HARTZ, Circuit Judge.

The parties to this trademark dispute make consumer products for rinsing sinus cavities. Med-Systems, Inc., the earlier entrant in this market, sells its products under the federally registered trademark SinuCleanse and two similar marks. Water Pik, Inc., which traditionally sold oral irrigators and showerheads, registered the trademark SINUSENSE with the intention of selling sinus-irrigation devices under the brand name "SinuSense." It brought an action against Med-Systems in the United States District Court for the District of Colorado, seeking a declaratory judgment that its use of the SinuSense brand name did not infringe on any of Med-Systems' marks. Med-Systems counterclaimed for trademark infringement and unfair competition under the Lanham Act, 15 U.S.C. §§ 1051-1127. Ruling that the SinuSense brand was not likely to cause consumer confusion, the district court awarded summary judgment to Water Pik on the counterclaims and dismissed Water Pik's declaratory-judgment claim as moot. We affirm.

I. BACKGROUND

Med-Systems was founded in 1997 by Dr. Diane Heatley and David Gallo. The company's business is selling nasal-irrigation products. One product line is a device known as a neti pot, which allows users to rinse their sinus passages with saline solution. Med-Systems also sells squeeze bottles and saline refill packets (which contain a powder to add to water to create the saline solution). These products are sold through a number of pharmacies and retail chains nationwide.

Med-Systems has obtained federal registration of three different trademarks to brand its products: the stylized word mark SinuCleanse, registered in 1998; the nonstylized word mark SINUCLEANSE, registered in 2008; and the nonstylized word mark SINUCLEANSE SQUEEZE, also registered in 2008. There is no dispute that these marks are statutorily incontestable under 15 U.S.C. § 1065 (2010). Photographs in the record indicate that the packaging for Med-Systems' products conspicuously features the stylized term "SinuCleanse," with the first component printed in plain type and the second component italicized.

Since the 1970s Water Pik has developed, manufactured, and sold oral irrigators and showerheads. In 2008 it sought to enter the market for sinus-irrigation products. It expressed interest in acquiring Med-Systems, but negotiations between the two parties failed in December 2008. Water Pik began developing its own nasal-irrigation line, which it ultimately decided to market under the name "Sinu-Sense." In May 2009 it filed an intent-to-use application for federal registration of the nonstylized word mark SINUSENSE, and in October 2009 the United States Patent and Trademark Office (PTO) published the mark for opposition. See 15 U.S.C. § 1051(b)(1) (2002) (allowing a person "who has a bona fide intention ... to use a trademark in commerce" to file for registration of the mark); id. § 1062(a) (2000) (providing that if it appears to the PTO that an applicant is entitled to registration, "the Director shall cause the mark to be published in the Official Gazette of the [PTO]"); id. § 1063(a) (2006) (allowing those who believe that they will be damaged by registration to file an opposition to registration of the mark). Med-Systems commenced an opposition proceeding against the SINUSENSE mark before the Trademark Trial and Appeal Board in December 2009, but that proceeding has been stayed pending resolution of this case.

Water Pik filed this action against Med-Systems in May 2010. It sought a declaratory judgment (1) that its use of the SINUSENSE mark did not infringe on any of Med-Systems' registered trademarks under § 32 of the Lanham Act, id. § 1114 (2005); (2) that the trade dress for its SinuSense-branded products did not infringe on the trade dress for any of Med-Systems' SinuCleanse-branded products under § 43(a) of the Lanham Act, id. § 1125(a) (2012); (3) that its use of the SINUSENSE mark did not constitute unfair competition against Med-Systems under § 43(a); (4) that its use of the mark did not constitute dilution of any of Med-Systems' trademarks under § 43(c) of the Lanham Act, id. § 1125(c); and (5) that it had a right to register the SINUSENSE mark and to use it in connection with its sinus-irrigation line.

Two months after Water Pik filed its complaint, SinuSense-branded products began appearing in stores. They included neti pots, squeeze bottles, electric nasal pulsators, and saline refill packets. Exhibits to Water Pik's complaint show the packaging for these products. They feature the word "waterpik" in large, unitalicized white letters against a blue background, and, directly below it in larger letters, the word "SinuSense" in unitalicized blue letters against a white background. The letter i in "waterpik" is modified to include three vertically stacked dots instead of one:

Med-Systems raised seven counterclaims: (1) that Water Pik's use of the SINUSENSE mark infringed on Med-Systems' registered trademarks in violation of § 32; (2) that the packaging for Water Pik's SinuSense-branded products constituted trade-dress infringement under § 43(a); (3) that Water Pik's use of the SINUSENSE mark constituted unfair competition under § 43(a); (4) that Water Pik's activities had diluted Med-Systems' trademarks under § 43(c); (5) that Med-Systems was entitled to an injunction against Water Pik's use of the SINUSENSE mark under 15 U.S.C. § 1116 (2008); (6) that Water Pik had tortiously interfered with Med-Systems' contractual relations with a national retailer by inducing the retailer to place SinuSense-branded products on its shelves in place of SinuCleanse-branded products; and (7) that through the same conduct, Water Pik had also tortiously interfered with Med-Systems' existing economic relations with that retailer. Med-Systems later dismissed the sixth and seventh counterclaims with prejudice, and the district court awarded summary judgment to Water Pik on the remaining counterclaims. The court dismissed Water Pik's claims for declaratory judgment as moot. See Water Pik, Inc. v. Med-Systems, Inc., 848 F.Supp.2d 1262, 1283 (D.Colo.2012).

Med-Systems' only challenges on appeal are to the summary judgment on its claims of trademark infringement and unfair competition. On those claims the district court ruled that Med-Systems had not produced adequate evidence that Water Pik's use of the SINUSENSE mark was likely to cause confusion among consumers. See id. at 1270-71. Med-Systems filed a motion to reconsider, arguing, among other things, that the court had overlooked a report on likelihood of confusion by one of its experts, Dr. Michael A. Belch. Dr. Belch's report was an analysis of a survey that measured consumer confusion arising from the SINUSENSE mark. After examining the survey and report, the court determined that they had too many methodological flaws to be of any probative value, so it did not revise its previous analysis.

On appeal Med-Systems argues that the district court erred in ruling that no genuine issue of fact existed regarding likelihood of confusion. It also suggests that an improper grant of summary judgment would violate its Seventh Amendment right to a jury trial; but we need not address that issue because the summary judgment was proper.

II. DISCUSSION

A. Likelihood of Confusion

Med-Systems claims that Water Pik's use of the SINUSENSE mark constitutes infringement of its registered trademarks as well as unfair competition, both of which are violations of the Lanham Act. Section 32 of the Act allows the owner of a registered mark to bring an infringement action against any person who

15 U.S.C. § 1114(1)(a) (emphasis added). And § 43(a) of the Act provides:

Id. § 1125(a) (emphasis added). Although Med-Systems does not clarify whether it is appealing summary judgment on its § 32 counterclaim, its § 43(a) counterclaim, or both, the central inquiry is the same: whether the junior user's mark is likely to cause confusion with the senior user's mark. See, e.g., Team Tires Plus, Ltd. v. Tires Plus, Inc., 394 F.3d 831, 832-33 (10th Cir.2005) (§ 32 claim); Cardtoons, L.C. v. Major League Baseball Players Ass'n., 95 F.3d 959, 966 (10th Cir.1996) (§ 43(a) claim); see also Jordache Enters., Inc. v. Hogg Wyld, Ltd., 828 F.2d 1482, 1484 (10th Cir.1987) (noting that similar tests govern claims under both provisions).

