Schwinn Bicycle Company v. Ross Bicycles, Inc.

COFFEY, Circuit Judge. Plaintiff-appellee Schwinn Bicycle Company brought suit against defendant-appellant Ross Bicycles, Inc., alleging that Ross had taken orders for a prototype of an exercise bicycle that unfairly copied the trade dress of an exercise bicycle produced *1178 by Schwinn in violation of § 43(a) of the Lanham Act, 15 U.S.C. § 1125(a). 1 Schwinn moved for a preliminary injunction requesting that Ross be prohibited from selling its exercise bicycle, and the district judge designated a magistrate to conduct a hearing and issue proposed findings of fact as well as a recommendation for disposition of the motion pursuant to 28 U.S.C. § 636. After a nine-day hearing the magistrate recommended denial of the preliminary injunction motion. The district judge, after conducting a de novo review of the record, disagreed with the magistrate’s recommendation and granted Schwinn’s preliminary injunction motion. 2 Ross appeals the district court’s order granting the preliminary injunction. We vacate and remand. I. FACTS Schwinn Bicycle Company (“Schwinn”), an Illinois corporation, is a bicycle manufacturer whose name and reputation are widely known in the industry. Ross Bicycles, Inc. (“Ross”) likewise is a well-established bicycle manufacturer. In 1977 Lindsay Hooper (an engineer for Repco, an Australian company) approached the executive vice president of Schwinn with a prototype for a new exercise bicycle. 3 This prototype was unique because, in addition to exercising the rider’s legs like other conventional exercise bicycles, it also enables the rider to exercise his/her upper body by moving the long handle arms back and forth with his/her arms and shoulders. The prototype had air vanes (resembling the blades on an ordinary freestanding fan) mounted on a front bicycle wheel to provide resistance to the pedals and handle arms as the wheel turned which, consequently, required the rider to expend energy (“exercise”). Thus, the resistance was generated by “air displacement.” 4 See Schwinn Bicycle Co. v. Ross Bicycles, Inc., 678 F.Supp. 1336, 1339 (N.D.Ill.1988). Hooper had obtained Australian and United States patents on the “linkage” mechanism connecting the foot pedals to the handle arms which allowed the handle arms to spin the pedals and, in turn, drive the front fan wheel. Hooper, before approaching Schwinn, assigned the patents to Repco (an Australian company). In effect, Hooper’s patented linkage mechanism allowed the front fan wheel to be driven by the handle bars alone only if the rider took his feet off the pedals. Ultimately, Schwinn obtained the exclusive right to use the linkage mechanism patent and, thereafter, modified the design of Hooper’s exercise bicycle prototype making it more attractive, adding a protective cage around the large (27-inch) front wheel (the “fan”), and adding a locking device to keep the wheel and pedals from inadvertently moving when not being ridden by an exerciser. Schwinn has successfully marketed its version of the air displacement exercise bicycle since late 1978 under the trade name “Air-Dyne” and has spent well over $4 million in advertising the product between 1980 and 1986. 5 The district court trial judge determined that “[u]ntil recently, ... there were more than 100 exercise bicycles on the market” and further that “no other manufacturer sold an exercise bicycle that looks or performs like the Air-Dyne.” 678 F.Supp. at 1339. The district court was silent concerning the reason for *1179 this lack of competition, but the magistrate reported that “a major reason for the lack of competition [with the Air-Dyne] was the Hooper patent.” In 1986 Jerry Ross, President and Chief Executive Officer of Ross Bicycles, Inc., designed and developed a linkage mechanism that avoided the Hooper patent. It is uncontested that Ross actually purchased a Schwinn Air-Dyne to aid him in designing an exercycle around the patent. 6 After Ross came up with a linkage mechanism design that did not infringe on Schwinn’s patent, it constructed a prototype of its version of an air displacement exercise bicycle and displayed its early prototypes at trade shows in late 1986 and early 1987. According to the district court, Ross intentionally copied the Air-Dyne’s “basic shape, silhouette, and primary features,” 678 F.Supp. at 1342, and stated that “[although some engineering improvements were made by Ross, the basic design — apart from the gear [the linkage mechanism] — was taken from Schwinn.” Id. at 1343 n. 10. Nonetheless, there are differences between Ross’ final prototype (the “Futura”) and Schwinn’s Air-Dyne. For example, Ross’ linkage mechanism design allows the rider to hold his feet stationary on the pedals while exercising only his upper body without the pedals rotating (“free-wheeling”). Schwinn’s linkage does not allow free-wheeling; thus, the pedals must rotate to turn the wheel because Schwinn’s linkage mechanism connects the handle arms and the pedals. There are foot rests on the Air-Dyne type for the rider to place his feet while exercising his upper body. Ross’ and Schwinn’s separately patented linkage mechanism designs require that their respective locations and housings (guards) be different: The Schwinn Air-Dyne’s connecting mechanism is between the pedals and the handle arms, but the Ross Futura’s connecting mechanism is in a relatively large enclosure on the front fan-type wheel. The meters that register, calculate and display the amount of time, distance and energy expended by the rider (ergonometers) are of different shapes and sizes. Moreover, the Futura features an electronic ergonometer while the Schwinn’s ergonometer is mechanical. There are also differences in the chain covers and, in addition, the locking devices for the “fans” are of different colors. The Futura’s handle arms are somewhat longer than the Air-Dyne’s; the Futura has toe straps on the pedals; the small wheels located at the front of the exercise bicycles for transporting the entire machine from one location to another are of different size and material; and while the Air-Dyne has black markings on its base, the Futura does not. The handle arms of the two exercise bicycles are of different diameters; the space between the handle arms varies between the Air-Dyne and the Futura; and the seat used on the Air-Dyne is not only larger than on the Futura but also has a different contour. Finally, the Schwinn exercise bicycle displays either “Air-Dyne” or “Schwinn” in ten places while the Ross exercise bicycle displays the name “Futu-ra” and “Ross” or their logo in thirteen places. The 27-inch fan wheels on both exercise bikes is located on the front in the same relation to the rider’s body as on a bicycle, and the district court stated: “[t]he Air-Dyne and the Ross bike are basically identical in shape, the arms of the two machines are shaped nearly the same, the wheel and enclosing cage are of the same size and configuration, and the rear and bases of the two bikes are nearly identical. While Ross uses the color creme where Schwinn has used chrome, and chrome where Schwinn uses white, the overall white, and chrome effect of the two exercisers is also similar.” 