IN THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF DELAWARE
COUPA SOFTWARE INC, and LLAMASOFT LLC, Plaintiffs and Counter- defendants, Civil Action No. 24-1275-RGA v. OPTILOGIC, INC., DONALD HICKS, and DOES 1-10, Defendants and Counterclaimants.
MEMORANDUM ORDER Before me is Plaintiffs Coupa Software and Llamasoft’s Motion for a Preliminary Injunction. (D.I. 22). I have considered the parties’ briefing. (D.I. 57, 116, 136, and 144). I heard oral argument on August 29, 2025. For the reasons set forth below, this motion is DENIED. Coupa produces “Supply Chain Design and Planning” (“SCDP”) software. (D.I. 57 at 2). Plaintiffs argue that “several versions of Coupa’s supply chain software are copyrighted.” (id). Plaintiffs also assert that Coupa “maintains the secrecy of two customizations to the algorithms that underlie SCDP: (1) the dynamic calculation of a value known as ‘Big M,’ and (2) the ability of customers to apply percentage-based ‘last mile constraints’ to their supply chain models using the underlying methodology.” (/d.). Plaintiffs argue, “Coupa’s particular implementation of these concepts” qualify as trade secrets. (/d.). Defendant Donald Hicks founded Llamasoft, a company focused on developing supply chain software. (/d. at 3). Coupa purchased Llamasoft in 2018. (Ud). In 2018, Donald Hicks began another company, Defendant Optilogic, which developed “Atlas,” a supply chain model
which Plaintiffs argue served as the foundation of “the Cosmic Frog supply chain design solution.” Plaintiffs argue that Cosmic Frog violates its trade secrets and improperly uses its copyrighted software. (/d. at 3-7). Plaintiffs seek a preliminary injunction “to enjoin Defendants Optilogic, Inc. and Donald Hicks from using Coupa’s copyrighted software and trade secrets.” (D.I. 22 at 1). A “preliminary injunction is an extraordinary and drastic remedy, one that should not be granted unless the movant, by a clear showing, carries the burden of persuasion.” Mazurek v. Armstrong, 520 U.S. 968, 972 (1997) (internal quotation omitted) (italics in original). A movant seeking a preliminary injunction must establish “that he is likely to succeed on the merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the balance of equities tips in his favor, and that an injunction is in the public interest.” Winter v. Nat. Res. Def’ Council, Inc., 555 US. 7, 20 (2008). The Third Circuit has held, “The first two factors are the ‘most critical.’” Del. State Sportsmen’s Ass’n v. Del. Dep’t of Safety & Homeland Sec., 108 F.4th 194, 202 (3d Cir. 2024) (citation omitted), cert. denied sub nom. Gray v. Jennings, 145 §. Ct. 1049 (2025). “If both are present, a court then balances all four factors.” Jd. The Third Circuit has explained that the first factor, likelihood of success on the merits, “requires a showing significantly better than negligible, but not necessarily more likely than not.” Reilly v. City of Harrisburg, 858 F.3d 173, 179 (3d Cir. 2017). The second factor, irreparable harm in the absence of preliminary relief, requires a showing that irreparable harm is “more likely than not.” Jd. If both the first and second factors are met, “a court then considers the remaining two factors and determines in its sound discretion if all four factors, taken together, balance in favor of granting the requested preliminary relief.” Id.
At the conclusion of the preliminary injunction hearing, I denied most of the motion except to the extent it was based on the “last mile constraint” technology. (D.I. 164 at 81:13-86:7). I now summarize my previous rulings and provide a ruling on the “last mile constraint” aspect. The first factor is the likelihood of success on the merits at trial. Winter, 555 U.S. at 20. There are two legal issues in this case, copyright and trade secrets. To succeed on a claim of copyright infringement, Plaintiffs must prove “(1) ownership of a valid copyright, and (2) copying of constituent elements of the work that are original.” Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 361 (1991). Plaintiffs argue that the data schema and user interface (“UI”) of Defendant’s software infringes three of Plaintiffs’ copyrights for SCDP software. (D.I. 57 at 12). As I ruled at the conclusion of the hearing, I do not think that Plaintiffs’ data schema argument has much likelihood of success because Plaintiffs only identify 590 rows of similar schema out of more than 7,200 rows of Plaintiffs’ schema and more than 5,450 rows of Defendants’ schema.! (D.I. 164 at 82:8-25). Similarly, I found that Plaintiffs do not have much likelihood of success on their UI copyright claim because Optilogic’s UI was created by an individual who never worked for Plaintiffs and was unfamiliar with Plaintiffs’ UJ. Wd. at 83:1—9). To succeed on a misappropriation of trade secrets claim, Plaintiffs must (1) identify the material protected as a trade secret under 18 U.S.C. § 1836 and § 1839, and (2) demonstrate Defendants misappropriated the trade secrets. Oakwood Lab’ys LLC v. Thanoo, 999 F.3d 892, 905 (3d Cir. 2021). Plaintiffs argue that two algorithmic customizations, the Big M method and the last-mile constraints, qualify as trade secrets and have been misappropriated by Defendants.