Likelihood of confusion is a question of fact, but the court may grant summary judgment in appropriate circumstances. See Sally Beauty Co., Inc. v. Beautyco, Inc., 304 F.3d 964, 972 (10th Cir.2002). In evaluating whether there is a likelihood of confusion, we examine six nonexhaustive factors: (1) evidence of actual confusion; (2) the strength of the contesting mark; (3) the degree of similarity between the competing marks; (4) the intent of the alleged infringer in adopting the contested mark; (5) the degree of care that consumers are likely to exercise in purchasing the parties' products; and (6) the similarity of the parties' products and the manner in which they market them. See Vail Assocs., Inc. v. Vend-Tel-Co., Ltd., 516 F.3d 853, 863 (10th Cir.2008). The importance of any particular factor in a specific case can depend on a variety of circumstances, including the force of another factor. For example, if the two marks are not at all similar, the degree of consumer care is unlikely to make a difference. As we have said, "These [six] factors are interrelated and no one factor is dispositive." Sally Beauty, 304 F.3d at 972. "At all times ... the key inquiry is whether the consumer is likely to be deceived or confused by the similarity of the marks." Team Tires, 394 F.3d at 833 (internal quotation marks omitted).

B. Standard of Review

We review a summary judgment de novo, applying the same standard that the district court should have applied. See id. at 832. "Summary judgment is appropriate if the pleadings, depositions, other discovery materials, and affidavits demonstrate the absence of a genuine issue of material fact and that the moving party is entitled to judgment as a matter of law." Sally Beauty, 304 F.3d at 971; see Fed. R.Civ.P. 56(a). A factual issue is genuine "if the evidence is such that a reasonable jury could return a verdict for the nonmoving party." Sally Beauty, 304 F.3d at 972 (internal quotation marks omitted). The nonmoving party is entitled to all reasonable inferences from the record; but if the nonmovant bears the burden of persuasion on a claim at trial, summary judgment may be warranted if the movant points out a lack of evidence to support an essential element of that claim and the nonmovant cannot identify specific facts that would create a genuine issue. See id. at 971-72. As the party alleging infringement under § 32 and unfair competition under § 43(a), Med-Systems bears the burden of proving a likelihood of confusion at trial. See John Allan Co. v. Craig Allen Co., 540 F.3d 1133, 1138 (10th Cir.2008) (§ 43(a) unfair-competition claim); Jordache, 828 F.2d at 1484 (§ 32 infringement claim).

C. Application to This Dispute

Med-Systems contends that summary judgment was inappropriate because likelihood of confusion was a genuine issue of fact. We apply our six-factor test to evaluate likelihood of confusion between the SinuCleanse and SinuSense marks, addressing Med-Systems' specific arguments challenging the district court's evaluation of the factors.1 As we shall see, only the sixth factor favors Med-Systems and the factors as a whole strongly favor Water Pik.

1. Actual Confusion

"[A]ctual confusion in the marketplace is often considered the best evidence of likelihood of confusion." King of the Mountain Sports, Inc. v. Chrysler Corp., 185 F.3d 1084, 1092 (10th Cir.1999) (internal quotation marks omitted). For one thing, such empirical data can be a reality check on the more theoretical analysis under the other factors. Med-Systems cites two different kinds of evidence of actual confusion: survey evidence gathered by Dr. Belch; and some emails, blog posts, and affidavits. We first address the survey evidence.

a. Survey Evidence

"Evidence of actual confusion is often introduced through the use of surveys, although their evidentiary value depends on the methodology and questions asked." Vail Assocs., 516 F.3d at 864 n. 8 (internal quotation marks omitted). To test consumer confusion, Dr. Belch designed a questionnaire that respondents completed over the Internet. After first answering a few questions for demographic identification and to establish their experience with purchasing healthcare products, respondents were shown the unitalicized typewritten words "SinuCleanse," "SinuSense," and "NeilMed" in varying orders and were asked (Question 3) what came to mind when they saw those marks. Aplt.App., Vol. 6 at 1434. (NeilMed is a third competitor in the sinus-irrigation market.) They then viewed packages for SinuCleanse-branded products, SinuSense-branded products, and NeilMed-branded products in varying orders and were asked to rank, on a scale of one to seven, how appealing they found the design of each package. These initial questions were to disguise the goal of the survey while focusing attention on the packages.

Next, respondents were shown packaging for saline refill packets that carried the SinuCleanse, SinuSense, and NeilMed brands. After viewing each brand's package in isolation, then all three packages side by side, respondents were given a sequence of three questions aimed at measuring confusion. The first asked whether they thought that two or more of the displayed products were made by the same company. The second asked whether they thought that one or more of the companies whose brands were displayed had a business relationship with one of the other companies. And the third asked whether they thought that one or more of the companies whose brands were displayed had obtained permission or approval from one of the other companies. Respondents were presented with the second question only if they responded no or not sure to the first, and they were presented with the third question only if they had responded no or not sure to the second. Respondents who answered yes to any one question were asked to identify which two or three brands they thought were related and why. The survey also used this same procedure, with the same questions, to gauge reactions to neti pots branded as SinuCleanse, SinuSense, and NeilMed. In both cases, after the three-question sequence all respondents answered a final question in which they were shown three pairs of typewritten words — "SinuCleanse and SinuSense," "NeilMed and SinuCleanse," and "SinuSense and NeilMed" — and were asked (Question 10) to rank the similarity between the paired words on a scale from 1 to 7. Id. at 1446.

The district court ruled that the results of Dr. Belch's survey were "devoid of any probative value and therefore irrelevant," id. Vol. 8 at 2244 (Order at 22, Water Pik, Inc. v. Med-Systems, Inc., Civ. No. 10-cr-01221-PAB-CBS, 2012 WL 2153162 (D.Colo. June 13, 2012)), because it perceived several serious methodological flaws in the survey: (1) Dr. Belch selected respondents from those whom he regarded as potential purchasers of sinus remedies in general, not potential purchasers of sinus-irrigation products in particular (which account for only 1% of sinus-remedy sales), thus creating an overinclusive survey universe; (2) Dr. Belch, based on the opinion of Med-Systems' CEO Gallo regarding the composition of the sinus-remedy market, skewed the survey population to ensure that 42% of respondents were older than 55, but an independent market study indicated that fewer than 29% of sinus-remedy consumers belonged to that age group; (3) Questions 3 and 10 did not present the marks as they would appear to a consumer because both marks were shown in a typewritten format, divorced from packaging, and without any italics; and the SinuSense mark was unaccompanied by the distinctive "waterpik" logo that appears on packaging for SinuSense products; (4) the survey presented images of Water Pik's products and images of Med-Systems' products side by side, thereby failing to recreate the true conditions of the marketplace; and (5) the questions were leading and suggestive because they encouraged respondents to entertain the possibility of an affiliation between SinuCleanse- and SinuSense-branded products when they might not have considered such a possibility on their own.