678 F.Supp. at 1339. The district court agreed with the magistrate’s report that the style and appearance of the two machines was very similar and stated that the appearance of the two machines is “strikingly similar." Nonetheless, the district judge did not agree with the magistrate’s recommendation that the preliminary injunction be denied and *1180 instead granted the same. The district judge specifically determined that Schwinn established that the appearance (i.e., “the overall configuration” or overall design) of the Air-Dyne had acquired secondary meaning; that Schwinn was likely to prove consumer confusion between the Air-Dyne and the Futura at trial; and, finally, that Ross failed to establish it was likely to prevail at trial on its defense that the overall design of the Air-Dyne was functional and thus not protected under federal trade law. Based on these findings, the district court judge granted a preliminary injunction against Ross prohibiting it from marketing its Futura design. Ross specifically contends on appeal that the district court erred only in determining (1) Schwinn was likely to prove consumer confusion; and (2) Ross failed to establish it was likely to prevail on its “functionality” defense. II. THE PRELIMINARY INJUNCTION AND OUR STANDARD OF REVIEW This is an interlocutory appeal from the district court’s grant of a preliminary injunction. Schwinn moved for a temporary restraining order and a preliminary injunction prohibiting the sale of Ross’ exercise bicycle. 7 The district court denied the temporary restraining order and designated a magistrate to hear evidence and make a report and recommendation on the motion for preliminary injunction pursuant to 28 U.S.C. § 636(b)(1)(B). After a nine-day hearing, the magistrate recommended denial of the preliminary injunction motion because Schwinn was not likely to establish some likelihood of consumer confusion at trial, and because Ross was likely to prevail on the issue of functionality. The respective parties objected to the magistrate’s report, on separate grounds, 8 pursuant to Rule 72(b), Fed.R.Civ.P.; the district judge made a de novo determination based on the record of the magistrate’s hearing pursuant to 28 U.S.C. § 636; and granted the preliminary injunction determining, contrary to the magistrate’s recommendation, that Schwinn was likely to establish consumer confusion at trial and that Ross was not likely to prevail on the issue of functionality. We review the Memorandum and Order of the district court judge granting the preliminary injunction. 9 The “scope of judicial review of a district judge’s decision to grant or deny a preliminary injunction is limited.” American Hosp. Supply v. Hospital Prod., Ltd., 780 F.2d 589, 594 (7th Cir.1986) (citations omitted). Findings of fact are reversible only if clearly erroneous, Vaughan Mfg. Co. v. Brikam Int’l., Inc., 814 F.2d 346, 349 (7th Cir.1987), and we will reverse an order granting or denying a preliminary injunction on the facts only if the district court abused its discretion. American Hosp. Supply v. Hospital Prod., Ltd., supra 10 ; Maxim’s Ltd. v. Badonsky, 772 F.2d 388, 390 (7th Cir.1985). Stated another way: “To reverse an order granting or denying a preliminary injunction ... it is not enough that we think we would have acted differently in the district judge’s shoes; we must have a strong conviction that he exceeded permissible bounds of judgment.” Ameri *1181 can Hosp. Supply, supra, at 595. 11 On the other hand, we have cautioned trial courts that “ 7't]he granting of a preliminary injunction is an exercise of a very far-reaching power, never to be indulged in except in a case clearly demanding it ’ ” presumably because it constrains one of the party’s freedom to engage in, as in this case, non-criminal behavior. Roland Machinery Co., 749 F.2d at 389 (quoting Warner Bros. Pictures, Inc. v. Gittone, 110 F.2d 292, 293 (3d Cir.1940) (per curiam) (citing inter alia 11 Wright & Miller, Federal Practice & Procedure § 2942, at 368 (1973)) (emphasis added). “The exercise of a power so far-reaching ought to be subject to effective and not merely perfunctory, appellate review.” Id. Thus, this Court does “not simply engage in a perfunctory rubber-stamping of the district court’s decision.” Olin Water Services v. Midland Research Laboratories, Inc., 774 F.2d 303, 307 n. 7 (8th Cir.1985). Conclusions of law, however, are reviewed de novo. Lawson Products Inc. v. Avnet, Inc., 782 F.2d 1429, 1437 (7th Cir.1986). Thus, “the review of a grant of preliminary injunction is mixed. Factual determinations are reviewed under a clearly erroneous standard; legal conclusions are reviewed de novo.” A.J. Canfield Co. v. Vess Beverages, Inc., 796 F.2d 903, 905 (7th Cir.1986). This court described the legal standard for determining whether a preliminary injunction is proper in both non-mathematical and mathematical terms, but both descriptions explain the same standard. In A.J. Canfield, supra, we stated that “[t]he district court must evaluate ... [preliminary injunction] case[s] under a well-delineated four-part test; we ... review that court’s analysis under that test for factual and legal error.” Id. at 906. In Roland Machinery Co. v. Dresser Ind., Inc., 749 F.2d 380, 386-88 (7th Cir.1984), we stated the four-part test in granting injunctive relief in non-mathematical form: (1) whether the plaintiff has an adequate remedy at law; (2) whether the plaintiff will suffer irreparable harm if the injunction is not granted and whether this harm is greater than any irreparable harm to defendant if the injunction is granted; (3) whether the plaintiff has some likelihood of success on the merits and how likely that success is “because this affects the balance of harms” between the parties; and (4) the consequences of an order granting or denying a motion for preliminary injunction on the public interest. See also A.J. Canfield, supra; American Hosp. Supply, supra, at 593-94. In American Hosp. Supply v. Hospital Products, Ltd., 780 F.2d 589, 593 (7th Cir.1986) we described the standard in a mathematical fashion analyzing the probable “costs” of the parties: “[T]he one likely to be less costly can be selected ... [and we can] grant the preliminary injunction if but only if ... the harm to the plaintiff if the injunction is denied, multiplied by the probability that the denial would be an error (that the plaintiff, in other words, will win at trial), exceeds the harm to the defendant if the injunction is granted, multiplied by the probability that granting the injunction would be an error.” *1182 III. TRADE DRESS INFRINGEMENT In this case, Schwinn contends that Ross’s Futura unfairly infringes on the “trade dress” of Schwinn’s Air-Dyne under § 43(a) of the Lanham Act, 15 U.S.C. § 1125(a). This court has stated that: “[a] product’s trade dress is the overall image used to present it to its purchasers; it could thus include, to give a partial list, the product’s size, shape, color, graphics, packaging, and label. See Blau Plumbing Inc. v. S.O.S. Fix-It, Inc., 781 F.2d 604 (7th Cir.1986), and LeSportsac, Inc. v. K Mart Corporation, 754 F.2d 71, 75 (2d Cir.1985). A trademark [on the other hand] is thought of as something more specific, such as a logo.” Vaughan Mfg. Co. v. Brikam Int’l., Inc., 814 F.2d 346, 348 n. 2 (7th Cir.1987) (emphasis added). Section 43(a) of the Lanham Act creates a federal remedy against “a false designation of origin, or any false description or representation, including words or other symbols tending falsely to describe or represent the same....” Its general purpose “is to reduce the cost of information to consumers by making it easy for them to identify the products or producers with which they have had either good experiences, so that they want to keep buying the product (or buying from that producer), or bad experiences, so that they want to avoid the product or the producer in the future.” W.T. Rogers Co. Inc. v. Keene, 778 F.2d 334, 338 (7th Cir.1986) citing Scandia Down Corp. v. Euroquilt, Inc., 772 F.2d 1423, 1429-30 (7th Cir.1985). Although § 43(a) is often thought of in a more narrow trademark context, its remedies are “not limited to trademark infringement ... [i]t provides a remedy for trademark infringement only because that is a type of unfair competition, and it provides a remedy for other types [of infringement] as well. In fact, it does not [specifically] mention trademarks.” Blau Plumbing, Inc. v. S.O.S. Fix-It, Inc., 781 F.2d 604, 608 (7th Cir.1986) (emphasis added). 