OPINION OF THE COURT

FISHER, Circuit Judge.

This appeal requires us to decide when the trade dress on the packaging of store-brand products is so similar to that of directly competing national-brand products as to create a likelihood of confusion among consumers. The plaintiff-appellant McNeil Nutritionals, LLC (McNeil) sells and markets Splenda, a highly successful national brand of sucralose, an artificial sweetener. The defendants-appellees Heartland Sweeteners LLC and Heartland Packaging Corp. (collectively, Heartland) package and distribute sucralose as store brands to a number of retail grocery chains. McNeil brought a trade dress infringement action against Heartland, alleging that Heartland's product packaging is confusingly similar to Splenda's. McNeil also filed a motion for preliminary injunction to enjoin Heartland from advertising, selling, or distributing the allegedly infringing products. The District Court denied the motion, and this appeal followed. For the reasons that follow, we will affirm the denial of the motion in part and reverse in part and remand to the District Court.

I

A. Presence and Effect of Private-label and Store-brand Products

Private-label products are typically manufactured by one company for sale to the consuming public under another company's name. Such products are generally made with the same active ingredient as the name-brand or national-brand product with which the private-label product competes. Private-label products are available in a wide range of industries and are often positioned as lower cost alternatives (about 25% less expensive) to national-brand products. As of 2005, private-label sales represented 20% of all U.S. supermarket, drug chain, and mass merchandiser sales and totaled $50 billion.

Store-brand products are a type of private-label product, in which the store or retail chain name is the brand name. Store-brand products have existed in retail chains since 1883, and consumers have become highly aware of them. Indeed, the Private Label Manufacturers Association reported that in 2005, more than 90% of consumers polled were familiar with store brands, and 83% bought them regularly.

Store brands are typically found on store shelves next to the analogous national-brand products. The packaging of store-brand products often includes reference points to invite the consumer to compare them to the national-brand ones. These reference points often include similar product packaging and "compare to" statements on the packaging. Stores also employ tags on store shelves (so-called shelf extenders or shelf talkers) that explicitly invite consumers to compare the store-brand product with the national-brand analog.

Consumers are generally aware of the name of the store in which they are shopping. They are aware that stores have private-label brands that in most cases are merchandised next to the national-brand products. Prices for the products are typically displayed prominently. Consumers can, therefore, see the cost difference between store brands and national brands.

B. Emergence of No-calorie Sweeteners

American consumers spend between $600 and $700 million annually on sugar substitutes, also known as artificial sweeteners. No-calorie sweeteners are a subset of artificial sweeteners. The market for no-calorie sweeteners is dominated by products that contain one of three sweetening ingredients: saccharin, aspartame, or sucralose. Saccharin was first marketed in the United States in 1957; the leading brand of artificial sweetener containing saccharin is Sweet `N Low. In 1982, the FDA approved aspartame for sale in the United States; the leading brand of artificial sweetener containing aspartame is Equal. The FDA approved sucralose as a general purpose sweetener in 1999. Sucralose is manufactured through a process in which the molecular structure of sugar itself is modified, making it more heat-resistant than saccharin and aspartame.

In September 2000, McNeil introduced Splenda, the first artificial sweetener in the United States made from sucralose. Sales of Splenda grew more than tenfold in just six years, from approximately $32 million in 2001 to approximately $410 million in 2006. Within a year of its introduction, Splenda captured 14% of the total U.S. market for no-calorie sweeteners (based on dollar volume). In 2006, Splenda captured approximately 60% of the no-calorie sweetener market, compared to approximately 15% for Equal and 14% for Sweet `N Low.

All three leading artificial sweeteners are sold in distinctive packaging that helps consumers identify and distinguish them. Sweet `N Low and other saccharin-based sweeteners use predominantly red and pink packaging. This practice informs consumers that the particular product is made primarily with saccharin and, in the case of store-brand products, that the item competes with Sweet `N Low. Equal and other aspartame-based sweeteners use primarily blue packaging. Nearly all grocery store chains sell private-label saccharin and aspartame sweeteners that compare to Sweet `N Low and Equal, respectively. Finally, Splenda uses primarily yellow packaging.

As of the time it commenced this litigation, McNeil had spent nearly $250 million to promote and publicize the Splenda brand to consumers. Through its campaign, McNeil has highlighted the yellow Splenda packaging which includes the Splenda trademark in gradated blue italicized lettering on a white cloud. A Splenda package has been featured in nearly every Splenda television commercial and print advertisement since its launch.

C. Splenda Packaging

McNeil began selling boxes of individual Splenda packets in 2000. The boxes come in 100 and 200-count sizes and are identical except for their size. The boxes are oriented horizontally. The background is yellow, while the lettering on the boxes is primarily blue. The trade name "Splenda" appears at the top-center of the front of the boxes, in italicized blue lettering. The trade name is surrounded by a white, oval-shaped cloud. On the front, lower-right side of the boxes, there is a photograph of a white cup of coffee and saucer, with an individual Splenda packet resting on the saucer. On the front, left side of the boxes, there is a photograph of a glass and pitcher of iced tea. On the bottom-left corner is a circular element that contains the blue all-caps text, "Made from Sugar, Tastes Like Sugar."

McNeil also sells granular Splenda in vertically-oriented bags. The front of the Splenda bag is exactly the same as that of the Splenda boxes, except that it displays different physical props: a piece of pie on a plate, behind which are a bowl of cereal and a scoop containing granular Splenda.

D. Ahold, Food Lion, and Safeway Packaging

In mid-2006, private-label and store-brand sucralose products began to appear in the market. Heartland manufactures and packages a number of store-brand artificial sweetener products for retailers including Giant, Stop Shop, Tops, Food Lion, Safeway, Albertson's, and Wal-Mart. Giant, Stop Shop, and Tops are all owned by Ahold, and the packages of the store-brand sucralose products sold by each of these stores are identical except that each packaging contains that respective store's name and/or logo.