Med-Systems argues that the district court erred in finding flaws and that far from being methodologically unsound, Dr. Belch's survey provides strong evidence that consumer confusion is likely.

"Before expert testimony can be admitted, Federal Rule of Evidence 702 requires the district court to `ensure that any and all scientific testimony or evidence admitted is not only relevant, but reliable.'" In re Williams Sec. Litig. -WCG Subclass, 558 F.3d 1130, 1137 (10th Cir. 2009) (quoting Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 589, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993)). "We review for an abuse of discretion the manner in which the district court exercises its Daubert. `gatekeeping' role in deciding whether to admit or exclude survey evidence." Vail Assocs., 516 F.3d at 864 n. 8.

In our view the district court's concerns about the survey's methodology were reasonable. And we see no error in its deciding that the survey failed to support a likelihood of confusion.

First, Questions 3 and 10 presented both marks in a format that did not accurately reproduce the manner in which consumers ordinarily view them. We address each mark in turn. The SinuCleanse mark was typewritten without italics. Med-Systems argues that use of the nonstylized mark was proper because consumers frequently encounter it in that form. But the record indicates otherwise. Every photograph of a Med-Systems package conspicuously features the stylized term "SinuCleanse," with the first component printed in plain type and the second component italicized. To be sure, the narrative on some packages contains a nonstylized mark in small type, and the nonstylized mark appears on Med-Systems' website and elsewhere on the Internet; but the nonstylized mark is generally accompanied by a more prominent stylized mark. These uses of the nonstylized mark are of minimal consequence. They are likely to escape the attention of consumers, as they escaped the attention of Med-Systems' CEO Gallo, who testified as follows at his deposition after being shown the stylized mark (Exhibit 4) and the nonstylized mark (Exhibit 5).

Aplt.App., Vol. 2 at 157-58.

As for the SinuSense mark, it was typewritten without the "waterpik" logo — with the three dots stacked vertically above the letter i — which appears prominently and in close proximity to the SinuSense mark on packaging for SinuSense products. The same logo appears prominently on packaging for other Water Pik products as well. In other words, the logo functions as a "house mark" — that is, "a mark used on several different goods or services which themselves use a particular `product mark.'" 4 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 23:43, at 23-218 (4th ed. 2013) (4 McCarthy). Examples of house marks appended to product marks include "APPLE MACINTOSH computer, FORD MUSTANG auto and VASELINE INTENSIVE CARE hand lotion." Id.

Marks should be compared "as a whole as they are encountered by consumers in the marketplace." King of the Mountain, 185 F.3d at 1090 (internal quotation marks omitted); cf. 6 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 32:163, at 32-357 (6 McCarthy) ("[T]he closer the survey methods mirror the situation in which the ordinary person would encounter the trademark, the greater the evidentiary weight of the survey results."). By presenting the SinuCleanse mark differently from the way that it actually appears on packaging, and by failing to include the "waterpik" house mark next to the SinuSense product mark, Questions 3 and 10 exaggerated the similarities between the two marks, likely increasing the confusion of the respondents.

Second, when the survey did ask respondents to react to the marks as they actually appeared on product packaging, it presented the marks side by side along with the NeilMed mark. But a side-by-side comparison ordinarily is not the proper method of determining likelihood of confusion. The marks "must be compared in light of what occurs in the marketplace, not in the courtroom." Beer Nuts, Inc. v. Clover Club Foods Co. (Beer Nuts I), 711 F.2d 934, 941 (10th Cir.1983) (internal quotation marks omitted). Thus, in Jordache, 828 F.2d at 1488, we affirmed the decision by the district court to give minimal weight to a survey that used a side-by-side comparison because it bore "little resemblance to the actual workings of the marketplace." But see SquirtCo v. Seven-Up Co., 628 F.2d 1086, 1089 n. 4, 1091 (8th Cir.1980) (using side-by-side comparison). See generally 6 McCarthy § 32:173.50 (noting that one expert approves of side-by-side comparisons to compare two weak marks if products are simultaneously or sequentially accessible in the marketplace). A side-by-side comparison can emphasize differences in marks that might not be noticed by the actual shopper (thereby decreasing the respondents' likelihood of confusion), or it may bring to mind the senior mark when the respondent would not otherwise have thought of it (thereby increasing the respondents' likelihood of confusion). As the district court noted, Med-Systems cited no evidence that SinuCleanse and SinuSense were sold side-by-side in stores. Although Med-Systems did point to screenshots of search results from online shopping websites, where both Sinu Cleanse products and SinuSense products are featured, these screenshots do not support its contention that a side-by-side comparison was the proper survey design. In none of the screenshots are SinuCleanse, SinuSense, and NeilMed products isolated and displayed next to one another; rather, they are scattered among a host of other sinus remedies, which bear names like "NasalCare," "SinuPulse," and "SinuAir." Aplt.App., Vol. 6 at 1464, 1467, 1469-70. The district court could properly distrust the survey's side-by-side format.

Third, the district court could reasonably view the survey questions as improperly leading. Respondents were shown only three products and were asked whether two or more of the products were made by the same company; whether two or more of the products' makers had a business affiliation; and whether one or more of the makers had received permission or approval from one of the others. They could answer yes, no, or not sure. "A survey question that begs its answer by suggesting a link between plaintiff and defendant cannot be a true indicator of the likelihood of consumer confusion." Scott Fetzer Co. v. House of Vacuums, Inc., 381 F.3d 477, 488 (5th Cir.2004). In Scott Fetzer the leading question was similar to what was asked in Med-Systems' survey. Respondents were shown an ad for the defendant and were asked whether the defendant was "in any way affiliated with, connected with, associated with or authorized by" the plaintiff's company. Id. at 487-88 (internal quotation marks omitted); see also Universal City Studios, Inc. v. Nintendo Co., Ltd., 746 F.2d 112, 118 (2d Cir.1984) ("To the best of your knowledge, was the Donkey Kong game made with the approval or under the authority of the people who produce the King Kong movies?" (internal quotation marks omitted)); IDV N. Am., Inc. v. S & M Brands, Inc., 26 F.Supp.2d 815, 831 (E.D.Va.1998) ("Do you think there is any connection between (BAILEYS Irish Cream/BAILEYS liqueur/BAILEYS alcohol, if mentioned) and Bailey's cigarettes?" (internal quotation marks omitted)); see generally 6 McCarthy §§ 32:172, 32:175; cf. id. § 32:175 at 32-405 ("In the author's view, it is not improperly leading to present the respondent with a fair and accurate representation of the contesting marks and ask in a neutral manner if the respondent thinks there is or is not some connection, affiliation or sponsorship relation between the owners of the marks — or that the respondent doesn't know.... However, it will probably be improperly leading to suggest the desired response in the form of a yes or no question.") By suggesting the possibility that SinuSense might be connected with another brand, and limiting the candidates to SinuCleanse and NeilMed, the survey questions risked sowing confusion between SinuSense and SinuCleanse when none would have arisen otherwise. The district court could reasonably conclude that the question suggested the very answer most helpful to Med-Systems' cause rather than eliciting responses as they might occur spontaneously in the marketplace.