12 In a trade dress infringement action the manufacturer (in this case Schwinn) must establish either that its trade dress has acquired “secondary meaning” 13 or that its trade dress is a “distinctive, identifying mark.” Blau Plumbing, Inc., supra, at 608. The manufacturer must also establish that consumers will be “likely confused” as to the source of the product because of the similarity of the products’ appearance. See Service Ideas, Inc. v. Traex Corp., 846 F.2d 1118 (7th Cir.1988). Finally, as the district court stated below: *1183 "[e]ven if both of ... [the first two] requirements are satisfied ... a plaintiff [Schwinn] cannot prevail if the alleged appropriated trade dress is found to be `functional.'" Schwinn Bicycle Co. v. Ross Bicycles, 678 F.Supp. 1336, 1341 (N.D.Ill.1988). With this general background we turn and consider Ross' contentions on appeal. IV. THE LIKELIHOOD OF "CONSUMER CONFUSION" Initially Ross contends the district court erred in granting Schwinn a preliminary injunction when it determined that Schwinn was likely to prove consumer confusion at trial. In the analytical framework applicable to the grant or denial of a preliminary injunction, the district court correctly enunciated the elements that Schwinn was required to establish in order that it might prevail on its claim for trade dress infringement. 14 The district court determined that Schwinn was likely to be able to establish at trial consumer confusion between the Air-Dyne and the Futura determining that Ross' intentional copying of the Air-Dyne's trade dress created a presumption of consumer confusion and, further, that Ross' evidence presented at the magistrate's hearing was insufficient to overcome the presumption with the application of the "digits of confusion" test. Ross does not contest the district judge's finding of "secondary meaning" for the present purpose of appealing the grant of the preliminary injunction. 15 On the other hand, Ross does challenge the district judge's use of a presumption of confusion based on the Air-Dyne's secondary meaning as related to Ross' copying of the Air-Dyne. See 678 F.Supp. at 1348. Further, Ross argues that the use of the presumption tainted the district court's application of the digits (factors) of confusion test, and in any event, that the district court misapplied the digits of confusion test. A. The Presumption of Consumer Confusion The district judge relied on Processed Plastic Co. v. Warner Communications, Inc., 675 F.2d 852 (7th Cir.1982), for the proposition that a presumption of consumer confusion exists in this circuit: "[i]n instances of intentional copying the second [manufacturer] is generally presumed to have intended to create a confusing similarity of appearance and to have succeeded in doing so." Id. at 857 (emphasis added). However, we believe the district judge placed far more emphasis on this quotation than it properly deserves when he presumed a likelihood of confusion based merely upon a showing that the defendant intentionally copied the plaintiff's trade dress because the sentence preceding it stated "another important factor bearing on the likelihood of confusion issue is [junior manufacturer's] admission that it deliberately copied the ... [item] to capitalize on the popularity of ... [a] TV show." Id. (emphasis added). This language fails to expressly adopt a presumption but rather states that deliberate copying is one of the factors to be considered within the framework of the "digits of confusion test." Convincing reasoning against the *1184 use of a presumption was stated by the Third Circuit on an appeal from a judgment denying relief for alleged trade dress infringement: “Appellant maintains that the district court committed clear error by not invoking a presumption of likelihood of confusion based on defendant’s deliberate decision to copy the color of ADVIL.... We disagree. In an action for unlawful imitation of product appearance, evidence of defendant’s intent does not relieve plaintiff of its burden of proving likelihood of confusion by preponderance of the evidence. At most, defendant’s intent is a factor tending to suggest likelihood of confusion.” American Home Products v. Barr Laboratories, 834 F.2d 368, 371 (3d Cir.1987) (emphasis added). Moreover, the exclusive use of a presumption of consumer confusion without other support in granting a preliminary injunction is at odds with the general rule that “[t]he granting of a preliminary injunction is an exercise of a very far-reaching power, never to be indulged in except in a case clearly demanding it.” Roland Machinery Co. v. Dresser Industries, 749 F.2d 380, 389 (7th Cir.1984) (emphasis added). Because intentional or deliberate copying is but one of the factors “bearing on the likelihood of confusion issue,” we are of the opinion that the district judge committed error when he presumed the likelihood of confusion element from his finding that Ross intentionally copied the Air-Dyne. 16 Processed Plastic Co. v. Warner Communications, Inc., supra, at 857. While the district court has discretion to determine the evidentiary weight to be given any single factor, we do not believe that the application of a presumption is a permissable or appropriate equivalent to weighing one factor more than another and, therefore, the use of a presumption was not appropriate in this case. We know of no case law that allows for the use of a presumption in this manner. Therefore, we set aside the ruling of the district court. In cases requiring determinations concerning the likelihood of confusion, the actual and reasoned weighing of the evidence is imperative and is inherent in a meaningful exercise of discretion. It is paramount for the trier of fact to look to the totality of the circumstances and evidence and then, after careful consideration and reflection, in its own sound judgment assess the likelihood of confusion. As this court has stated: “A variety of factors may be material in assessing the likelihood of confusion ... [and] this circuit has considered several *1185 factors to be important-” McGraw-Edison Co. v. Walt Disney Productions, 787 F.2d 1163, 1167 (7th Cir.1986) (citations omitted) (emphasis added). However, “[n]one of these factors by itself is disposi-tive of the likelihood of confusion question, and different factors will weigh more heavily from case to case depending upon the particular facts and circumstances involved.” Id. (citations omitted, emphasis added). The importance of the weighing process leads us to conclude that the district court also erred as a matter of law in the manner in which it applied the “digits of confusion test.” B. The Factors (Digits) of Confusion This court has specifically delineated several important factors to be considered by the trial court under the digits of confusion test. In Wesley-Jessen, Div. v. Bausch & Lomb Inc., 698 F.2d 862, 866 (7th Cir.1983), we delineated some of the factors to be considered and commented on their use: “ ‘In determining whether the likelihood of confusion exists, courts consider such factors as the type of trademark in issue, the similarity of design, similarity of products, identity of retail outlets and purchasers, identity of the advertising media utilized, [the alleged infringer’s] intent, and actual confusion.’ ” Id. (bracketed material in original) quoting Union Carbide Corp. v. Ever-Ready Inc., 531 F.2d 366, 381-82 (7th Cir.), cert. denied, 429 U.S. 830, 97 S.Ct. 91, 50 L.Ed.2d 94 (1976). The role of the confusion test in the district court’s analysis is unclear because the judge applied the presumption of confusion in determining Schwinn’s likelihood of success on consumer confusion before addressing the merits of the case under the factors of the confusion test: “The magistrate’s finding that Ross intended to exploit and capitalize on the good will of the Schwinn Air-Dyne ... especially when viewed in conjunction with the equally well-supported finding of secondary meaning, suggests that there is more than a negligible chance that Schwinn will be able to prevail in establishing a likelihood of confusion at trial.” 678 F.Supp. at 1344-45. It at least appears that the factors of confusion may have been in balance, analyzed in the context of rebutting a presumption because “[i]n making her findings with regard to ... [the likelihood of consumer confusion], the magistrate accorded insufficient weight to a number of factors-in cluding a presumption which arises from the evidence of secondary meaning and the Air-Dyne’s distinctive appearance, coupled with the evidence of Ross’ intentional copying.” 678 F.Supp. at 1344 (emphasis added). We have previously determined the use of the presumption of consumer confusion from the finding of secondary meaning was error; consequently, the district court erred as a matter of law because the confusion test is not or “must” not be influenced by a presumption. Even without the use of the presumption, however, the district court failed to properly apply the confusion test when it found, with the exception of actual confusion, that each digit of confusion was supported in the evidence. First, the district court disagreed with the magistrate and deemphasized evidence relating to the conditions of purchase, the price of the products and brand labeling because it determined as a matter of law that cases relying on those factors lacked a separate finding of secondary meaning. See, e.g., Litton Systems Inc. v. Whirlpool Corp., 728 F.2d 1423 (Fed.Cir.1984). The district court stated in its discussion of labeling, for example, that “the fact that Air-Dyne in this case has acquired secondary meaning renders the mere fact of labeling significantly less important here than it was in Litton.” 17 *1186 It is correct that in its discussion of labeling the Third Circuit stated in IAtton that “[the district] court failed to note, in particular, that the similarity of design ..., except as to very few items, [is] a similarity shared by many microwave ovens on the market ..thus implying the ovens were neither distinctive nor had acquired secondary meaning. However, we do not interpret this language as mandating that a finding of secondary meaning automatically obviates the label analysis under the confusion element of trade dress infringement, and the IAtton court said nothing contrary to our interpretation requiring a separate analysis of the secondary meaning and consumer confusion elements. In fact, the court in Litton confined its review to the consumer confusion element and did not even consider either the secondary meaning or functionality elements indicating a distinct analysis. Id. at 1445. Moreover, the order in which the Litton court prepared its list of necessary trade dress infringement elements also supports our interpretation. The court stated: “In order to prove a violation of section 43(a) of the Lanham Act, a plaintiff must establish the existence of three elements: (1) that the trade dress or product configuration of the two competing products is confusingly similar; (2) that the appropriated features of the trade dress or product configuration are primarily nonfunctional; (3) that the trade dress or product configuration has obtained secondary meaning. Id. (emphasis added). If the order in which the elements were listed is given any import, a determination concerning secondary meaning could have no effect on the determination of the other elements because the secondary meaning element is listed last. While it is unlikely that the Third Circuit meant to imply that the elements be considered in any given order or that one element is of more significance than another, its separate analysis of the consumer confusion element emphasizes the importance of each separate element. Simply, each of the three elements must be established separately, and, therefore, labeling may be an important factual factor under the facts and circumstances of a particular case and may not, as a matter of law, be deemphasized merely because the element of secondary meaning is established. In this case the district court erred in deem-phasizing, as a matter of law, the importance of the conditions of purchase (including labeling) solely because secondary meaning had been determined. 18 Therefore, we agree with the magistrate’s reasoning and disagree with the district court that the magistrate misapplied the case law in focusing on evidence regarding “the *1187 manner in which consumers actually purchase” a product, 19 and the district court committed error in deemphasizing the labeling as a matter of law in reliance on such cases as Litton. Even so, we recognize that the same considerations supporting a factual finding of secondary meaning may mitigate against determining as a factual matter that the labels alone prohibit confusion because some of the facts supporting a finding of strong secondary meaning are the same as facts used in other cases to find consumer confusion even with unique labeling (e.g., Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254 (2d Cir.1987)). Nonetheless, these facts must be considered separately in the factual context of consumer confusion rather than being swept aside as a matter of law based on a finding of secondary meaning as the district court did in this instance. Our emphasis on the modifying word “alone” in stating that it is possible that “labels alone” may not prohibit consumer confusion in some cases belies the second problem with the district judge’s approach in applying the confusion test; that is, the serial and mechanical use of the factors as being totally independent of one another. As we stated previously it is imperative for the trier of fact to look to the totality of the circumstances and evidence and in its own sound judgment assess the likelihood of confusion. We agree with the district •court that the plaintiff need not prove each and every factor in order to prevail. E.g., McGraw-Edison Co. v. Walt Disney Productions, 787 F.2d 1163, 1168 (7th Cir.1986), quoting Marathon Mfg. Co. v. Enerlite Products Corp., 767 F.2d 214, 218 (5th Cir.1985). However, the converse is also true; neither is it required that the defendant refute each and every factor. The weight and totality of the most important factors in each case will ultimately be determinative of the likelihood of confusion, not whether the majority of the factors tilt the scale in favor of one side or the other. 20 Regarding actual confusion the district judge stated, “The law is clear that a plaintiff need not demonstrate actual confusion in order to show a likelihood of confusion.” 