The Giant, Stop Shop, and Tops (collectively, Ahold) store-brand boxes of individual sucralose packets are oriented horizontally. The boxes come in 100 and 200-count sizes and are identical except for their size. The background is yellow, while the lettering on the boxes is either blue or white. The product name "Sweetener" appears at the top-center of the front of the boxes, in italicized blue font. The product name is outlined in white, but not by a cloud. On the lower-right corner is a photograph of a white cup of coffee and saucer, a glass of lemonade, and several fruits further off to the right side. There is a white rectangular border surrounding the front of the boxes. The store logo (regardless of the store name) appears just above the product name.

The Ahold stores also sell granular sucralose in vertically-oriented bags. The front of the Ahold bags is exactly the same as that of the Ahold boxes, except that it displays different physical props: a slice of cheesecake on a plate, a bowl of cereal with raspberries, and a white cup of coffee and saucer.

The Food Lion store-brand 100-count box of individual sucralose packets is oriented horizontally. The background is yellow, while the lettering on the box is either blue or black. The product name "Sweet Choice" appears on the bottom of the front of the box, in italicized blue font. The product name is not surrounded by a cloud. The front of the box contains a vertical design element that divides it into two portions. The left portion is in a darker yellow than the right, and includes the Food Lion logo (black) and store name (black) at the top. Food Lion uses this vertical design feature in some of its other store-brand packaging. The front-right portion contains a photograph of a white cup of coffee, saucer, and teaspoon, behind which are a pitcher of lemonade, two glasses of lemonade, and sliced lemons.

Food Lion also sells granular sucralose in vertically-oriented bags. The front of the Food Lion bag is exactly the same as that of the Food Lion box, except that it displays different physical props: a loaf of banana nut bread, a container of sucralose with a scoop, and a bowl of mixed berries.

The Safeway store-brand boxes of individual sucralose packets are oriented horizontally. The boxes come in 100 and 200-count sizes and are identical except for their size. The background is yellow, while the lettering on the boxes is mostly blue. The product name "Sucralose" appears on the bottom-left of the front of the boxes, in italicized blue font. Each letter in the product name is surrounded by a white shadow, but not all together by a cloud. The front of the boxes contains a white S-shaped design element that divides it into two portions. The S-shaped element leads to a display of the Safeway name (black) and logo (red) on the bottom-right corner. Safeway uses this S-shaped design feature in some of its other store-brand packaging. To the right of the S-shaped element is a circular element containing the count-size of the box, also found on the Safeway saccharin boxes. To the left of the S-shaped element is a photograph of a white cup of coffee, bowl of strawberries, and packet caddy containing individual packets of "Sucralose."

E. Additional Findings of Fact

A 100-count box of Splenda costs approximately $5.00, while the comparable store-brand sucralose products vary in price from approximately $4.00 to $4.60. The Heartland products are merchandised next to Splenda. The majority of sugar and artificial sweetener packages contain photographs of foods and drinks in which the sweetener is an ingredient. For example, packages typically depict hot and cold beverages, such as coffee, tea, iced tea, or lemonade; fruit; cereal; and baked goods, such as cake, bread, or pie.

In December 2006, Margaret Grossman, a California consumer, mistakenly purchased Safeway's "Sucralose" when she intended to purchase Splenda. At the time, she was "just buzzing through the market." She did not look at pricing, but just grabbed the box and ran. She is a self-described "surgical strike" shopper, intending to shop at a faster pace than others. She is aware that store-brand products exist; however, because she is not a comparison shopper, she is not aware that they are less expensive than national-brand products. Her yearly household income exceeds $300,000, far above the national median. She was not wearing her reading glasses while making the inadvertent purchase. She continued to use the product for three weeks before noticing that it was Safeway's "Sucralose."

F. Procedural History in the District Court

The procedural history of this case is straightforward. On December 5, 2006, McNeil filed a seven-count complaint against Heartland under federal and state law. Initially, McNeil sought relief for both the Splenda trade dress and the slogan "Made from Sugar, Tastes Like Sugar." On December 14, 2006, McNeil filed its motion for preliminary injunction. On February 7, 2007, the District Court held an evidentiary hearing on the motion, followed by oral argument on March 13, 2007. On May 21, 2007, the District Court denied the motion with respect to all of the allegedly infringing packages in a written memorandum and order. Eight days later, McNeil filed a timely notice of appeal from that decision. On appeal, McNeil does not raise any arguments with respect to the "Tastes Like Sugar" slogan or any arguments under state law, so its arguments and thus our analysis are limited to the single issue of trade dress infringement under federal law.

II.

The District Court had jurisdiction under 15 U.S.C. § 1121(a) and 28 U.S.C. § 1338. We have jurisdiction under 15 U.S.C. § 1121(a) and 28 U.S.C. § 1292(a)(1). To prevail on a motion for preliminary injunctive relief, the moving party must demonstrate that each of the following factors favors the requested relief:

"(1) the likelihood that the moving party will succeed on the merits;

(2) the extent to which the moving party will suffer irreparable harm without injunctive relief;

(3) the extent to which the nonmoving party will suffer irreparable harm if the injunction is issued; and

(4) the public interest."

Shire U.S. Inc. v. Barr Labs. Inc., 329 F.3d 348, 352 (3d Cir. 2003).

We review the denial of a preliminary injunction for abuse of discretion. However, "any determination that is a pre-requisite to the issuance of an injunction is reviewed according to the standard applicable to that particular determination." Id. (internal ellipses and citation omitted). Thus, we review a district court's factual findings for clear error and its legal conclusions de novo.

The District Court denied injunctive relief only on the basis that McNeil did not demonstrate a likelihood of success on the merits, and McNeil raises appellate arguments limited to that basis, so we shall similarly limit our analysis in this opinion.