Moreover, as pointed out by Water Pik's expert in the district court, Med-Systems' survey, even if assumed to be methodologically sound, does not show a likelihood of confusion. To arrive at this conclusion, we must discuss how likelihood of confusion is computed from surveys. The key questions in Med-Systems' survey were those that asked whether the SinuCleanse and SinuSense products were made by the same company, whether the companies had a business relationship, or whether one of the companies had obtained permission or approval from the other to make its product. The raw confusion rate was the percentage of respondents who answered yes to one of these questions. For the trial in which respondents viewed packaging for saline packets, 203 respondents participated and 47 answered yes to one of the three questions, yielding a raw confusion rate of about 23%. For the trial involving packaging for neti pots, 198 participated and 29 answered yes, producing a raw confusion rate of 14.6%.

Such raw figures are usually unhelpful, however, in predicting likelihood of confusion, because they can be inflated by "background noise," or false positives arising from something other than the particular confusion that is alleged and that the survey aims to capture. 6 McCarthy § 32:187, at 32-433. "In any survey, there will always be some interviewees who are bored, hurried or just plain contrary and whose responses must be filtered out...." Id. (We would add another cause of false positives — poor survey design that stimulates confusion.) Surveys can adjust for this background noise by using controls. To isolate confusion arising specifically from the contested mark, survey designers will substitute for the contested mark a control mark that shares as many characteristics with the contested mark as possible, "with the key exception of the characteristic whose influence is being assessed." THOIP v. Walt Disney Co., 690 F.Supp.2d 218, 240 (S.D.N.Y.2010) (internal quotation marks omitted); see, e.g., First Nat'l Bank in Sioux Falls v. First Nat'l Bank S.D., 679 F.3d 763, 770-71 (8th Cir.2012) (when two marks at issue were "First National Bank South Dakota" and "First National Bank in Sioux Falls," the use of "First Bank & Trust" as a control was proper (internal quotation marks omitted)); AHP Subsidiary Holding Co. v. Stuart Hale Co., 1 F.3d 611, 613 (7th Cir.1993) (using "Mazola" and "Wesson-Lite" cooking sprays as controls when the contested mark was "Pan-Lite" and the contesting mark was "PAM"). By discounting for confusion arising from the control, the survey can measure net confusion, or "the difference between the raw confusion percent and the control confusion percent." 6 McCarthy § 32:187, at 32-435; see AHP, 1 F.3d at 613-14 (control confusion of 7% was subtracted from raw confusion of 45%, yielding net rate of 38%).

Here, the control used in the survey was NeilMed. The net confusion rate in each trial, then, was the rate of confusion between SinuCleanse and SinuSense, minus the rate of confusion between SinuCleanse and NeilMed. Stated differently, the net rate of confusion was the net number of confused respondents divided by the total number of respondents. This number was insubstantial for both trials. In the saline-packet trial, 34 of the 203 respondents, or 16.7%, expressed confusion between SinuCleanse and NeilMed. Subtracting these respondents from the 47 who were confused between SinuCleanse and SinuSense, one arrives at a net confusion rate of 13 divided by 203, or less than 6.5%.2

In the neti-pot trial, 29 of the 198 total respondents, or 14.6%, expressed confusion between SinuCleanse and SinuSense. But 30 respondents (15.2%) expressed confusion between SinuCleanse and NeilMed. Therefore, the net confusion rate for the neti-pot trial was negative 0.5% (1 divided by 198). Net confusion rates of 6.5% and negative 0.5% do not create a genuine factual issue of likelihood of confusion. See 1-800 Contacts, Inc. v. Lens.com, Inc., 22 F.3d 1229, 1246-49, Nos. 11-4114, 11-4204, 12-4022, 2013 WL 3665627, at **13-15 (10th Cir. July 16, 2013) (net confusion rates of 7.5% and 7.3% were too low); and because the methodological shortcomings identified by the district court probably increased the net confusion rate, the survey is even more harmful to Med-Systems.

Dr. Belch's report cited net confusion rates higher than the ones calculated above, but only by using a nonsensical analysis. Rather than repeating all his computations, we present only one, which is typical and illustrative. In the saline-packet trial, 22 of the 203 respondents answered that two or more brands had a business affiliation; 14 said that the two brands were SinuCleanse and SinuSense, whereas 6 said that they were SinuCleanse and NeilMed. The proper analysis would be that the raw confusion rate from this question was 14 over 203, or roughly 6.9%, and that the net confusion rate was the net number of confused respondents (14 minus 6) divided by the total number of respondents (203), or 3.9%.

Dr. Belch, however, divided the net number of respondents confused between SinuCleanse and SinuSense (14 minus 6) by the number of respondents confused between any pair of brands (or by all three) — that is, those who answered that any two or more of the three brands had a business relationship. The number of such confused respondents was 22, so he arrived at a net confusion rate of 8 over 22, or 36%. This approach greatly exaggerates the confusion rate. The 36% figure does not measure the net percentage of consumers who were confused between SinuCleanse and SinuSense; as stated above, that would be 3.9%. The absurdity of Dr. Belch's approach can be shown by an illustration. Say, 2 respondents out of 203 (or out of 1,000, for that matter) confused SinuCleanse and SinuSense while 1 respondent confused SinuCleanse and NeilMed. Dr. Belch's net confusion rate would then be 33%, the net number of confused respondents (2 minus 1) divided by the total number of confused respondents (3). Yet the true raw confusion rate would be less than 1% (2 divided by 203) and the true net confusion rate would be less than 0.5% (1 divided by 203). Dr. Belch's calculation is irrelevant.

We also hold that the likelihood of confusion is not shown by the respondents' answers to Question 10. Question 10 presented respondents with three different pairs of marks at once — SinuCleanse (not stylized) and SinuSense; SinuCleanse (not stylized) and NeilMed; and SinuSense and NeilMed — and asked them to rate the similarity of each pair on a scale from 1 ("Not at all similar") to 7 ("Very similar"). Aplt. App., Vol. 2 at 404. We have already noted the problem created by not comparing the marks as they actually appear' (with SinuCleanse stylized and SinuSense accompanied by the waterpik house mark). But there is an even more fundamental problem. Dr. Belch never explained, and we have no way of knowing, whether any particular score indicates a likelihood of confusion and why. We may say that siblings look very similar yet never confuse them. Cf. Beer Nuts I, 711 F.2d at 942 (commenting that the district court "erroneously equated likelihood of confusion with similarity"); J.B. Williams Co., Inc. v. Le Conté Cosmetics, Inc., 523 F.2d 187, 192 (9th Cir.1975) ("The marks may be similar in appearance ... yet not likely to cause confusion as to their source, particularly when all the factors are considered."). All that can be drawn from the data is the unsurprising conclusion that people saw SinuCleanse and SinuSense as more similar than SinuCleanse and NeilMed or Sinu-Sense and NeilMed. This conclusion is of little assistance in assessing whether confusion was likely.

In short, the Med-Systems survey strongly suggests an insignificant likelihood of confusion.

b. Nonsurvey Evidence

In addition to Dr. Belch's survey, Med-Systems cites four pieces of evidence that it says show actual confusion: (1) a text message to cofounder Gallo in February 2011 from a friend and customer who claimed to have seen an ad for SinuCleanse products on the Weather Channel, where Med-Systems apparently does not advertise (although there is no evidence that the ad seen by the friend was for SinuSense); (2) a March 2011 post by a coupon blog that featured a photograph of a SinuSense-branded squeeze bottle along with the caption "WaterPik SinuCleanse"3; (3) a June 2011 post on the online auction site eBay featuring a photograph of Sinu-Sense-branded saline packets along with the caption "Sinucleanse 60 Soothing Saline Packs Easy Pour Waterpik"; and (4) a declaration by a consumer that she was confused between the two brands when she purchased a neti pot at Walgreens.