understand Defendants’ argument to be that there is only a small percentage of overlap between the parties’ schema, and within that overlap, there are a limited number of ways to structure the schema for use in supply chain modeling. Thus, as I understand the argument, the similarity of the sort that Plaintiffs point out is not really indicative of anything.
(D.I. 57 at 9). As explained at the conclusion of the hearing, I do not think Plaintiffs have much likelihood of success proving that Big M qualifies as a trade secret because Big M is based on well-known supply optimization concepts, most of the factors used to create the equations in the Big M method appear in a general sense on Plaintiffs’ website, and there is a relatively small universe of whole numbers that can be used to complete the non-public aspect of the Big M equations. (D.I. 164 at 81:17-82:3). The “last mile constraint” is an algorithm in the source code that “allows customers to create a constraint [requiring] a percentage of demand [generated by the SCDP model to] be satisfied within either a specific distance or time.” (D.I. 57 at 7). Plaintiffs admit that the last mile constraint is a publicly-known feature that is discussed on their website. (D.I. 164 at □□□□□ 8). However, Plaintiffs argue that the implementation of the feature is a secret. (/d. at 12:8—-10). Both parties agree that the last mile constraint is a mathematical formula, eventually represented by the source code as an inequality, which allows a user to add a requirement that a percentage of the demand, measured in weight or “cubic,” to be satisfied within a set distance or time. (See □□□□ 57 at 7; D.I. 116 at 11.). The factors that go into this equation—percentage, the time or distance, the measurement in weight or cubic, and the overall flow—are all factors publicly available on Plaintiffs’ website. (D.I. 116 at 11 (not contested by Plaintiffs)). It is simple algebra to construct an equation using these publicly-known factors. I think it is possible that Defendants utilized the □
Free access — add to your briefcase to read the full text and ask questions with AI
IN THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF DELAWARE
COUPA SOFTWARE INC, and LLAMASOFT LLC, Plaintiffs and Counter- defendants, Civil Action No. 24-1275-RGA v. OPTILOGIC, INC., DONALD HICKS, and DOES 1-10, Defendants and Counterclaimants.
MEMORANDUM ORDER Before me is Plaintiffs Coupa Software and Llamasoft’s Motion for a Preliminary Injunction. (D.I. 22). I have considered the parties’ briefing. (D.I. 57, 116, 136, and 144). I heard oral argument on August 29, 2025. For the reasons set forth below, this motion is DENIED. Coupa produces “Supply Chain Design and Planning” (“SCDP”) software. (D.I. 57 at 2). Plaintiffs argue that “several versions of Coupa’s supply chain software are copyrighted.” (id). Plaintiffs also assert that Coupa “maintains the secrecy of two customizations to the algorithms that underlie SCDP: (1) the dynamic calculation of a value known as ‘Big M,’ and (2) the ability of customers to apply percentage-based ‘last mile constraints’ to their supply chain models using the underlying methodology.” (/d.). Plaintiffs argue, “Coupa’s particular implementation of these concepts” qualify as trade secrets. (/d.). Defendant Donald Hicks founded Llamasoft, a company focused on developing supply chain software. (/d. at 3). Coupa purchased Llamasoft in 2018. (Ud). In 2018, Donald Hicks began another company, Defendant Optilogic, which developed “Atlas,” a supply chain model
which Plaintiffs argue served as the foundation of “the Cosmic Frog supply chain design solution.” Plaintiffs argue that Cosmic Frog violates its trade secrets and improperly uses its copyrighted software. (/d. at 3-7). Plaintiffs seek a preliminary injunction “to enjoin Defendants Optilogic, Inc. and Donald Hicks from using Coupa’s copyrighted software and trade secrets.” (D.I. 22 at 1). A “preliminary injunction is an extraordinary and drastic remedy, one that should not be granted unless the movant, by a clear showing, carries the burden of persuasion.” Mazurek v. Armstrong, 520 U.S. 968, 972 (1997) (internal quotation omitted) (italics in original). A movant seeking a preliminary injunction must establish “that he is likely to succeed on the merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the balance of equities tips in his favor, and that an injunction is in the public interest.” Winter v. Nat. Res. Def’ Council, Inc., 555 US. 7, 20 (2008). The Third Circuit has held, “The first two factors are the ‘most critical.’” Del. State Sportsmen’s Ass’n v. Del. Dep’t of Safety & Homeland Sec., 108 F.4th 194, 202 (3d Cir. 2024) (citation omitted), cert. denied sub nom. Gray v. Jennings, 145 §. Ct. 1049 (2025). “If both are present, a court then balances all four factors.” Jd. The Third Circuit has explained that the first factor, likelihood of success on the merits, “requires a showing significantly better than negligible, but not necessarily more likely than not.” Reilly v. City of Harrisburg, 858 F.3d 173, 179 (3d Cir. 2017). The second factor, irreparable harm in the absence of preliminary relief, requires a showing that irreparable harm is “more likely than not.” Jd. If both the first and second factors are met, “a court then considers the remaining two factors and determines in its sound discretion if all four factors, taken together, balance in favor of granting the requested preliminary relief.” Id.