678 F.Supp. at 1345 (emphasis in original). Concerning the conditions of purchase the district court stated, “[i]t is well settled, however, that increased consumer care in purchasing higher-priced products does not necessarily eliminate the likelihood of confusion,” Id. (emphasis added) (the retail price of both exercycles is around $600), and further that “a label cannot altogether preclude the possibility of likelihood of confusion_” Id. at 1347 (emphasis added). Each of these statements is an accurate statement of the law. However, just because each factor may not alone or of itself be conclusive of consumer confusion does not mean their cumulative effect need not be considered. Therefore, we believe that in this case the district court failed to properly apply the consumer confusion test because it rigidly applied the factors (or “digits”) in a pedantic matter without assessing the totality of the circumstances. This is a misapplication of the test, and this misapplication constitutes an error of law. Moreover, the pedantic use of the factors foreclosed the judge from exercising his sound discretion on the facts and, therefore, it constitutes a factual abuse of discretion. In explaining the test the district court stated: *1188 “A plaintiff need not establish that each digit of confusion weighs in its favor in order to establish a likelihood of confusion. See Sicilia [Di R. Biebow & Co. v. Cox], supra [732 F.2d 417], at 434 [5th Cir.1984], Recently, for example, a likelihood of confusion was found based on the presence of seven digits, excluding only the digit of actual confusion. Chemlawn Services, supra, at 1393 [the full citation offered by the district court in a prior passage of his opinion was 652 F.Supp. 1382 (S.D.Tex.1987), rev’d on other grounds, 823 F.2d 515 (Fed.Cir.1987) ].” 678 F.Supp. at 1345 (emphasis in original). Thus, the order granting the preliminary injunction must be vacated because we have determined, for the reasons delineated above, that the district court committed three errors in interpreting and applying the confusion test to the facts and circumstances of this case: (1) The district court erred as a matter of law in its application of a presumption of likelihood of confusion; (2) This error tainted the district court’s determination that the “digits of confusion” test supported the granting of the injunction because it failed to weigh all relevant evidence; and (3) The mechanical misapplication of the test, in the absence of consideration of the totality of the circumstances, constituted both an error of law and an abuse of discretion. V. THE FUNCTIONALITY DEFENSE Ross contends that the district court erred in finding that it failed to establish it was likely to prevail at trial on its “functionality defense.” For purposes of a defense against trade dress infringement, “functional” means not simply that the feature serves a function, but that the feature is necessary to afford a competitor the means to compete effectively. We explored the contours of the functionality defense in W.T Rogers Co. v. Keene, 778 F.2d 334 (7th Cir.1985), and settled on a functional definition of “functionality”: “If the feature is ornamental, fanciful, decorative, like the patterns on a piece of china or of silverware, then the manufacturer can use it as his name, his symbol, his identifying mark. Ornamental, fanciful shapes and patterns are not in short supply, so appropriating one of them does not take away from any competitor something that he needs in order to make a competing brand. But if the feature is not ornamental or fanciful or whimsical or arbitrary, but is somehow intrinsic to the entire product consisting of this manufacturer’s brands and his rivals’ brands, trademark protection will be denied.... A firm that makes footballs could not use as its trademark the characteristic oval shape of the football, thereby forcing its rivals to find another shape for their footballs; since they wouldn’t be able to sell any round or oblong or hexagonal footballs, that firm would have, not an identifying mark, but a product monopoly, and a product monopoly not for a term of years as under the patent laws but forever. “The football’s oval shape is ‘functional’ in the following practical sense: it would be found in most or all brands of the product even if no producer had any desire to have his brand mistaken for that of another producer.... To put this differently, a functional feature is one which competitors would have to spend money not to copy but to design around, as they would have to do if they wanted to come up with a nonoval substitute for a football. It is something costly to do without (like the hood [of a car]), rather than costly to have (like the statue of Mercury [on the hood of a Rolls Royce]).” So the shape of the Schwinn machine is functional if it is something Ross must pay extra to design around. The application of the “functionality” test is complicated by different verbal formulations, sometimes appearing in the same opinions that recite the approach of W.T. Rogers. For example, our opinion in Vaughan Manufacturing Co. v. Brikam International, Inc., 814 F.2d 346, 350 (7th Cir.1987), quoted the central language of W.T. Rogers and immediately followed with this language from Sicilia Di R. Biebow & *1189 Co. v. Cox, 732 F.2d 417, 429 (5th Cir.1984): “To achieve the status of ‘functional,’ a design or feature must be superior or optimal in terms of manufacture, or accommodation of utilitarian function or performance.” Service Ideas, Inc. v. Traex Corp., 846 F.2d 1118, 1123 (7th Cir.1988), then quoted the language from Cox as if it were the holding of Vaughan and followed with a paraphrase from W.T. Rogers. This would not be troubling if “superior or optimal” were the same thing as “something costly to do without,” but it is not. “Superior” could mean something that is costly to do without, but “optimal” implies that unless the feature in question is the best possible way to achieve a result, it is not “functional.” As the copier bears the burden of showing that the feature is functional, the burden of showing that a competitor’s design is the “best possible” could extend the scope of trademark protection into the domain that, we held in W.T. Rogers, is within the bounds of lawful competition. In addition, our reference to an “essential” feature further complicates the functionality test by creating the additional task of attaching an unusual meaning to a common term. See Service Ideas, Inc., 846 F.2d at 1123. We therefore reiterate the approach of W.T. Rogers, which this circuit adopted after considering the views of other courts of appeals. A feature is functional if it is one that it costly to design around or do without, rather than one that is costly to have. To the extent the district court believed that the junior user must demonstrate that even after spending substantial sums it could not duplicate the utilitarian features of the Air-Dyne without duplicating aspects of its appearance, it was led astray by words such as “optimal,” “superior” and “essential” that are not integral to this circuit’s approach. The district court determined that Ross was unlikely to prove its functionality defense at trial because: “Ross’ president, who designed the Ross exerciser, admitted from the beginning, he never considered using any alternative design other than the one that duplicated the appearance of the Air-Dyne ... that it conducted virtually no engineering studies, that it prepared no technical drawings during the development of its initial prototypes, that it conducted no marketing studies, that it made no study of manufacturing costs, and that it did not even calculate its cost of production.” 678 F.Supp. at 1349 (emphasis in original). 