We have consistently recognized, however, that isolated, anecdotal instances of actual confusion may be de minimis and may be disregarded in the confusion analysis. What is required for a claim of trademark infringement under the Lanham Act is a likelihood of confusion, not merely the possibility of confusion. See 4 McCarthy § 23:3. In Universal Money Centers, Inc. v. AT & T Co., 22 F.3d 1527, 1535 (10th Cir.1994), for example, the plaintiff submitted an affidavit by a customer that he had been confused by the defendant's mark, and two affidavits by the plaintiff's employees that they had received a number of accounts of similar confusion from consumers. We ruled that this evidence, together with survey data reflecting a 2.6% rate of actual confusion, was de minimis. See id. at 1534-35. And in King of the Mountain we ruled that evidence of seven episodes of actual confusion was de minimis. See 185 F.3d at 1092-93.

The district court characterized Med-Systems' nonsurvey evidence as de minimis, see Water Pik, 848 F.Supp.2d at 1274, and we agree. As the D.C. Circuit observed in a trademark-infringement case predating the Lanham Act, "Probable confusion cannot be shown by pointing out that at some place, at some time, someone made a false identification." McGraw-Hill Publ'g Co. v. Am. Aviation Assocs., 117 F.2d 293, 295 (D.C.Cir.1940) (Vinson, J.). The four instances cited by Med-Systems are isolated episodes with minimal probative value on whether reasonable consumers as a whole are actually confused by Water Pik's trademark.

Med-Systems has wholly failed to show sufficient actual confusion to support a claim of likelihood of confusion.4 This factor strongly favors Water Pik.

2. Strength of Med-Systems' Mark

Likelihood of confusion depends partly on the senior mark's strength — that is, its "capacity to indicate the source of the goods or services with which it is used." Restatement (Third) of Unfair Competition § 21 cmt. i (1995). Strength has two aspects: conceptual strength, or the mark's place on the spectrum of distinctiveness, and commercial strength, or its level of recognition in the marketplace. See King of the Mountain, 185 F.3d at 1093.

The strength of the contesting mark can be important in several contexts. For example, if the mark is particularly strong, use of that mark by another can give rise to likelihood of confusion even if the use occurs in connection with a totally distinct product line. The Fourth Circuit has offered a helpful illustration: "If the trade mark is a coined word, such as Kodak, it is more probable that all goods on which a similar designation is used will be regarded as emanating from the same source than when the trade-mark is one in common use on the variety of goods, such as `Gold Seal' or `Excelsior.'" Arrow Distilleries, Inc. v. Globe Brewing Co., 117 F.2d 347, 349 (4th Cir.1941) (internal quotation marks omitted); see id. at 350 ("English courts have prevented one from putting forth Kodak bicycles, at the suit of the originator of the name for a totally different article." (internal quotation marks omitted)). On the other hand, if a mark is weak, use of a similar mark even on similar goods is unlikely to cause confusion if minor differences distinguish one party's mark from another. See First Sav. Bank, 101 F.3d at 655 ("When the primary term [in a mark] is weakly protected to begin with, minor alterations may effectively negate any confusing similarity between the two marks."). The latter principle is what applies here. As we shall see, the SinuCleanse mark is a weak one. The differences between it and Water Pik's mark therefore make confusion highly unlikely.

We begin by analyzing the conceptual strength of the SinuCleanse mark.

a. Conceptual Strength

In measuring conceptual strength, the spectrum of distinctiveness ranges along the following five categories of marks (from least to most distinctive): (1) generic, meaning that the mark identifies only a general class of goods to which a specific product may belong; (2) descriptive, meaning that it reflects one or more of the product's characteristics or qualities; (3) suggestive, meaning that it suggests or evokes, rather than describes, the nature of the product, requiring "the consumer to use imagination and perception to determine the product's nature"; (4) arbitrary, meaning that the mark is a word or symbol already in common use that does not have any apparent relation to the product (as in Apple computers); and (5) fanciful, meaning that the mark is a novel word or design that has been coined for the sole purpose of serving as a trademark (as in Kodak or Exxon). Sally Beauty, 304 F.3d at 975-76. Generic marks are not protected as trademarks, and a descriptive mark can receive trademark protection "only when it has acquired a secondary meaning by becoming distinctive of the [owner's] goods in commerce." Id. at 976 (internal quotation marks omitted). Unlike descriptive marks, "[s]uggestive, fanciful, and arbitrary marks are considered inherently distinctive" and need not acquire secondary meaning to warrant protection as trademarks. Id.

The mark "SinuCleanse" is a weak one. For two reasons. First, the "Sinu" component of SinuCleanse is a widely used prefix that is essentially generic. See First Savings Bank, 101 F.3d at 655 (noting that the word BANK "is clearly generic"). Although Med-Systems disputes Water Pik's assertion that the prefix "Sinu" is commonly used by third parties, Water Pik produced scores of third-party trademark registrations in which sinu or sinus appeared, including several sinus-irrigation products. Med-Systems does not challenge their existence or authenticity. Further, Med-Systems CEO Gallo answered yes at his deposition to the question, "[A]s the marketplace exists today, there are a number of companies that use the prefix `Sinu' to describe some type of product dealing with your sinuses?" Aplt.App., Vol. 2 at 159-60. This extensive third-party use of a component of a disputed term undermines the strength of the term as a whole. See First Savings, 101 F.3d at 655 (widespread third-party use of "FIRST BANK" and variations made it a weak term, undermining the strength of the mark "First Bank System"). In particular, "third-party registrations are relevant to prove that some segment of the composite marks which both contesting parties use has a normally understood and well-recognized descriptive or suggestive meaning, leading to the conclusion that that segment is relatively weak." Id. at 654 (internal quotation marks omitted); see 2 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 11:90 (4th ed. 2013) (2 McCarthy).

More importantly, SinuCleanse is a weak mark because it is merely descriptive, rather than suggestive (as Med-Systems contends). Although "[t]he distinction between descriptive and suggestive marks is difficult" and "often based on intuitive reactions rather than analytical reasoning," we have endorsed a helpful test for distinguishing between the two categories: "suggestive terms are those which require the buyer to use thought, imagination, or perception to connect the mark with the goods," whereas "[d]escriptive terms are those which directly convey to the buyer the ingredients, qualities, or characteristics of the product." Educ. Dev. Corp. v. Econ. Co., 562 F.2d 26, 29 (10th Cir.1977) (emphasis added). "SinuCleanse" falls on the side of descriptiveness. It is a term previously unknown in the English language, but it is plainly a composite of the common words "sinus" and "cleanse," which, when fused together, straightforwardly communicate the nature and purpose of Med-Systems' products: they cleanse the sinuses. Indeed, Med-Systems' cofounder Dr. Heatley testified that the founders "wanted a name that described what it was" and that SinuCleanse "described better [than a previously selected name] the cleaning or cleansing process of putting a solution through the nose to clear your sinuses out." Aplt.App., Vol. 2 at 117. The ordinary consumer has little need for imagination to reach this conclusion.