III.

The Lanham Act, 15 U.S.C. § 1125(a), establishes a cause of action for trade dress infringement. TrafFix Devices, Inc. v. Mktg. Displays, Inc., 532 U.S. 23, 28-29, 121 S.Ct. 1255, 149 L.Ed.2d 164 (2001). "`Trade dress' refers to the design or packaging of a product which serves to identify the product's source." Shire US, 329 F.3d at 353. It is "the total image or overall appearance of a product, and includes, but is not limited to, such features as size, shape, color or color combinations, texture, graphics, or even a particular sales technique." Rose Art Indus., Inc. v. Swanson, 235 F.3d 165, 171 (3d Cir. 2000). The purpose of trade dress protection is to "secure the owner of the trade dress the goodwill of his business and to protect the ability of consumers to distinguish among competing producers." Shire US, 329 F.3d at 353 (internal brackets and citation omitted).

To establish trade dress infringement under the Lanham Act, a plaintiff must prove that (1) the allegedly infringing design is non-functional; (2) the design is inherently distinctive or has acquired secondary meaning; and (3) consumers are likely to confuse the source of the plaintiff's product with that of the defendant's product. Id. In the case at bar, the District Court analyzed the merits in favor of Heartland only on the basis that there was no likelihood of source confusion, and we shall so limit our analysis.

A likelihood of confusion exists when consumers viewing the defendant's trade dress probably would assume that the product it represents is associated with the source of a different product identified by the plaintiff's similar trade dress. A H Sportswear, Inc. v. Victoria's Secret Stores, Inc. (A H II), 237 F.3d 198, 211 (3d Cir. 2000). The likelihood of confusion between two trade dresses is a question of fact. A H Sportswear, Inc. v. Victoria's Secret Stores, Inc. (A H I), 166 F.3d 191, 194 (3d Cir. 1999); Am. Home Prods. Corp. v. Barr Labs., Inc., 834 F.2d 368, 370 n. 2 (3d Cir. 1987); see also A H II, 237 F.3d at 237 ("The question of likelihood of confusion is ultimately one of fact, and we cannot roll up our sleeves and engage in the balancing ourselves.").

One type of confusion is initial interest confusion, which is "confusion that creates initial customer interest, even though no actual sale is finally completed as a result of the confusion." 4 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 23:6 (4th ed. 2007). Courts have applied the initial interest approach "to the situation of look-alike private branded products." Id. For example, one court:

"found look-alike packaging on a nonprescription medicinal product to be infringing because it served to `hook customers at the initial point of contact with the product, thus initially drawing the customers to its product through the similarity in trade dress.' Even though the customer subsequently may realize the true source of the product before the sale is consummated, the confusing similarity of appearance has attracted the consumer and induced him or her to consider purchasing the product. `Defendant has already accomplished what it set out to do, which is to confuse the consumer at the point when he first reaches for the product on the shelf. It is at that point that the damage is done.'"

Id. (internal citation omitted).

We have held that "initial interest confusion is actionable under the Lanham Act." Checkpoint Sys., Inc. v. Check Point Software Techs., Inc., 269 F.3d 270, 292 (3d Cir. 2001). We reaffirm the holding that initial interest confusion is an independently sufficient theory that may be used to prove likelihood of confusion. See Australian Gold, Inc. v. Hatfield, 436 F.3d 1228, 1238-39 (10th Cir. 2006); Nissan Motor Co. v. Nissan Computer Corp., 378 F.3d 1002, 1018 (9th Cir. 2004); Promatek Indus., Ltd. v. Equitrac Corp., 300 F.3d 808, 812 (7th Cir. 2002); Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254, 260 (2d Cir. 1987). Another type of confusion, alleged in most Lanham Act cases, is point-of-sale confusion, which occurs or remains at the time of purchase. 4 McCarthy, supra, § 23:5.

Whether a plaintiff alleges initial interest or point-of-sale confusion or both, a district court should employ the factors we announced in Interpace Corp. v. Lapp, Inc., 721 F.2d 460 (3d Cir. 1983), to decide whether there is a likelihood of confusion. Checkpoint Sys., 269 F.3d at 297. These factors, as modified semantically in the trade dress context, are:

(1) the degree of similarity between the plaintiff's trade dress and the allegedly infringing trade dress;

(2) the strength of the plaintiff's trade dress;

(3) the price of the goods and other factors indicative of the care and attention expected of consumers when making a purchase;

(4) the length of time the defendant has used its trade dress without evidence of actual confusion arising;

(5) the intent of the defendant in adopting its trade dress;

(6) the evidence of actual confusion;

(7) whether the goods, though not competing, are marketed through the same channels of trade and advertised through the same media;

(8) the extent to which the targets of the parties' sales efforts are the same;

(9) the relationship of the goods in the minds of consumers because of the similarity of function;

(10) other facts suggesting that the consuming public might expect the plaintiff to manufacture a product in the defendant's market, or that the plaintiff is likely to expand into that market.

Freedom Card, Inc. v. JPMorgan Chase Co., 432 F.3d 463, 471 (3d Cir. 2005). "[T]he Lapp test is a qualitative inquiry. Not all factors will be relevant in all cases; further, the different factors may properly be accorded different weights depending on the particular factual setting." Id. (internal citation omitted).

IV.

On appeal in this case, McNeil argues that the District Court: (1) misapplied the first Lapp factor with respect to those products for which it weighed the factor in favor of Heartland; (2) erred in neglecting to consider the effect of color coding on the sweetener industry; (3) misapplied the third Lapp factor on consumers' degree of care when selecting sweeteners; (4) misapplied the sixth Lapp factor in disregarding evidence of actual confusion; and (5) clearly erred in concluding there was no likelihood of confusion with respect to those products for which it weighed the first Lapp factor in favor of McNeil. We address each of these arguments in turn.