21 The district court then delineated three important areas of design that Ross failed to establish which would likely be found “functional” (i.e., for effective competition) at trial: (1) the size of the wheel used; (2) the nature and composition of the wire wheel guards; and (3) the overall design. 678 F.Supp. at 1349-52. There is some confusion as to exactly what Schwinn claims as trade dress in this case, and, consequently, what elements of the design Ross must establish are essential to produce a competitive exercise bicycle (i.e., what design elements are “functional”). In this regard the district court stated: “Schwinn’s claimed trade dress appears to relate primarily to the front-end of the bike (comprising the large bicycle wheel fan run by the wire fan guard or cage, with the long handle arm levers that separately move back and forth). Schwinn’s counsel conceded at the hearing that the rear frame of the Air-Dyne is not particularly distinctive, and indeed, the back end parts of the bike have been used for a number of years on other exercise bikes made by Schwinn, Ross and several other competitors.” 678 F.Supp. at 1340 (emphasis added). Similarly, the magistrate stated: *1190 "It is conceded in this case that the rear position of the Air-Dyne and the base are in the public domain, or otherwise are so commonly used on bicycles or exercisers that Schwinn cannot prevent Ross from utilizing these factors-which both companies have used on other exercisers-on its new exerciser." Nonetheless, on appeal Schwinn argues that "both the front and rear portions of the Air-Dyne are part of the `overall configuration' ... [the] `overall look' which Schwinn seeks to protect." Schwinn's argument in this regard misses the mark because it is well established that trade dress protection does not extend to features already shared by different brands on the market. 22 Moreover, the district court determined that the rear end design and mechanism have been used by Schwinn and its competitors on numerous other exercise bicycles for years. This finding of fact is certainly not clearly erroneous, and Schwinn does not allege any mistake of law specifically relating to the district court's finding that the design of the back portion of the Air-Dyne is "not particularly distinctive." Schwinn cites Hartford House Ltd. v. Hallmark Cards, 846 F.2d 1268 (10th Cir.1988) as support for seeking protection of the Air-Dyne's overall design including the back end. In Hartford House, the Tenth Circuit held that a trade dress "`claim may still be made without a particular identifying design feature.'" Id. at 1273 quoting Brunswick Corp. v. Spinit Reel Co., 832 F.2d 513, 519 (10th Cir.1987). However, in the present case, the district court determined the front portion of the Air-Dyne is its primary design feature, and its existence distinguishes the present case from Hartford House. Thus, it is appropriate to look to the front end design (the particular identifying design feature) to determine wheth-er it is functional and the district court's focus on the "distinctive" front end design was proper. 23 Therefore, we review the question of whether the district court erred in determining that the distinctive feature was not functional. We review the district court's factual determinations under the clearly erroneous standard and its findings of law de novo. Ross bears the burden of proving the functionality of the design. Vaughan Mfg., supra, at 350. The district court's analysis in determining that the front end design of Schwinn's Air-Dyne was not functional because Ross failed to meet its burden of demonstrating the unavailability of alternate configurations that would have avoided the virtual copying of Schwinn's trade dress was flawed in two respects: (1) It erred as a matter of law in not considering whether the design and placement of the large front wheel was aesthetically functional; and (2) It is not entirely clear whether the district court considered the ultimate question of whether Ross would be able to "compete effectively" using any design other than Schwinn's Air-Dyne design. The magistrate considered "the attractiveness of the Air-Dyne and its popularity with consumers as evidence of functionality" when recommending that both the 27-inch wheel and the wire wheel guard were functional. See 678 F.Supp. at 1352 n. 25. The district court did not even consider these aesthetic aspects of the design in determining that the designs were not functional, stating: "Some courts have considered the aesthetic appeal of a product to be one of the indicia of functionality.... The Seventh Circuit, however, has not adopted this approach." Id. The district court then quoted W.H. Brady Co. v. Lem Products, Inc., 659 F.Supp. 1355, 1365 (N.D.Ill.1987), as support: *1191 “ ‘[a] feature may serve an aesthetic function, making the product more pleasing to the consuming public which is not indifferent to such things. However, the fact that a feature makes a product more attractive does not automatically mean that the feature is “functional.” ’ W.H. Brady Co., supra, at 1365, citing W.T. Rogers, supra, at 343.” Id. (bracketed material in original, emphasis added). We add that neither does the fact that a feature makes a product more attractive automatically mean that it is nonfunctional and therefore subject to trade dress protection. It was an error of law for the district court to ignore the “pleasing” aspect in determining the functionality of the overall design because we have stated that attractiveness may, at some point, become functional: “Though a producer does not lose a design trademark just because the public finds it to be pleasing, there may come a point where the design feature is so important to the value of the product to consumers that continued trademark protection would deprive them of competitive alternatives; and at that point protection ceases.” W.T. Rogers, supra, at 347 (emphasis added). Thus, Ross carries the heavy burden of proving that Schwinn’s design is so attractive that trade dress protection deprives consumers of competitive alternatives. Whether it carried this burden is a question of fact that the district court failed to address and its failure is an error as a matter of law. We also must reiterate that effective competition is the standard used to determine functionality because in determining that the design was not functional the district court emphasized Ross’ failure to establish that other competitive designs were not available: “The evidence presented by Ross at the hearing largely ignored the evidence of available design alternatives introduced by Schwinn. The possibility that alternative configurations may exist that are feasible suggests that the Air-Dyne’s trade dress is not entirely functional, and that Ross did not have to copy it in order to compete effectively in the market.” The district court also stated that Ross failed “to introduce any cost data from which any comparative findings between the Air-Dyne and the Ross could be made_” 678 F.Supp. at 1351. The relevant cost comparison, however, is not between the Air-Dyne and the present Futu-ra design but between the Air-Dyne and any alternative designs. This comparison is merely evidence to be used to assess whether any manufacturing cost differential translates to a price differential to the consumer and, in turn ultimately, whether the price differential prohibits the junior competitor (in this case Ross) from effectively competing in the marketplace. The fact that Ross failed to introduce cost studies may very well be determinative of whether Ross has sustained its burden on the question of functionality; however, the district court did not analyze this fact within the proper