SUMMARY ORDER
Following a one-day bench trial in this trademark infringement case, plaintiff-counter-defendant-appellant Joules Limited (“Joules”) appeals from the district court’s judgment entered August 5, 2016 in favor of defendant-counter-claimant-appellee Macjfs Merchandising Group, Inc. (“MMG”). By opinion and order entered August 2, 2016, the court granted MMG’s request for a declaration that its MAISON JULES mark did not infringe the JOULES mark. We assume the parties’ familiarity with the underlying facts, procedural history, and issues on appeal.
BACKGROUND
Joules is an England-based retailer, founded by Tom Joule, that sells women’s clothing, shoes, and accessories to retailers in the United States. Joules owns U.S. *635Trademark Registration No. 3,696,021, which protects the use of the JOULES mark in connection with women’s clothing, shoes, and accessories. The United States Patent & Trademark Office (“USPTO”) issued the registration in October 2009 without proof of secondary meaning. Joules has used the mark in the marketplace since 2004 and on its product tags, clothing labels, and websites, www.joules.com and www.joulesusa.com, primarily in the stylized script shown below:
[[Image here]]
S. App. 3.
MMG is a New York-based subsidiary of Macy’s, Inc. (“Macy’s”), which operates department stores throughout the United States and sells products online at www. macy’s.com. MMG designed “Maison Jules” as a private brand line of clothing for Macy’s. MMG selected the name “Mai-son Jules” (1) to evoke an association with French fashion, as “maison” means “house” in French and is used to refer to French fashion houses, and the popular French film “Jules and Jim"; (2) for the dual-gender quality of the name “Jules,” so the brand could expand into men’s apparel; and (3) to create an easily pronounceable name that was unlike the name of other Macy’s brands targeting the same consumers. After the Maison Jules line launched in July 2013, MMG used the MAISON JULES mark on product tags, clothing labels, Macy’s website, and physical objects in the Maison Jules section in Macy’s stores. MMG has used the mark in the marketplace in this partially scripted form:
[[Image here]]
S. App. 8.
The Joules and Maison Jules product lines target similar audiences and are generally sold at separate locations. The Joules line targets women consumers between the ages of twenty-five and forty and the Maison Jules line targets women consumers between the ages of eighteen and thirty. Both lines feature workplace- and weekend-appropriate clothing of comparable price and quality. Joules clothing is sold in retail stores such as Nordstrom, and online at www.joulesusa.com and third-party websites such as www.amazon. com. No Joules products are sold in Macy’s stores or at www.macys.com. In contrast, Maison Jules clothing is sold almost exclusively in Macy’s stores and at www.macys.com. A small subset of Maison Jules products is sold at third-party websites such as www.amazon.com.
In 2012, MMG performed a trademark search for conflicting marks and found a result for Joules’s U.K. website but not its U.S. website. MMG filed an intent-to-use application and sought to register the *636MAISON JULES mark for use in connection with women’s clothing and other goods. In early 2013, Joules sent cease- and-desist letters to Macy’s and MMG requesting withdrawal of the application. The USPTO issued a statement that it found no conflicting marks that would bar registration of the MAISON JULES mark and issued a notice of publication. Joules filed a Notice Opposition and, in 2015, filed the complaint in this action.
On August 2, 2016, the district court held, after a bench trial, that there was no likelihood of confusion between the JOULES and MAISON JULES marks. It entered judgment for MMG. This appeal followed.
DISCUSSION
“When reviewing a judgment following a bench trial, we review a district court’s findings of facts for clear error and its conclusions of law de novo.” Process Am., Inc. v. Cynergy Holdings, LLC, 839 F.3d 125, 141 (2d Cir. 2016). We will not find clear error in the findings “[w]here there are two permissible views of the evidence.” Id. (alteration in original) (quoting Diesel Props S.R.L. v. Greystone Bus. Credit II LLC, 631 F.3d 42, 52 (2d Cir. 2011)).