A. First Lapp Factor: Degree of Similarity Between McNeil's Trade Dress and Heartland's Various Trade Dresses

We have held that "[t]he single most important factor in determining likelihood of confusion is [trade dress] similarity." A H II, 237 F.3d at 216. "The proper test is not side-by-side comparison but whether the [trade dresses] create the same overall impression when viewed separately." Kos Pharms., Inc. v. Andrx Corp., 369 F.3d 700, 713 (3d Cir. 2004) (internal quotation marks and citation omitted). However, if buyers typically see the two products side-by-side, as is true in this case, then a side-by-side comparison may be appropriate. See A H II, 237 F.3d at 216. Here, the District Court conducted a side-by-side comparison, and McNeil does not challenge its choice to do so. Following its analysis, the District Court concluded that "the similarity of trade dress factor weighs in favor of finding that there is no likelihood of consumer confusion for all of Heartland's products except for the Ahold 100 and 200 count boxes of individual packets, and the Ahold bag[s] of granular sucralose."

1. Heartland Products for Which the District Court Weighed the Degree of Similarity Against Finding a Likelihood of Confusion: Food Lion Box, Food Lion Bag, and Safeway Boxes

McNeil challenges the District Court's analysis of the first Lapp factor as to the Food Lion box and bag and the Safeway boxes. McNeil argues that the similarities the District Court found should have weighed more heavily than differences. The majority of the courts of appeals, including ours, has not adopted this rule. Instead, we have stated that forceful and distinctive design features should be weighed more heavily because they are more likely to impact the overall impression, A H II, 237 F.3d at 216, regardless of whether they happen to be similarities or differences.

Further, although it is legal error to engage in a "detailed analysis of the differences in the marks rather than focusing on the overall impression created by them," Fisons Horticulture, Inc. v. Vigoro Indus., Inc., 30 F.3d 466, 478 (3d Cir. 1994), it likewise would be legal error to engage in a skewed analysis of the similarities in the trade dresses rather than focusing on the overall impressions they create. See Bristol-Myers Squibb Co. v. McNeil-P.P. C, Inc., 973 F.2d 1033, 1046 (2d Cir. 1992) ("in a trade dress infringement case the question is not how many points of similarity exist between the two packages but rather whether the two trade dresses create the same general overall impression" (internal quotation marks and citation omitted)). McNeil's proposed minority rule would improperly take the focus away from the overall impression.

We review for clear error the District Court's determination of whether the overall impressions created by two trade dresses are similar. See Kos Pharms., 369 F.3d at 715; A H II, 237 F.3d at 218. "Clear error exists when, giving all deference to the opportunity of the trial judge to evaluate the credibility of witnesses and to weigh the evidence, we are left with a definite and firm conviction that a mistake has been committed." A H I, 166 F.3d at 194 (internal quotation marks and citation omitted).

In the case at bar, we cannot say with a definite and firm conviction that the District Court committed a mistake. With respect to all three packages at issue, the most important difference is that the trade name "Splenda" is not present, but the name and logo of the respective stores are. The leading commentator has observed:

"A vexing problem in trade dress cases is the extent to which prominent placement of a defendant's own word or design marks on look-alike trade dress will serve to prevent a likelihood of confusion. The majority of cases hold that labeling an otherwise infringing look-alike product does not prevent infringement. However, other cases in various factual settings, have found that the use of defendant's name on a look-alike product or package is sufficient in the overall context of the case to prevent a likelihood of confusion."

1 McCarthy, supra, § 8:16. Indeed, the absence of the "Splenda" label from the Food Lion and Safeway packages is not sufficient to cure an otherwise infringing trade dress, nor is the mere presence of another label. See Bristol-Myers Squibb, 973 F.2d at 1047.

However, the prominent presence of another well-known word or design mark might be sufficient. Fun-Damental Too, Ltd. v. Gemmy Indus. Corp., 111 F.3d 993, 1003 (2d Cir. 1997); August Storck K.G. v. Nabisco, Inc., 59 F.3d 616, 618 (7th Cir. 1995); Bristol-Myers Squibb, 973 F.2d at 1046 ("The presence and prominence of markings tending to dispel confusion as to the origin, sponsorship or approval of the goods in question is highly relevant to an inquiry concerning the similarity of the two dresses. When prominently displayed it can go far towards eliminating any possible confusion."); Fisher Stoves, Inc. v. All Nighter Stove Works, Inc., 626 F.2d 193, 194-95 (1st Cir. 1980) ("Each manufacturer displayed its name and logo prominently . . ., clearly identifying [the product's] origin[, so] . . . there is hardly likelihood of confusion or palming off when the name of the manufacturer is clearly displayed." (internal citation omitted)). We made a similar statement in Versa Products Co. v. Bifold Co., 50 F.3d 189, 203 (3d Cir. 1995), though limited in the product configuration (as opposed to product packaging) context. Although we reaffirm Versa Products in the sense that the Bristol-Myers Squibb principle is strongest in product configuration cases, the principle does not lose all applicability in packaging cases, especially because Bristol-Myers Squibb and August Storck were both product packaging cases.

On facts even more akin to those before us, the Court of Appeals for the Federal Circuit reached the same holding in Conopco, Inc. v. May Department Stores Co., 46 F.3d 1556, 1568 (Fed. Cir. 1994) (collecting supporting cases). There, the court applied Eighth Circuit law in a store-brand product packaging trade dress case and held that there was no likelihood of confusion because "[a] factor more probative of that issue . . . is the significance of the black and white diagonally-striped Venture [the name of the store] logo prominently situated on the front of the original and relaunched Venture products." Id. at 1566.

