature of the proceedings. We cannot assume that the size of the record indicates the extent of remaining evidence that may be uncovered during discovery, nor do we see a basis for assuming there is an inverse relationship between the size of a preliminary injunction record and the level of risk that a defendant will be wrongfully enjoined.

56 Project, 137 S. Ct. 2080, 2087 (2017) (quoting 11A C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure § 2947, at 115 (3d ed. 2013)). District courts should engage in a case-specific analysis that accounts for the factual circumstances of the parties, the nature of the case and competing harms, and the scope and potential impact of the injunction, and they should place on the record their reasons for setting a bond amount, so as to provide a meaningful basis for appellate review. Cf. Coyne-Delany Co., Inc. v. Cap. Dev. Bd., 717 F.2d 385, 392 (7th Cir. 1983) (“[T]he ingredients of a proper decision are objective factors—such as the resources of the parties, the defendant’s efforts or lack thereof to mitigate his damages, and the outcome of the underlying suit— accessible to the judgment of a reviewing court. … [T]he district court’s decision cannot stand” when “it fails to consider and evaluate the full range of factors … that would be relevant under the proper standard[.]”).

Here, the District Court’s stated reason that a $500,000 bond was “reasonable, proper[,] and sufficient” was that $500,000 was “toward the high-end of those [bonds] to have been imposed” in similar cases.37 (J.A. at 45 n.2.) And it

37 The District Court “surveyed federal court decisions in like cases, from across the country,” but there is nothing in the record or the Court’s decision indicating the number of cases surveyed, any case names or citations, or how the Court determined which decisions to include in its set of purportedly similar cases. (J.A. at 45 n.2.) So assuming that a survey of “like cases” might inform a court’s exercise of discretion in setting a bond amount, the absence of any information about the content of that survey in this case remains a problem. And in the absence of that information, we cannot conclude that the

57 rejected the Defendants request for a bond exceeding $21.5 million since that amount “seem[ed] astronomical by comparison.” (J.A. at 45 n.2.) The Court further alluded to a warning it had previously given, advising the parties “that it would ‘summarily … adopt the [proposals] it believe[d to be] most reasonable, appropriate and consistent with the Court’s [preliminary injunction] rulings.’” (J.A. at 45 n.2 (alterations in original) (quoting ECF No. 109 at 27).) And after contemplating that “a splitting-of-the-baby” strategy may have influenced the Defendants’ bond proposal, it faulted the Defendants for not heeding its warning. (J.A. at 45 n.2.) But frustration with a party, even if it may be justified, does not relieve a district court of the obligation to engage in a case- specific analysis, looking at potential harm from the injunction. Considering comparable cases – if they are truly comparable – is not inappropriate, but there still must be an analysis of the particular facts at hand. See Kos Pharms., Inc., 369 F.3d at 732 n.28. If the District Court decides to again grant a preliminary injunction on remand in this case, it should consider anew the appropriate bond amount.

III. CONCLUSION

For the foregoing reasons, we will vacate the preliminary injunction and remand for further proceedings.

enormous gap between the Defendants’ bond proposal and bond amounts in other cases means that what the Defendants proposed was unreasonable.