A plaintiff asserting either a trademark infringement claim under federal or New York law or a federal claim of unfair competition must establish a likelihood of consumer confusion as to the source or sponsorship of defendant’s products. 1-800 Contacts, Inc. v. WhenU.Com, Inc., 414 F.3d 400, 406-07 (2d Cir. 2005) (federal trademark infringement); EMI Catalogue P’ship v. Hill, Holliday, Connors, Cosmopulos Inc., 228 F.3d 56, 61 (2d Cir. 2000) (federal unfair competition); Std. & Poor’s Corp. v. Commodity Exch., Inc., 683 F.2d 704, 708 (2d Cir. 1982) (trademark infringement under federal and New York law). We assess the likelihood of confusion under the Polaroid factors: (1) the strength of plaintiffs mark, (2) the similarity of the two marks, (3) the competitive proximity of the products, (4) the likelihood that plaintiff will “bridge the gap” and offer a product like defendant’s, (5)-actual confusion between the products, (6) defendant’s good faith, (7) the comparable quality of defendant’s product, and (8) purchaser sophistication. Streetwise Maps, Inc. v. VanDam, Inc., 159 F.3d 739, 743 (2d Cir. 1998) (quoting Polaroid Corp. v. Polarad Elecs. Corp., 287 F.2d 492, 495 (2d Cir. 1961)). We review the findings as to each factor for clear error and the balancing of the factors de novo. Starbucks Corp. v. Wolfe’s Borough Coffee, Inc., 588 F.3d 97, 105 (2d Cir. 2009).
The district court found that the strength, similarity, competitive proximity, actual confusion, and good faith factors favored MMG; the quality factor favored Joules; and the likelihood of bridging the gap and sophistication factors were neutral, It balanced the factors, held there was no likelihood of confusion, and dismissed Joules’s claims. On appeal, Joules challenges its assessment of the competitive proximity, actual confusion, strength, and similarity factors.
A. Competitive proximity
Competitive proximity “’concerns whether and to what extent the two products compete with each other’ and ’the nature of the products themselves and the structure of the relevant market.’” Morningside Grp. Ltd. v. Morningside Capital Grp., L.L.C., 182 F.3d 133, 140 (2d Cir. 1999) (quoting Cadbury Beverages, Inc. v. Cott Corp., 73 F.3d 474, 480 (2d Cir. 1996)). “Among the considerations germane to the structure of the market are the class of customers-to whom the goods are sold, the manner in which the products are advertised, and the channels through which the goods are sold.” Cadbury, 73 F.3d at 480.
*637The district court found that the competitive proximity factor weighed “clearly in favor of MMG.” S. App. 17. We do not agree. Both sets of products are sold in brick-and-mortar department stores (albeit not the same ones); sold online through the parties’ respective websites; and, for a small subset of both products, sold online through the same or similar third-party websites. Both product lines target and are sold to young women who seek work- and weekend-appropriate clothing and who are likely to visit both sets of sales locations. These considerations show competitive proximity between the products, even if the vast majority of products are not sold side-by-side in the same stores or on the same websites. See Patsy’s Brand, Inc. v. I.O.B. Realty, Inc., 317 F.3d 209, 218 (2d Cir. 2003) (finding competitive proximity, even though products were sold in separate stores, because products appealed to same consumers, stores were geographically close, and consumers were reasonably likely to visit both stores).
The different nature of the products, however, tempers the finding of competitive proximity. Although both product lines feature clothing of comparable quality and price, the court found that “a significant percentage” of Joules sales in the United States are for rain boots and rainwear, the majority of large national retailers carrying Joules products carry only their rain boots and rainwear, and there are no rain boots or rainwear in the Maison Jules line. S. App. 17. This evidence indicated that a significant portion of the Joules products sold are different in nature from the Mai-son Jules products sold, a conclusion that weakens the finding of competitive proximity.