analytical structure whose ultimate focus is on effective competition. The district court’s failure to do so is an error of law even though it may not be outcome determinative in this case. YI. For the reasons stated, we have determined that the district court erred as a matter of law in presuming a likelihood of confusion based on its factual determination of secondary meaning and that this error tainted the district court’s determination that the “digits of confusion” test supported the granting of the preliminary injunction. Moreover, the mechanical application of the confusion test constituted both an error of law and an abuse of discretion. The district court also erred as a matter of law in not considering whether the appearance of the Air-Dyne was functional. Finally, the district court’s reasoning based on the lack of comparative design cost studies is flawed because it focused on Ross’ failure to research alternative designs and conduct cost studies rather than the required ultimate question of whether Ross would be able to “effectively compete” if it used an alternative design. Therefore, we vacate the order granting the preliminary injunction and remand the *1192 matter to the district court for reconsideration consistent with this opinion. The order' is VACATED AND REMANDED. Footnotes [1] Schwinn's complaint also alleges violation of state trade law including Ch. 140 Ill.Rev.Stat. (the Illinois Anti-Dilution Statute) § 22; Ch. 121‘/2 Ill.Rev.Stat. (the Illinois Deceptive Trade Practices Act) § 312; and common law trademark protection. The only claim subject to the immediate appeal is the Lanham Act claim. [2] The district court’s memorandum opinion and order is reported at Schwinn Bicycle Co. v. Ross Bicycles, Inc., 678 F.Supp. 1336 (N.D.Ill.1988). For a more detailed discussion of 28 U.S.C. § 636 see section II of this opinion. [3] "Exercise bicycles” are also commonly known as "exercycles” or "stationary bicycles.” This nomenclature is used interchangeably herein. [4] The air vanes (like fan blades) "pushed air.” This air movement provides resistance. Hooper’s prototype did not have a protective cage around the fan wheel on which the air vanes were mounted. Schwinn later enclosed the fan wheel with a protective wire cage (similar to the type of wire cage surrounding a fan). [5] Sales of Air-Dyne have increased steadily; in 1978 Schwinn sold 91 units; in 1986 it sold 66,655. [6] Ross has also patented its linkage mechanism. [7] The denial of the temporary restraining order is not contested on appeal. [8] Ross’ objection to the report was limited to the admission of survey evidence for purposes of establishing secondary meaning and consumer confusion and would not have changed the outcome of the magistrate's recommendation. The district court’s resolution of the survey evidence issue is not on appeal. [9] It is the district court’s memorandum opinion, not the magistrate’s proposed findings and recommendation, that serves as the findings of fact and conclusions of law in this case pursuant to Rule 52(a), Fed.R.Civ.P. In United States v. Raddatz, 447 U.S. 667, 681-82, 100 S.Ct. 2406, 2415, 65 L.Ed.2d 424, rehearing denied, 448 U.S. 916, 101 S.Ct. 36, 65 L.Ed.2d 1179 (1980), the Supreme Court explained § 636 stating: " ‘[t]he authority — and the responsibility — to make an informed final determination ... remains with the judge.’ ” Id. (editing in original) (quoting Mathews v. Weber, 423 U.S. 261, 271, 96 S.Ct. 549, 554, 46 L.Ed.2d 483 (1976)). [10] We stated that the abuse of discretion standard “covers a family of review standards rather than a single standard, and a family whose members differ greatly in the actual stringency of review." American Hosp. Supply, 780 F.2d at 594 (citations omitted). [11] In both Roland Machinery Co. v. Dresser Indus., 749 F.2d 380, 390, rehearing denied (7th Cir.1984), and American Hosp. Supply, supra, at 595, we stated that one of the reasons deference is given to the district judge’s finding is his ability to observe the witnesses and the parties. For example, in Roland we stated: "Although, considering the nature of the judge’s determination and the stakes to the parties, we do not think the term abuse of discretion can be limited to cases where the judge can be said to have acted irrationally or fancifully (Delmo), we also do not think, considering the imponderable character of the balancing process and the judge’s superior feel for the issues which a cold transcript may not fully communicate to the reviewing court, that we are entitled to substitute our judgment for the district judge’s." Id. (citations omitted) (emphasis added). We note that in the case at hand the record consists primarily of the magistrate’s hearing and, therefore, the district court judge did not have the full advantage of a "superior feel” of the facts based on the hearing. However, the 1985 amendment to Rule 52(a), Fed.R.Civ.P., "made clear that a district court’s conclusions cannot be regarded as pertaining to a ‘question of law’ —and thus reviewed de novo — merely because they are based on evidence that is theoretically susceptible of independent evaluation by an appellate court." American Home Products v. Barr Laboratories, 834 F.2d 368, 370 (3d Cir.1987). [12] As one commentator has stated: "The emphasis and thrust of trademark protection is deciding whether a given symbol in fact functions as a mark and whether defendant's mark is likely to cause confusion. Thus, for trademark infringement purposes, it is only the symbol characterized as a "trademark” which is juxtaposed against defendant’s usage. On the other hand, unfair competition protection is not so limited in scope. The court need not focus on merely one facet of plaintiffs total selling "image,” as in trademark law. To determine unfair competition, the court must consider the total image of plaintiffs product, package and advertising and compare this with defendant’s image. If defendant’s trade dress is likely to cause confusion with plaintiffs trade dress, then a finding of unfair competition is warranted. The ultimate protection desired — prevention of a likelihood of confusion of buyers — is just as important when a junior user presents a package which in toto is confusingly similar to plaintiffs, as when a mere part of the package (such as a word mark) is simulated.” McCarthy, Trademarks and Unfair Competition (2d ed.), § 8:1 p. 282-283 (1984). [13] As we stated in Vaughan Mfg. Co., supra, at 348: “ ‘Secondary meaning’ denotes an association in the mind of the consumer between the trade dress of a product and a particular producer. Secondary meaning need not be shown where the trade dress, by itself, is a distinctive, identifying mark, but must be shown where, as here, the trade dress is not inherently distinctive but is claimed to have become identified with the producer over time." (Citations omitted). Some courts have employed a presumption of "secondary meaning” where the second manufacturer of a product deliberately and closely copies the trade dress of the first manufacturer of a product. We have not gone so far; rather, proof of intentional copying is probative evidence of “secondary meaning.” Id. at 349. [14] The district court properly noted that in order for a manufacturer to prevail on a claim for trade dress infringement, it must first establish that its trade dress is "inherently distinctive" or that it has acquired "secondary meaning" to consumers; the manufacturer must also show some "likelihood of confusion" on the part of consumers as to the source of the product; finally, even if both of these requirements are satisfied, a manufacturer cannot prevail if the alleged appropriated trade dress is found to be "functional." Each of these elements-secondary meaning, likelihood of confusion, and functionality-must be separately considered. 