True, the term "SinuCleanse" by itself may not tell consumers every detail of the associated products — for example, that a neti pot is used to pass saline solution through one's nose. But to be descriptive, a mark "need not directly convey all of a product's characteristics, uses, or functions"; rather, it "need only impart directly a crucial, important aspect of the product." Educ. Dev. Corp., 562 F.2d at 29 (emphasis added). In Educational Development Corp. the senior user had adopted the mark "`Continuous Progress'" for its educational product, "a rectangular box holding various cards, pamphlets, and cassette tapes," or what was "known in the educational publishing industry as a `laboratory.'" Id. at 28. We said that "Continuous Progress" was descriptive because (1) the term had been used among educators for decades to describe a concept of teaching that "allows a student to commence a program of study at a level best suited to his knowledge of the subject and to progress at a rate suited to his individual capability," id. at 27, and (2) the term therefore "directly conveyed to educators the most important characteristic of [the] product," id. at 29, even though it did not, by itself, "indicate the format, laboratory nature, or other features of the product," id. Similarly, in Bardahl Oil Co. v. Atomic Oil. Co. of Okla., Inc., 351 F.2d 148, 150 (10th Cir.1965), we said that it was "perfectly clear to us that the word `Savmotor,'" when used in connection with motor-oil concentrates and additives, was "descriptive of the qualities and the functions of the products."

"SinuCleanse" is at least as descriptive as "Continuous Progress" and "Savmotor." It clearly conveys the sinus-cleansing function of Med-Systems' products, their most salient property. That it does not fully explain how the products work does not prevent it from being a descriptive mark. We note that Med-Systems produced no evidence that its mark was suggestive rather than descriptive. We conclude that the term "SinuCleanse" is descriptive and thus has little conceptual strength.

b. Commercial Strength

Commercial strength is "the marketplace recognition value of the mark." King of the Mountain, 185 F.3d at 1093 (internal quotation marks omitted). Evidence of a mark's commercial strength can make up for conceptual weakness because a conceptually weak mark may become strong by virtue of acquired consumer awareness. "For example, marks such as AMERICAN airlines, PAYLESS shoe stores, FORD autos and KENTUCKY FRIED CHICKEN fast-food outlets would, at birth, have been characterized as inherently `weak' terms. But they have all become famous and well recognized by the consuming public." 2 McCarthy § 11:83, at 11-232. By the same token, a mark with conceptual strength may ultimately be weak if its commercial strength is negligible. See George & Co. LLC v. Imagination Entm't Ltd., 575 F.3d 383, 395-96 (4th Cir.2009) (assuming that the mark "LCR" was suggestive but ruling that it was weak because the record lacked adequate evidence that consumers associated it with plaintiff).

Commercial strength is a concept analogous to secondary meaning. Secondary meaning refers to the level of distinctiveness that a descriptive mark must attain in the minds of consumers before it is eligible for trademark protection. See Donchez v. Coors Brewing Co., 392 F.3d 1211, 1216 (10th Cir.2004). "To acquire secondary meaning, a descriptive mark must have been used so long and so exclusively by one producer with reference to his goods or articles that, in the trade and to that branch of the purchasing public, the mark has come to mean that the article is his product." J.M. Huber Corp. v. Lowery Wellheads, Inc., 778 F.2d 1467, 1470 (10th Cir.1985) (brackets and internal quotation marks omitted). The difference between commercial strength and secondary meaning is that the former is a range, while the latter is a threshold: a mark may enjoy anything from a high degree of commercial strength to a low degree, but either it has secondary meaning or it does not. See generally 2 McCarthy § 11:82. Nevertheless, "the same evidentiary findings" are generally probative of both; "[t]he stronger the evidence of secondary meaning, the stronger the mark, and the more likely is confusion."5 Vail Assocs., 516 F.3d at 866 (internal quotation marks omitted); see George & Co., 575 F.3d at 395 ("The second step [in the strength analysis] considers the mark's commercial strength, a concept similar to the `secondary meaning' inquiry considered in evaluating a mark's validity."); Herman Miller, Inc. v. Palazzetti Imports & Exports, Inc., 270 F.3d 298, 316 n. 8 (6th Cir.2001) ("When determining the commercial or marketplace strength of a mark, the courts look to the same kind of evidence of real world recognition of the mark as is used to decide the presence or absence of secondary meaning to determine whether a non-inherently distinctive designation is or is not a valid mark." (internal quotation marks omitted)).

The factors we have identified as helpful in evaluating secondary meaning are "direct evidence, such as consumer surveys or testimony from consumers," and "circumstantial evidence regarding: (1) the length and manner of [the mark's] use, (2) the nature and extent of advertising and promotion of the mark and (3) the efforts made in the direction of promoting a conscious connection, in the public's mind, between the name or mark and a particular product or venture." Donchez, 392 F.3d at 1218 (internal quotation marks omitted).

The evidence in district court did not indicate commercial strength. Med-Systems produced evidence that it has sold products under the SinuCleanse name since 1997 but it has wholly failed to show the nature and extent of its advertising and promotion of the mark or efforts to promote a conscious connection in the public's mind between its mark and its products. Evidence that its products had millions of users and that its products were sold through well-known retailers does not tell us whether the sales were stimulated by the mark. That its products were shown on the Oprah Winfrey Show on at least one occasion is not enough to establish that consumers thereafter connected the mark and the product. Even CEO Gallo's testimony that Med-Systems had set its advertising budget at 30% of sales since 1997 does not tell us how much was spent on advertising, and even if that figure had been established, "evidence of a mark's promotion [although] relevant to prove the strength of a mark, standing alone without a context ... may not be sufficient to prove that a mark is very strong." Vail Assocs., 516 F.3d at 867 (internal quotation marks omitted); see Bose Corp. v. QSC Audio Products, Inc., 293 F.3d 1367, 1375 (Fed.Cir.2002) ("Raw numbers of product sales and advertising expenses may have sufficed in the past to prove fame of a mark, but raw numbers alone in today's world may be misleading.... Consequently, some context in which to place raw statistics is reasonable."). We have no evidence of the amount of funds and effort expended by Med-Systems to acquire consumer recognition, the nature of any such advertising, or the effect of its actions.6 Indeed, Dr. Heatley conceded that "the average person on the street" would not "know what SinuCleanse is." Aplt.App., Vol. 2 at 131. And Med-Systems' marketing consultant had no data on whether the average consumer was aware of the brand. A jury could only guess whether SinuCleanse enjoys substantial consumer recognition. The commercial-strength component, like the conceptual-strength component, favors Water Pik.

3. Similarity of the Marks

Given the weakness of the SinuCleanse mark, the likelihood of confusion is small unless the challenged mark is very similar. See First Sav. Bank, 101 F.3d at 655. The requisite similarity is missing here.