To be sure, we do not suggest that the prominent presence of another well-known mark is an affirmative defense to every trade dress infringement action. Sunbeam Prods. Inc. v. W. Bend Co., 123 F.3d 246, 259 n. 25 (5th Cir. 1997), abrogated on other grounds by TrafFix Devices, 532 U.S. at 28, 121 S.Ct. 1255. But this fact unquestionably plays a role in a district court's analysis of the first Lapp factor, such that it may cause the overall impressions created by two trade dresses to be different enough for the first Lapp factor to be weighed in a defendant's favor. When it is relevant, district courts should consider this fact as part of its analysis of the degree of similarity (first Lapp) factor, as was done in Fun-Damental Too, 111 F.3d at 1003, Versa Products, 50 F.3d at 202-03, and Bristol-Myers Squibb, 973 F.2d at 1046, as opposed to some sort of independent defense.

Applying these principles to the case at bar, we conclude that there was no clear error. First, "Food Lion" and "Safeway" are well-known because they are well-known to the consumers who shop in the stores with those same names. Second, the stores are represented prominently on their respective packages. For example, the Food Lion name and logo in black (a color with virtually no presence on the front of Splenda packages) are displayed in the top-left corner. As importantly, a vertical design element runs through the front of the package, visually dividing it between a dark yellow bar and a light yellow canvas, in a way found on other Food Lion store-brand products. The yellow color aside, these features are far more similar to other Food Lion store-brand packaging features and therefore distinguish themselves from any feature present on the Splenda packages. These distinguishing elements are also found on the Food Lion bag of granular sucralose.

The District Court found additional differences between the Food Lion trade dress and Splenda trade dress. First, the Food Lion product name "Sweet Choice" is shown rather than "Splenda." Second, "Sweet Choice" is positioned at the bottom of the box, rather than at the top. Third, whereas "Sweet Choice" is not surrounded by a white cloud, "Splenda" is. Fourth, missing from the Food Lion packages is the circular element with the slogan "Made From Sugar, Tastes Like Sugar." When combined with the distinguishing use of the Food Lion name and logo, these differences are not minute ones found only upon examination with a microscope. See n. 2, supra.

As for the Safeway boxes, the question is closer, but again not enough for us to conclude that the District Court committed clear error. Here, although the "Safeway" name and logo are not prominently displayed on its "Sucralose" product, the packages contain an S-shaped element that divides the entire front of the box vertically, so it is prominently displayed and visually striking. True, this S-shaped element is not as well-known to a Safeway shopper as the store's name, but its presence on other Safeway store-brand products renders it well-known as a store-specific signature that alerts the shopper that the product is more associated with Safeway than Splenda or anything else for that matter. A closer look at the S-shape would also cause the shopper to see the Safeway name and logo to which the S-shape leads, so on balance, the District Court could find that the S-shape design unique to Safeway materially differentiates the overall impressions created by the Splenda and Safeway packages, per Bristol-Myers Squibb.

As with Food Lion, the District Court found additional differences between the Safeway trade dress and Splenda trade dress. First, the product name "Sucralose" is shown rather than "Splenda." Second, "Sucralose" is positioned at the bottom of the box, rather than at the top. Third, whereas "Sucralose" is not surrounded by a white cloud, "Splenda" is. Fourth, missing from the Safeway packages is the circular element with the slogan "Made From Sugar, Tastes Like Sugar." Instead, Safeway's packaging contains a circular element with the box's count size. Again, the District Court could find that these cumulative differences are fairly forceful and distinctive. See A H II, 237 F.3d at 216.

When we apply the clear error standard of review to these facts, we cannot say with a firm and definite conviction that the District Court made a mistake in finding that the overall impression created by the Food Lion and Safeway packages is not similar to that created by Splenda's. This is especially true because in Bristol-Myers Squibb, August Storck, and Conopco, the Courts of Appeals for the Second, Seventh, and Federal Circuits, respectively, found clear error when the district courts weighed the degree of similarity factor in favor of plaintiffs where the defendants' own well-known marks were prominently displayed. See 973 F.2d at 1046, 59 F.3d at 620, 46 F.3d at 1566. Here, where the District Court weighed that factor in favor of the defendant by way of reasoning similar to that of our sister courts in those cases, we would be contradicting them were we to intrude upon the District Court's factfinding role in the opposite direction and say that it may not weigh the first Lapp factor in favor of the defendant under these circumstances. Therefore, the District Court did not clearly err in weighing the first Lapp factor in favor of Heartland as to both the Food Lion and Safeway products.

2. Heartland Products for Which the District Court Weighed the Degree of Similarity in Favor of Finding a Likelihood of Confusion: Ahold Boxes and Ahold Bags

With respect to the Heartland products (the Ahold boxes and bags) for which the District Court concluded that the first Lapp factor favors McNeil, McNeil argues that the District Court erred in not weighing this factor more than the others in the ultimate balancing. See Kos Pharms., 369 F.3d at 709 ("each factor must be weighed and balanced one against the other"). We will address this argument after examining whether the District Court misapplied any of the other Lapp factors. See section IV.E., infra. For now it suffices to say that, because Heartland does not challenge the District Court's finding with respect to the first Lapp factor, we will assume that it weighs in favor of McNeil as to those packages. B. Color Coding

McNeil argues that the history of color coding in the sweetener industry increases the likelihood of consumer confusion among sucralose-based products. A restaurant consumer, for example, encounters a range of sweeteners organized by packet color: white (and possibly brown) for sugar, pink for saccharin, blue for aspartame, and yellow for Splenda. According to McNeil, these colors have become a "shorthand" by which consumers identify their sweetener of choice. Moreover, there is apparently a history of the manufacturers of Equal and Sweet `N Low waiting too long to challenge imitators of their respective colors, whereas McNeil has been challenging its imitators since Splenda's inception. Therefore, McNeil argues that yellow does not signify sucralose; it signifies Splenda.