678 F.Supp. at 1341. [15] "The secondary meaning denotes an association in the mind of the consumer between the trade dress of a product and a particular producer ... [it] must be shown where, as here, the trade dress is not inherently distinctive but is claimed to have become identified with the producer over time." Vaughan Mfg., 814 F.2d at 348. Neither does Ross assign error in the finding that "the public interest will not be disserved by the issuance of a preliminary injunction." Schwinn Bicycle Co. v. Ross Bicycles, Inc., 678 F.Supp. at 1340 (the fourth factor of the four-part preliminary injunction test). [16] The use of a presumption is particularly troubling in an instance such as this, where the product that has acquired secondary meaning is the first of its type (i.e., the first air displacement exercise bicycle) to be marketed; this seems to be a presumption in favor of product monopoly rather than the protection of trade dress. We recognize, however, that this concern is somewhat ameliorated due to the fact that the functionality defense allows manufacturers to copy functional features from competitors’ products if the feature is one that the manufacturer would find necessary to incorporate into its product in order to compete effectively. Vaughan Mfg., 814 F.2d at 349. The functionality defense is discussed in more detail in section V. of this opinion. Additionally, we note that the use of this presumption borders on a "presumption on a presumption.” We have not directly addressed whether there is a presumption of secondary meaning "where a junior competitor deliberately and closely copies the trade dress of a senior competitor in the market.” Vaughan, supra, at 349. However, we have stated that we at least are in accord that intentional copying is probative evidence. Id. Here the district court judge determined that there was a likelihood of establishing secondary meaning. 678 F.Supp. at 1342 (footnotes omitted). At the very least this puts heavy, if not determinative, evidentiary weight on intentional copying for two of the three elements necessary to prove trade dress infringement (i.e., secondary meaning and consumer confusion), and the defendant bears the burden of establishing the third (the functionality defense). Moreover, this is particularly important in the context of a preliminary injunction interlocutory appeal because two of the four requirements necessary for the issuance of a preliminary injunction are also presumed. “In this case as in the usual trade dress case, lack of adequate remedy at law and irreparable harm are presumed.” Id. at 1340. In other words, if there are presumptions of secondary meaning based on intentional copying and of confusion based on secondary meaning, most of the plaintiffs required proof for both trademark infringement and for preliminary injunction can be accomplished by simply filing suit and proving intentional copying. [17] In discussing brand name labeling the district court noted: "Although Ross introduced evidence that brand name was important with respect to the sale of bicycles-which basically all look alike -exercisers, unlike bicycles, vary dramatically in appearance, making brand name identification considerably less important." *1186 678 F.Supp. at 1348 n. 19. We do not know the level on which the district court is making the comparison. Many brand name manufacturers probably do maintain competitive classes of bicycles (e.g., 10-speed and single gear racing bikes, multiple-speed mountain bikes, bikes for young children with training wheels, small bikes used for trick riding, bicycles built for two, and bikes that look like they could have been ridden out of the television show “Leave It to Beaver”). We also do not know as a factual matter whether there are classes of exercycles which would make the comparison valid. Additionally, as the listed classes indicate, it would be highly unusual and error to find that all bicycles, regardless of their competitive class, look alike. However, even if the district court erred in this comparison, we refuse to speculate what effect that error had on the ultimate findings of fact in this case. [18] 678 F.Supp. at 1348 n. 19. We also note that the district court’s reliance on Chemlawn Services v. GNC Pumps Inc., 652 F.Supp. 1382, 1393 (S.D.Tex.1987), rev’d 823 F.2d 515 (Fed.Cir.1987), is at best questionable. The Chemlawn case was reversed by the Federal Circuit Court of Appeals because: “[F]irst, the court failed to issue findings of fact and conclusions of law with its preliminary injunction as required by Rule 52(a). Second, the court attempted to cure that error at a time when it no longer had jurisdiction over the case because GNC had already filed an appeal to this court." 823 F.2d at 517 (emphasis added). Thus, the opinion cited by the district court (652 F.Supp. 1382) was issued absent jurisdiction and was not reviewed by the appellate court. The Federal Circuit explained: "The findings of fact and conclusions of law are improper because they were entered at a time when the District Court no longer had jurisdiction over the case.... It would be improper to accept, in the name of economy of judicial resources, these invalid determinations. ...” Id. at 518 (emphasis added). [19] We agree with the district court that ultimately: “The test for [trade dress] infringement is generally stated to be ‘likelihood of confusion’ of ordinary purchasers purchasing in the ordinary manner.... The test is not simply a ‘side-by-side one', made by the court through a personal comparison, but rather is one of consumer confusion, in light of the manner in which consumers purchased these products.” 678 F.Supp. at 1346 (citations omitted). [20] In McGraw-Edison Co. v. Walt Disney Productions, 787 F.2d 1163, 1167 (7th Cir.1986) we stated "different factors will weigh more heavily from case to case depending upon the particular facts and circumstances involved." (Citations omitted) (emphasis added). This quotation supports the fact that the importance of any single factor may vary between cases. We also note that if the factors themselves are applied without thoughtful analysis and precision, they become virtually meaningless if not damaging to the integrity of the ultimate determination. [21] The district court’s conclusion was again contrary to the magistrate’s recommendation because it determined that the rationale for the locations of features of the Ross exercise bicycle had little to do with whether Ross had to copy the Schwinn Air-Dyne trade dress in order to complete effectively. The magistrate recommended "[t]hat, apart from color, the features of the Air-Dyne allegedly copied by Ross are likely to be found functional.” [22] See Service Ideas, Inc. v. Traex Corp., 846 F.2d 1118, 1123 (7th Cir.1988); WT. Rogers Co., Inc. v. Keene, 778 F.2d 334, 339 (7th Cir.1985). [23] It is true that had Ross' wheel and wheel placement been designed differently, it might have altered the appearance of the "back-end"; however, the back-end design utilized by Ross is not the problem here because it is commonly used by other manufacturers. An alternative analysis that reaches the same conclusion is that Schwinn may claim the "entire design" but the "back-end" design is not protectible because it is commonly used on other exercise bicycles; therefore, the trier of fact is left with determining if the remaining design features (of the front end) are functional.