At the conclusion of the preliminary injunction hearing, I denied most of the motion except to the extent it was based on the “last mile constraint” technology. (D.I. 164 at 81:13-86:7). I now summarize my previous rulings and provide a ruling on the “last mile constraint” aspect. The first factor is the likelihood of success on the merits at trial. Winter, 555 U.S. at 20. There are two legal issues in this case, copyright and trade secrets. To succeed on a claim of copyright infringement, Plaintiffs must prove “(1) ownership of a valid copyright, and (2) copying of constituent elements of the work that are original.” Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 361 (1991). Plaintiffs argue that the data schema and user interface (“UI”) of Defendant’s software infringes three of Plaintiffs’ copyrights for SCDP software. (D.I. 57 at 12). As I ruled at the conclusion of the hearing, I do not think that Plaintiffs’ data schema argument has much likelihood of success because Plaintiffs only identify 590 rows of similar schema out of more than 7,200 rows of Plaintiffs’ schema and more than 5,450 rows of Defendants’ schema.! (D.I. 164 at 82:8-25). Similarly, I found that Plaintiffs do not have much likelihood of success on their UI copyright claim because Optilogic’s UI was created by an individual who never worked for Plaintiffs and was unfamiliar with Plaintiffs’ UJ. Wd. at 83:1—9). To succeed on a misappropriation of trade secrets claim, Plaintiffs must (1) identify the material protected as a trade secret under 18 U.S.C. § 1836 and § 1839, and (2) demonstrate Defendants misappropriated the trade secrets. Oakwood Lab’ys LLC v. Thanoo, 999 F.3d 892, 905 (3d Cir. 2021). Plaintiffs argue that two algorithmic customizations, the Big M method and the last-mile constraints, qualify as trade secrets and have been misappropriated by Defendants.
understand Defendants’ argument to be that there is only a small percentage of overlap between the parties’ schema, and within that overlap, there are a limited number of ways to structure the schema for use in supply chain modeling. Thus, as I understand the argument, the similarity of the sort that Plaintiffs point out is not really indicative of anything.
(D.I. 57 at 9). As explained at the conclusion of the hearing, I do not think Plaintiffs have much likelihood of success proving that Big M qualifies as a trade secret because Big M is based on well-known supply optimization concepts, most of the factors used to create the equations in the Big M method appear in a general sense on Plaintiffs’ website, and there is a relatively small universe of whole numbers that can be used to complete the non-public aspect of the Big M equations. (D.I. 164 at 81:17-82:3). The “last mile constraint” is an algorithm in the source code that “allows customers to create a constraint [requiring] a percentage of demand [generated by the SCDP model to] be satisfied within either a specific distance or time.” (D.I. 57 at 7). Plaintiffs admit that the last mile constraint is a publicly-known feature that is discussed on their website. (D.I. 164 at □□□□□ 8). However, Plaintiffs argue that the implementation of the feature is a secret. (/d. at 12:8—-10). Both parties agree that the last mile constraint is a mathematical formula, eventually represented by the source code as an inequality, which allows a user to add a requirement that a percentage of the demand, measured in weight or “cubic,” to be satisfied within a set distance or time. (See □□□□ 57 at 7; D.I. 116 at 11.). The factors that go into this equation—percentage, the time or distance, the measurement in weight or cubic, and the overall flow—are all factors publicly available on Plaintiffs’ website. (D.I. 116 at 11 (not contested by Plaintiffs)). It is simple algebra to construct an equation using these publicly-known factors. I think it is possible that Defendants utilized the □
? Plaintiffs’ source code does not contain the inequality. (D.I. 164 at 13:24~14:20). Rather, the code creates the inequality when prompted. (/d.). Defendants argue that the inequality is therefore not actually used by Plaintiffs’ product and cannot be classified as a trade secret. (D.I. 166 at 11). I disagree. Plaintiffs have presented sufficient evidence to argue at the preliminary injunction stage that the inequality is used in their product. The parties’ primary dispute at this stage is whether the inequality is a trade secret at all. (See D.I. 53:4-54:18). 3] think “cubic” refers to volume. After stating “‘weight’ or ‘cubic’” in their briefing, Defendants include a declaration which references “weight or volume.” (D.I. 116 p. 11 (citing D.I. 59 at 59)). Plaintiffs never dispute this.