There are several problems with this argument. First, McNeil focuses it on the first Lapp factor. Yet, the District Court's analysis of the first Lapp factor amply considers yellow color as a point of similarity. Insofar as the District Court weighed the factor against McNeil, it is because the overall impressions of the packages are different, notwithstanding their uniform use of yellow. As we already have analyzed, the District Court did not clearly err on that score.

Second, just because a consumer sees yellow packaging in the sugar aisle does not mean that she believes McNeil or Splenda to be the source, especially because consumers are generally aware of the use of pink and blue by manufacturers other than those of Sweet `N Low and Equal, respectively. The sugar aisle in a representative grocery store also contains yellow packages of products other than sucralose, including sugar itself. In this factual context, we cannot conclude that whenever any other sucralose producer uses yellow packaging, consumers are likely to associate that product with Splenda.

C. Third Lapp Factor: Degree of Consumer Care

The third Lapp factor focuses on "the price of the goods and other factors indicative of the care and attention expected of consumers when making a purchase." 721 F.2d at 463.

"The following non-exhaustive considerations should guide a court's determination of the standard of ordinary care for a particular product. Inexpensive goods require consumers to exercise less care in their selection than expensive ones. The more important the use of a product, the more care that must be exercised in its selection. In addition, the degree of caution used depends on the relevant buying class. That is, some buyer classes, for example, professional buyers will be held to a higher standard of care than others. Where the buyer class consists of both professional buyers and consumers, the standard of care to be exercised by the reasonably prudent purchaser will be equal to that of the least sophisticated consumer in the class."

Versa Prods., 50 F.3d at 204-05 (internal quotation marks, ellipses, and citation omitted). Here, McNeil argues that the District Court erred in attributing a heightened level of care to consumers of no-calorie sweeteners. McNeil quotes from Versa Products and argues that: (1) because the products at issue cost on average between $4.00 and $5.00, they are inexpensive; and (2) the least sophisticated consumer in the buying class is not buying sucralose for health reasons and thus is not likely to exercise heightened care and attention.

Initially, we note that McNeil inaccurately characterizes the District Court's analysis as applying a "presumption" that sucralose consumers exercise special care. Here, the District Court did not presume but instead concluded — based on two affidavits — that the level of care and attention a consumer uses when purchasing no-calorie sweeteners are heightened because she often purchases them for health reasons: "blood-sugar disorders, including diabetes; obesity; weight loss; fitness; and tooth decay." The affidavits also refute McNeil's statement that there is no record evidence indicating that consumers who purchase a no-calorie sweetener apply a heightened level of care.

On the merits of McNeil's argument, the District Court did not err in defining the consumer class as consisting mostly of people who buy sucralose to improve their health. Versa Products' instruction to look to the "least sophisticated consumer in the class" is limited on its own terms to "[w]here the buyer class consists of both professional buyers and consumers." Id. at 205; accord Checkpoint Sys., 269 F.3d at 285 (applying doctrine where "both average consumers and specialized commercial purchasers buy goods"). Professional buyers are generally wholesalers, retailers, and long-established buyers. 4 McCarthy, supra, § 23:100. McNeil cites Ford Motor Co. v. Summit Motor Products, Inc., 930 F.2d 277, 293 (3d Cir. 1991), for the proposition that we have never held that professional buyers are required to create a mixed class. This is simply not true. In that case we specifically defined the buyer class as "consist[ing] of both professional buyers [in that case auto dealers] and consumers." Id.

Even more tellingly, the case on which we relied in Ford Motor to craft the "least sophisticated consumer" doctrine in the first place, Worthington Foods, Inc. v. Kellogg Co., 732 F.Supp. 1417 (S.D.Ohio 1990), specifically found that a "heightened awareness of health and healthy foods raises the standard of care which the reasonable purchaser of the parties' products would exercise." Id. at 1448. Worthington demonstrates that it is not proper, let alone necessary, to equate "health-conscious" with "professional." The ordinary consumer class consists of both health-conscious people (more of them in this case) and less health-conscious people (fewer of them in this case); this collection of consumers does not constitute the sort of "mixed class" we discussed in Ford Motor. Here, it is undisputed that the buyer class does not include professional buyers, so the District Court did not need to discern a "least sophisticated buyer" within a class consisting solely of ordinary consumers.

McNeil has one more arrow in its quiver, however. It argues that the health-conscious consumers distinguish not between brands of sucralose (both "healthy"), but only between sugar ("unhealthy") and sucralose ("healthy"). When faced with just sucralose products, the argument goes, these consumers need not exercise any heightened care when choosing which one to buy. We also reject this argument. First, it was never raised before the District Court. As such, Freedom Card counsels that we decline to address it. See 432 F.3d at 477-78. Second, even were we to address it, because the representative retail grocery store places sugar right next to sucralose, and sugar packages are often yellow, even when purportedly choosing one sucralose product over another, the reasonably prudent consumer might be careful not to pick up sugar by accident.

In sum, not presiding over the preliminary injunction hearings ourselves, we cannot say that the District Court clearly erred in finding that the reasonably prudent consumer in this case exercises some heightened care and attention when buying sucralose because her health considerations typically override the products' low cost. Therefore, the District Court did not clearly err in stating that the third Lapp factor did not weigh in McNeil's favor.

D. Sixth Lapp Factor: Evidence of Actual Confusion

On this factor McNeil argues that the District Court erred in concluding that McNeil failed to produce any evidence of actual consumer confusion. McNeil states that Ms. Grossman's testimony requires the District Court to weigh this factor in favor of McNeil. There are two problems with this argument. First, the District Court did not "completely disregard" Grossman's testimony. In fact, the District Court began by citing Versa Products for the proposition that "proof of actual confusion is not required for a successful trade dress infringement action under the Lanham Act." 50 F.3d at 205. The District Court then proceeded to consider thoroughly Grossman's testimony, but concluded — based on her own admissions (that she is a speed "surgical strike" shopper, that she is not a comparison shopper, that her yearly household income exceeds $300,000, that she did not look at the prices on the day she made her inadvertent purchase, and that she was not wearing her reading glasses that day) — that Grossman was not representative of the kind of shopper ordinarily purchasing sucralose.