"We test the degree of similarity between marks on three levels: sight, sound, and meaning." King of the Mountain, 185 F.3d at 1090. In conducting the comparison, we do not independently examine each syllable of the marks but consider "the marks as a whole as they are encountered by consumers in the marketplace." Id. (internal quotation marks omitted). This rule "is based on a common sense observation of customer behavior: the typical shopper does not retain all of the individual details of a composite mark in his or her mind, but retains only an overall, general impression created by the composite as a whole." 4 McCarthy § 23:41, at 23-213. Thus, in Universal Money Centers, 22 F.3d at 1531, in which both marks used the term "UNIVERSAL," we rejected the argument that because the additional words "MONEY CARD," "MONEY CENTER," and "MONEY" on the plaintiff's cards had been disclaimed by the PTO as descriptive, they should be omitted from the comparison. Likewise, in Sabinsa Corp. v. Creative Compounds, LLC, 609 F.3d 175, 181 (3d Cir.2010), where the issue was whether the defendant's use of the mark "Forsthin" in connection with a weight-loss product was likely to create confusion with the plaintiff's "ForsLean" mark, the Third Circuit held that the district court had erred in basing its comparison on only the components "thin" and "lean" rather than the marks as a whole. Id. at 184 (internal quotation marks omitted). We therefore compare the full words "SinuCleanse" and "SinuSense," not just the components "Cleanse" and "Sense."

Comparing "SinuCleanse" and "SinuSense" in appearance, we grant that the two marks have several visual similarities. Both appear in a serif font; both begin with "Sinu" in unitalicized type; both feature a capital letter after this term; and both end in "nse." But the italicization of "Cleanse" in Med-Systems' mark is a dramatic distinction, as Water Pik does not italicize "Sense" in its mark. As for the "sound" factor, both marks feature "Sinu," and in both marks the second component contains a short "e" sound. But in "SinuCleanse" the short "e" occurs between a hard "kl" sound and a "z" sound, whereas in "SinuSense" it comes between two "s" sounds. These differences are significant to the ear.

With respect to "meaning," the marks are less similar still. "SinuCleanse" conveys an accurate impression of what a sinus-irrigation product actually does — namely, cleanses the sinuses. By contrast, "SinuSense" is vague: whether the word "Sense" functions as a transitive verb, as in "sensing the sinuses," or as a noun, as in "good sense for the sinuses," it hardly offers a descriptive account of what the product does or how it works, instead leaving such matters to the consumer's imagination.

Overall, there is more than enough distinction between the two marks to prevent any likelihood of confusion.

Med-Systems argues that the report of one of its experts, Dr. David Yerkes, should change our analysis. Dr. Yerkes is a professor of English whose report explains, at great length and with reference to numerous authorities on linguistics and psychology, why the marks in this case are similar in sight, sound, and meaning. Although we do not doubt Dr. Yerkes's credentials as a scholar, the report covers a matter on which the trier of fact does not need expert advice. Cf. Fed. R.Evid. 702(a) (permitting expert testimony if "the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue" (emphasis added)). In Chesebrough-Pond's, Inc. v. Faberge, Inc., 666 F.2d 393, 397-98 (9th Cir.1982), the Ninth Circuit upheld a summary judgment on the issue of likelihood of confusion even though the plaintiff's expert, an associate professor of English, had opined that the words "`Match'" and "`Macho'" were similar in sight, sound, and meaning. The court said that "[t]he similarity or dissimilarity of these two words is a matter easily evaluated by laymen within the realm of their common knowledge and experience"; that "[i]t is highly doubtful that expert testimony on the subject, of the type presented by the [expert's] affidavit, would be of any real assistance to the trier of fact"; and that it would have been "well within the discretion of the trial judge to exclude" this evidence under Rule 702. Id. at 398. Even though the district court apparently had not excluded the report, the Ninth Circuit nevertheless ruled that because "the issue of fact concerned matters of common knowledge and experience," the court was free to draw its own conclusions on similarity and to grant summary judgment even though its opinion differed from that of the expert. Id. A leading treatise agrees with the Ninth Circuit's approach. See 4 McCarthy § 23:2.75, at 23-23 ("It is proper for a judge to give little weight to expert testimony on an obvious question, such as an analysis of how the appearance and pronunciation of the conflicting marks differ. Such expert evidence cannot prevent a summary judgment to the contrary of the expert's view."). We also agree.

Moreover, another consideration strongly supports a determination that the marks here are not similar. As we have explained, and as the district court noted in criticizing Questions 3 and 10 of Dr. Belch's survey, the SinuSense mark does not appear in isolation when used in commerce. Rather, it is routinely accompanied by the "waterpik" house mark. "Conflicting marks must be compared in their entirety, including any `house mark' which one party may append to its mark." Id. § 23:43, at 23-217 to 218. Thus, in Universal Money Centers we analyzed whether the defendant's "`AT & T Universal Card'" was similar in sight, sound, and meaning to the plaintiff's "UNIVERSAL MONEY," "UNIVERSAL MONEY CARD," and "UNIVERSAL MONEY CENTER" trademarks. 22 F.3d at 1531. Emphasizing that "each mark is to be considered as a whole," we noted "the distinctive AT & T house mark prominently displayed on the front of" the defendant's card. Id. The presence of this house mark figured heavily in our ultimate ruling that "the differences between the marks greatly outweigh[ed] any similarities." Id. Just as the presence of the house mark in Universal Money Centers helped persuade us that the defendant's mark was not similar to the plaintiff's, the presence of the "waterpik" house mark confirms our view that the similarity factor weighs in Water Pik's favor.

4. Water Pik's Intent

The district court ruled that Water Pik's intent favored Med-Systems because Water Pik knew of the SinuCleanse mark before adopting the brand name "SinuSense," and it developed its SinuSense brand after unsuccessfully trying to acquire Med-Systems' line of sinus-irrigation products. See Water Pik, 848 F.Supp.2d at 1273. The court noted in particular an email sent in 2008 by Michael Wakeman, vice president of marketing for Water Pik, which said: "[R]egarding squeeze bottle designs, let's pursue both a `hybrid' of the NeilMed/SinuCleanse designs and a `gooseneck' angular nozzle design (note the suggested improvements for each in the attached report)." Aplt.App., Vol. 3 at 775.

Water Pik argues that the district court erred in its analysis of this factor because it mistakenly focused on whether Water Pik intended to copy Med-Systems' products, not whether it intended to deceive consumers about the source of those products by copying Med-Systems' trademarks. Water Pik contends that Med-Systems has cited no evidence of this latter intent.

We agree with Water Pik's description of the intent that should be the central concern of this factor. When we have said that evidence of intent to copy may justify an inference of likelihood of confusion, we have been referring to copying a particular mark, not copying a competitor's product. See, e.g., Utah Lighthouse Ministry, 527 F.3d at 1055 ("Evidence that the alleged infringer chose a mark with the intent to copy, rather than randomly or by accident, typically supports an inference of likelihood of confusion." (emphasis added)); Sally Beauty, 304 F.3d at 973 ("Proof that a defendant chose a mark with the intent of copying the plaintiff's mark may, standing alone, justify an inference of likelihood of confusion." (emphasis added)). And more specifically, the intent in copying the mark should be "the intent to derive benefit from the reputation or goodwill of plaintiff." Id.; see Universal Money Centers, 22 F.3d at 1532 (stressing that the record contained "absolutely no facts to support an inference that AT & T adopted the word `Universal' with the intent to derive benefit from [the plaintiff's] marks").