publicly-available information about Plaintiffs’ last mile constraint, along with generally-known information about network optimization, to creates the equation that is represented by the inequality used in Optilogic’s product. Even if Plaintiffs prove the last mile constraint inequality is a trade secret, there is little evidence that any of Defendants’ employees had knowledge of the inequality while employed by Plaintiffs. Thus, I do not think Plaintiffs have much likelihood of success on the last mile constraint trade secret claims. As I do not think Plaintiffs have much likelihood of success on any of their copyright or trade secret claims, the first factor weighs against a preliminary injunction. The second factor is the likelihood Plaintiffs will suffer irreparable harm in the absence of preliminary relief. Winter, 555 U.S. at 20. AsI ruled at the conclusion of the hearing, Plaintiffs have not made a strong case for irreparable harm in the absence of a preliminary injunction. (D.I. 164 at 83:14-84:12). While Plaintiffs are likely to suffer economic harm from direct competition or other business challenges, they have not attempted provide an estimate of the loss or much evidence explaining the irreparable harm caused by its brand. (/d. at 83:14-21). This weighs against Plaintiffs as there is no “proof of actual or imminent harm which cannot otherwise be compensated by money damages.” Frank’s GMC Truck Cir., Inc. v. Gen. Motors Corp., 847 F.2d 100, 102 (3d Cir. 1988). Plaintiffs became aware of Cosmic Frog in November of 2022, but waited approximately two years to file suit, and then waited an additional five months to file a motion for preliminary injunction. (/d. at 83:22-84:12).4 While there may be valid reasons for delay, delay generally
4 After I orally denied most of the preliminary injunction motion, I encouraged the parties to consider an expedited schedule. (D.I. 164 at 86:8-87:14). It is now more than three months later, and, based on the absence of entries on the docket, the parties have taken zero steps to move the case along.
weighs against a finding of irreparable harm. See Air Canada & Aeroplan Inc. v. Localhost LLC, 2024 WL 1251286, at *6 (D. Del. Mar. 14, 2024); Chestnut Hill Sound Inc. v. Apple Inc., 2015 WL 6870037, at 4 (D. Del. Nov. 6, 2015). Thus, the second factor weighs against a preliminary injunction. The third factor looks at the balance of equities between the parties. Winter, 555 U.S. at 20. To do this, “a court must balance the competing claims of injury and must consider the effect on each party of the granting or withholding of the requested relief.” Amoco Prod. Co. v. Vill. of Gambell, AK, 480 U.S. 531, 542 (1987). As I explained at the conclusion of the hearing, the balance of hardships weighs heavily in favor of Defendants. (D.I. 164 at 84:13-85:4). Ifa preliminary injunction is granted, the impact will go beyond maintaining the status quo and functionally allow Plaintiffs to win the case now. (/d. at 84:15-19). Ifa preliminary injunction is denied, Plaintiffs may suffer some amount of monetary harm and other losses. (/d.). On the other hand, Optilogic only has one product, and, if I were to grant the preliminary injunction, Optilogic would have no products. (/d. at 84:22—23). This would likely result in Defendants’ business closing. (Id. at 84:24-85:7). The risks of injury to Defendants are much greater than the risks of injury to Plaintiffs. Thus, the third factor weighs against a preliminary injunction. The last factor examines the impact a preliminary injunction could have on the public interest. Winter, 555 U.S. at 20. As I explained at the conclusion of the hearing, I think this factor is neutral. The law places the emphasis on the first two factors, only balancing all four if the first two factors support a preliminary injunction. Del. State Sportsmen's Ass'n, 108 F.4th at 202. Neither of the first two factors support a preliminary injunction; I do not think there is much likelihood of success on the claims, nor do I think Plaintiffs have shown irreparable harm in the absence of a
preliminary injunction. Thus, while I have explained my view of the third and fourth factors, I do not need to consider them in the overall balancing. If] did, the third factor, balance of the equities, strongly weighs against a preliminary injunction. Thus, Plaintiffs’ motion for a preliminary injunction (D.I. 22) is denied.
IT IS SO ORDERED.
Entered this day of December, 2025
United a District Judge