Whatever the merits of this analysis, McNeil overstates its argument by averring that the District Court disregarded Grossman's testimony in its entirety. The relevant passage from the District Court's memorandum is as follows:

"Heartland's allegedly infringing products were introduced in mid-2006. This relatively short period of time and the fact that the products at issue are inexpensive, may explain why McNeil has not been able to produce credible evidence of actual consumer confusion. Therefore, even though McNeil has not produced any evidence of actual consumer confusion, we find it inappropriate to draw an inference that is unlikely to be able to do so. Consequently, we find that factor four does not favor Heartland or McNeil."

The District Court's statement that "McNeil has not produced any evidence of actual consumer confusion" is therefore an inadvertent mistake because the District Court also states: "McNeil has not been able to produce credible evidence of actual consumer confusion." As the quoted passage reveals, the District Court then weighed the fourth Lapp factor in favor of neither litigant. We find unavailing, then, McNeil's attempt to turn the District Court's technical misstatement on a separate Lapp factor (length of time defendant uses trade dress without evidence of actual confusion) into a ruling against it with respect to the factor McNeil actually argues on appeal (evidence of actual confusion).

McNeil fares no better on the merits. It cites precedent for the proposition that "evidence of actual confusion may be highly probative of the likelihood of confusion." Checkpoint Sys., 269 F.3d at 291. This is an unremarkable proposition. The District Court nonetheless has the ability — indeed, the duty — to weigh that evidence in analyzing the sixth Lapp factor. See id. at 297; A H II, 237 F.3d at 227. Many courts of appeals have endorsed the view that "[e]vidence of only a small number of instances of actual confusion can be dismissed as inconsequential or de minimis." 4 McCarthy, supra, § 23:14 (collecting cases). For example:

"Just as one tree does not constitute a forest, an isolated instance of confusion does not prove probable confusion. To the contrary, the law has long demanded a showing that the allegedly infringing conduct carries with it a likelihood of confounding an appreciable number of reasonably prudent purchasers exercising ordinary care."

Id. (quoting Int'l Ass'n of Machinists Aerospace Workers v. Winship Green Nursing Ctr., 103 F.3d 196, 200-01 (1st Cir. 1996)). It is still true that "it takes very little evidence to establish the existence of the actual confusion factor," id. (quoting AmBrit, Inc. v. Kraft, Inc., 812 F.2d 1531, 1544 (11th Cir. 1986)), but a district court may weigh the sixth Lapp factor in favor of a defendant when it concludes that "the evidence of actual confusion was isolated and idiosyncratic." A H II, 237 F.3d at 227; see also Checkpoint Sys., 269 F.3d at 298; Ziebart Int'l Corp. v. After Mkt. Assocs., Inc., 802 F.2d 220, 228 (7th Cir. 1986) ("a `single misdirected communication' would not show actual confusion of the marks by the consuming public"). Further, "[a]s to those few customers who do not care about brands and do not pay any attention to them, such `brand indifferent' customers do not count in the equation of likelihood of confusion." 4 McCarthy, supra, § 23:5 (collecting cases).

Applying these principles to the case at hand, we conclude that the District Court did not improperly weigh Grossman's testimony. Grossman's own admissions indicate that she is brand indifferent and otherwise unrepresentative of the typical shopper, and McNeil presented no other customer testimony. As in A H II, then, "the District Court's conclusion as to the absence of [probative evidence of] actual confusion was supported by the record." 237 F.3d at 227. McNeil would have us hold that even one piece of evidence of actual confusion compels a weighing of the sixth Lapp factor in favor of a plaintiff; neither precedent nor common sense supports such a result.

E. Balancing the Lapp Factors to Reach an Ultimate Factual Finding on Likelihood of Confusion; the District Court Clearly Erred in Not Finding a Likelihood of Confusion as to the Packages for Which It Weighed the First Lapp Factor in McNeil's Favor

Having concluded that the District Court did not misapply any of the Lapp factors for which McNeil alleges error, we have one final though critical question to answer: whether, for the products for which the District Court weighed the first Lapp factor in favor of McNeil, an intermediary finding that is easily justifiable and also not challenged on appeal, see section IV.A.2. n. 4, supra, it then committed clear error in not ultimately finding likelihood of confusion. The products, again, are the Ahold 100 and 200-count boxes and the Ahold bags of granular sucralose.

Here, the District Court committed clear error. On the record before us, there is no way the District Court could have ultimately balanced the Lapp factors against McNeil after weighing the first, second, seventh, eighth, and ninth Lapp factors in its favor (and the third, fourth, and tenth factors in favor of neither party) with respect to the Ahold boxes and bags. See Kos Pharms., 369 F.3d at 725 ("a remand for reweighing would waste judicial resources and unnecessarily delay the proceedings further"). The District Court clearly erred in the ultimate balancing of the Lapp factors because it did not adequately heed our oft-repeated statement that "[t]he single most important factor in determining likelihood of confusion is [degree of] similarity." A H II, 237 F.3d at 216; Fisons Horticulture, 30 F.3d at 476. Further, "when goods are directly competing [as is undisputed in the instant case], both precedent and common sense counsel that the similarity of the marks takes on great prominence." A H II, 237 F.3d at 214; accord Kos Pharms., 369 F.3d at 713.