This focus on an intent to cause confusion by adopting a similar mark reflects the language of the Lanham Act. The Act speaks in terms of whether the use of a mark is likely to confuse consumers, not necessarily whether the design of a product is likely to do so.7 Section 32 proscribes the "use in commerce" of "any reproduction, counterfeit, copy, or colorable imitation of a registered mark" that "is likely to cause confusion, or to cause mistake, or to deceive." 15 U.S.C. § 1114(1)(a) (emphasis added). Likewise, § 43(a) imposes liability on any person who "uses in commerce any word, term, name, symbol, device, or any combination thereof," that "is likely to cause confusion, or to cause mistake, or to deceive" as to affiliation, connection, origin, sponsorship, or approval of the person's product. Id. § 1125(a). Such language reflects an interest in preventing competitors from using deceptive marks to mislead consumers about the source of their goods or services. The relevance of intent is that one can infer a likelihood of confusion from a defendant's selection of a mark with the intent to cause confusion.

Our fellow circuits have made the same point. For example, in Streetwise Maps, Inc. v. VanDam, Inc., 159 F.3d 739, 741-42 (2d Cir.1998), the plaintiff appealed the rejection at a bench trial of its claim that its "Streetwise" trademark, used in connection with foldable, laminated street maps, was infringed by the defendant's use of "StreetSmart" in connection with similar maps. The district court found, among other things, that the defendant had not adopted its mark in bad faith; it explained that the defendant had "probably copied the size and concept of [the] Streetwise[] map, but not in an effort to mislead consumers into believing that they were buying" the plaintiff's product. Id. at 745. The Second Circuit agreed with this reasoning, noting: "The intent to compete by imitating the successful features of another's product is vastly different from the intent to deceive purchasers as to the source of the product." Id.

In A & H Sportswear, Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198, 206 (3d Cir.2000), the district court found after a bench trial that the defendant's use of "Miracle Bra" in connection with swimwear was not likely to cause confusion with the plaintiff's "Miraclesuit" mark, also used for swimwear. Addressing the district court's finding that the defendant had not adopted its mark with the intent to confuse consumers, the Third Circuit explained:

Id. at 225-26 (emphasis added). The appeals court held that the district court's finding on this factor was not clear error because the record indicated that the defendant had "brought the mark into swimwear because of its success in lingerie," not "in an effort to ride on [the plaintiff's] good name." Id. at 226.

And in Sensient Technologies Corp. v. SensoryEffects Flavor Co., 613 F.3d 754, 758-60 (8th Cir.2010), the Eighth Circuit, affirming a summary judgment that the defendant's "SensoryEffects" mark was not likely to cause confusion with the plaintiff's "Sensient Flavors" mark, agreed with the district court that even if the defendant had selected its mark with advance knowledge of the plaintiff's mark, such knowledge alone did not compel the conclusion that the defendant possessed the intent relevant to the likelihood-of-confusion analysis. See id. at 766-68. Instead, the appeals court emphasized that "knowledge of another's product and an intent to compete does not correspond with an intent to mislead." Id. at 767.

Med-Systems has failed to establish the requisite intent by Water Pik. Water Pik points to evidence that the process of choosing the SinuSense name was innocuous. Michael Wakeman, vice president of marketing for Water Pik, swore in an affidavit that Water Pik had "utilized the Sterling Rice Group and Blue Sky Strategies, two outside vendors, to assist it in creating, testing (through consumer focus groups and on-line surveys) and developing the SinuSense ™ brand." Aplt.App., Vol. 2 at 334. He testified in a deposition that the name "SinuSense" was first proposed by Steve Costello of Blue Sky Strategies and that it was one of multiple names vetted by Sterling-Rice. In an October 2009 email to Wakeman, Chris Lohmann of Sterling-Rice stated: "[Two colleagues] and I just met to review the names. We all feel strongly that SinuSense is by far the best option." Id., Vol. 3 at 720. One of Wakeman's emails earlier in the same email chain suggested that Wakeman actually wished to avoid using any terminology too similar to an existing mark: he explained that although "technically we could use Sinus Rinse/SinusRinse, we do not want to as NeilMed uses it prominently on their package." Id. at 721. Moreover, Water Pik added the Water Pik housemark prominently to all SinuSense packaging, suggesting an effort to avoid confusion rather than create it. This evidence supports Water Pik's contention that the SinuSense name arose not from an intent to confuse consumers by copying SinuCleanse but from "a thorough process ... to create a trademark." Aplee. Br. at 35.

Med-Systems' evidence does not contradict Water Pik's. Much of it suggests only that Water Pik intended to manufacture products that were similar to the SinuCleanse line and to sell them to similar consumers. For example, Water Pik apparently targeted most of the same distributors and retailers as Med-Systems; manufactured the same types of sinus-irrigation items and sold saline packets in the same denomination as Med-Systems; and designed its squeeze bottle to resemble a hybrid of Med-Systems' and NeilMed's. But none of this evidence relates to the SinuSense brand name; it would not allow a reasonable juror to infer that Water Pik intended to confuse consumers through its adoption and use of that brand name.

And Med-Systems reads too much into an email that Wakeman received in May 2009 from an executive assistant to Water Pik's CEO. The executive assistant wrote that although she liked the name "SinuSense," she was "concerned that people may confuse it with SinuCleanse." Aplt. App., Vol. 3 at 704. There is no evidence, however, that the Water Pik decisionmakers decided that there would be such confusion, or that they considered the confusion to be helpful rather than harmful.8

Given the weakness of the evidence of an improper intent, this factor is neutral.

5. Degree of Care Likely to Be Exercised by Purchasers

If consumers are likely to exercise a high degree of care in deciding whose product to buy, the likelihood of confusion is reduced. See Heartsprings, Inc. v. Heartspring, Inc., 143 F.3d 550, 557 (10th Cir.1998). The district court ruled that this factor favored Water Pik because Med-Systems did not dispute Water Pik's contention that consumers are likely to be more discriminating than usual when they purchase healthcare products. See Water Pik, 848 F.Supp.2d at 1276-77. Although Med-Systems' opening brief on appeal contains passing remarks that consumers may not exercise much care when shopping for sinus-irrigation products, these remarks are confined to the statement of facts, and Med-Systems does not challenge the court's ruling until its reply brief. "Arguments inadequately briefed in the opening brief are waived." Utah Lighthouse Ministry, 527 F.3d at 1049 n. 1. We will not revisit the district court's ruling on this factor. The high degree of care that consumers are likely to exercise favors Water Pik.

6. Similarity of Products and Manner of Marketing

The evidence in the district court established that Water Pik's sinus-irrigation products are very similar to Med-Systems' and are marketed through nearly identical channels. This factor favors Med-Systems.

7. Overall Likelihood of Confusion

Our analysis has shown that only the last factor — similarity of products and marketing — favors Med-Systems. The evidence of actual confusion strongly favors Water Pik. And the weakness of the SinuCleanse mark, combined with the notable respects in which the SinuSense mark (particularly when considered with the Water Pik house mark) differs from that mark, suggests that the absence of evidence of actual confusion is no accident. Med-Systems has failed to raise a genuine factual issue regarding likelihood of confusion. The district court properly awarded summary judgment to Water Pik.

III. CONCLUSION

We AFFIRM the district court's grant of summary judgment to Water Pik. We GRANT the stipulated motion to supplement Water Pik's supplemental appendix.