It is true that "[n]ot all factors will be relevant in all cases; further, the different factors may properly be accorded different weights depending on the particular factual setting." Freedom Card, 432 F.3d at 471. However, the first Lapp factor is always relevant, and when it favors the plaintiff in a directly competing goods scenario, especially in which the District Court found only the fifth and sixth Lapp factors to favor the defendant outright, the defendant attempting to rebut the likelihood of confusion has a high hurdle to overcome. Here, the District Court apparently considered the following facts as rebuttal:

"Consumers are highly aware of the existence of store-brand products; when they are shopping in a particular store they are aware of the store's name; each of the Heartland products on sale in grocery stores displays the store name/logo; the Heartland and Splenda products typically appear next to each other; and there are other signals to the consumer on grocery store shelves; such as price differentials and shelf-talkers inviting consumer[s] to compare and save, that indicate to the consumer that the Heartland and Splenda products are not the same."

But with respect to the products for which the District Court weighed the first and several other Lapp factors in McNeil's favor, these considerations — even when combined with a correct finding in Heartland's favor on the sixth Lapp factor and what we assume is a correct finding in Heartland's favor on the fifth Lapp factor (which McNeil does not challenge on appeal) — are insufficient on the record before us to overcome the likelihood of confusion. See Kos Pharms., 369 F.3d at 725 (finding clear error in any ultimate balancing that results in no likelihood of confusion when the pro-plaintiff first Lapp factor in a competing goods situation is pitted against the pro-defendant third Lapp factor and neutral fourth and sixth Lapp factors).

The danger in the District Court's result is that producers of store-brand products will be held to a lower standard of infringing behavior, that is, they effectively would acquire per se immunity as long as the store brand's name or logo appears somewhere on the allegedly infringing package, even when the name or logo is tiny. The Lanham Act does not support such a per se rule.

We are not saying that the explanation provided by the District Court carries no weight. The fact that the products at issue were store-brand products does play a role in the inquiry. For example, in Conopco, the Court of Appeals for the Federal Circuit reasoned:

"The unique and extensive appearance of that logo in the store parking lot, on store signs, on employees' badges, in Venture's frequent and periodic print and television advertisements, and on other private label items sold by Venture, the large volume of Venture's annual sales ($1.3 billion in 1990), and the dearth of evidence that consumers ever purchased the Venture brand thinking it to have originated from Conopco [the national-brand product manufacturer] despite the extended (over 10 year) period over which the original and relaunched products were sold alongside the Venture brand, give rise to the expectation that consumers identify the logo with Venture, rather than Conopco, and use that logo to successfully distinguish between the two brands. The fact that Venture (and other retailers) compete in the manner described, which is all that the evidence establishes, is simply insufficient to amount to proof that there is in the minds of consumers a likelihood of confusion about whose product is whose."

46 F.3d at 1568. This reasoning is similar to the District Court's here, and it further supports our affirmance of the District Court on the products for which it weighed the first Lapp factor against McNeil.

However, it fails to save the District Court's conclusion on the Ahold boxes and bags. In contrast to the products in Conopco, see 46 F.3d at 1567 (product photos included in the opinion), the store name and logo are not prominently displayed on the Ahold packaging. Indeed, as already explained, the Food Lion and Safeway packages are in this sense much closer to the Conopco packages because a store-specific signature is prominently displayed on them, thereby substantially reducing the degree of similarity and hence the likelihood of confusion. The same simply cannot be said for the Ahold packages, and the District Court itself so found. To repeat, the District Court may take Conopco into account when analyzing the Lapp factors, in particular the first Lapp factor because under the Bristol-Myers Squibb framework, the more a store's name and/or logo are present around that store's shoppers, the more likely those shoppers will know well that name and/or logo, which in turn may serve to differentiate materially a store-brand packaging that displays them prominently. The District Court may not, however, consider the Conopco reasoning as an independent defense that altogether overrides the Lapp factors.

The nuanced distinction we make also brings the trade dress infringement issue back to its proper focus: just how similar the trade dresses are. See, e.g., Andrew Corydon Finch, Comment, When Imitation Is the Sincerest Form of Flattery: Private Label Products and the Role of Intention in Determining Trade Dress Infringement, 63 U. CHI. L.REV. 1243, 1276 (1996) ("courts should focus their efforts on the real issue: whether two trade dresses are so similar as to create a genuine likelihood of confusion among consumers"). Arguably under our holding, store brands can "get away" with a little more similarity than other defendants' products when they display prominently a well-known label, e.g., a store-specific signature, on their packages, but they cannot copy the national brands to such a degree of similarity, then merely affix a tiny differentiating label, as to become entirely immune to infringement actions. Compare TrafFix Devices, 532 U.S. at 29, 121 S.Ct. 1255 ("Trade dress protection must subsist with the recognition that in many instances there is no prohibition against copying goods and products."), with Two Pesos, Inc. v. Toco Cabana, Inc., 505 U.S. 763, 774, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992) ("Protection of trade dress, no less than of trademarks, serves the [Lanham] Act's purpose to secure to the owner of the mark the goodwill of his business and to protect the ability of consumers to distinguish among competing producers." (internal quotation mark omitted)).

Finally, our reversal of the District Court's clear error as to the Ahold boxes and bags based on its insufficient weighing of the first Lapp factor permits McNeil in its motion for preliminary injunction to proceed under an initial interest confusion theory (as well as a point-of-sale confusion theory) as to those products.

V.

For the foregoing reasons, we will affirm the District Court's denial of preliminary injunctive relief as to the Food Lion and Safeway products, but reverse the denial as to the Ahold 100 and 200-count boxes and the Ahold bags of granular sucralose. McNeil has demonstrated a likelihood of success on the merits for purposes of preliminary injunctive relief with respect to the third element of trade dress infringement, viz., under the Lanham Act, there is a likelihood of confusion between these products' trade dresses and the analogous Splenda trade dress. As to these Ahold products only, we will therefore remand for the District Court to consider whether McNeil establishes a likelihood of success on the remaining elements of trade dress infringement under the Lanham Act, as well as the remaining factors for preliminary injunctive relief.