On balance, we conclude the competitive proximity factor weighs slightly in Joules’s favor. Nonetheless, assuming the district court clearly erred in finding that the factor clearly favored MMG, as discussed below, we agree with its overall conclusion as to the likelihood of confusion.
B. Actual confusion
Evidence of actual confusion is “convincing evidence that confusion is likely.” Morningside, 182 F.3d at 141. The court concluded that MMG’s survey evidence of confusion was more credible than Joules’s evidence and that the actual confusion factor clearly favored MMG. This finding was not clearly erroneous.
“To ,be probative, a survey must ’have been fairly prepared and its results directed to the relevant issues.’” Sterling Drug, Inc. v. Bayer AG, 14 F.3d 733, 741 (2d Cir. 1994) (quoting Universal City Studios, Inc. v. Nintendo Co., 746 F.2d 112, 118 (2d Cir. 1984)). Here, the court found Joules’s survey evidence flawed and its finding of 24.3% confusion unpersuasive because (1) the line-up method forced participants to consider the marks in artificially close proximity; (2) the Maison Jules website shown to participants had been scrubbed of all Macy’s references and context; (3) the control stimulus was too dissimilar to the Maison Jules website; and (4) participants were'not limited to potential Maison Jules consumers, see id. (noting that where there is confusion as to source of junior user’s products, survey participants should be junior user’s consumers). The court found that MMG’s survey evidence more closely replicated market conditions and that its finding of no confusion was more reliable, because it did not use the line-up method, tested for both in-store and online confusion, and corrected for errors arising from the low level of recognition of the JOULES mark in the marketplace. Because there is evidentiary support for the court’s reading of the survey evidence and because we will not “second-guess the court’s credibility assessments,” Process Am., 839 F.3d at 141 (quoting Diesel Props, 631 F.3d at 52), we conclude there *638was no clear error in the court’s analysis of the survey evidence.
Accordingly, it was not clearly erroneous for the district court to find that the actual confusion factor weighed clearly in MMG’s favor.
C. Strength of the JOULES mark
“The strength of a mark is determined by its tendency to uniquely identify the source of the product.” Star Indus., Inc. v. Bacardi & Co., 412 F.3d 373, 384 (2d Cir. 2005). “To gauge a mark’s strength, we consider two factors: its inherent distinctiveness, and its distinctiveness in the marketplace.” Streetwise, 159 F.3d at 743.
We agree with the district court that the strength factor weighs in MMG’s favor. First, the JOULES mark is inherently strong because the USPTO registered it without proof of secondary meaning, so we presume “that the mark is more than merely descriptive” and “is inherently distinctive.” Lane Capital Mgmt., Inc. v. Lane Capital Mgmt., Inc., 192 F.3d 337, 345 (2d Cir. 1999). Second, there is evidence of the mark’s weakness in the marketplace. Joules’s expert opined that only 16.2% of participants in his survey, which targeted women between the ages of sixteen and thirty-five who had recently shopped for clothing, had seen, heard of, or purchased Joules products. Joules admitted that at least twelve other trademark registrations and websites use “Jules” or a homophone to offer products for sale in the market for women’s clothing. The court did not clearly err in finding that the commercial weakness of the JOULES mark undercut its inherent strength. Thus, we find no clear error in the finding that the strength factor favored MMG.
D. Similarity of marks
“In assessing similarity, courts look to the overall impression created by the logos and the context in which they are found and consider the totality of factors.” Star Indus., 412 F.3d at 386 (quoting Gruner + Jahr USA Publ’g, a Div. of Gruner + Jahr Printing & Publ’g Co. v. Meredith Corp., 991 F.2d 1072, 1078 (2d Cir. 1993)). A side-by-side comparison is a useful means of analyzing similarities in the respective designs, so long as the focus is on the ultimate issue of the likelihood of consumer confusion. Louis Vuitton Malletier v. Dooney & Bourke, Inc., 454 F.3d 108, 117 (2d Cir. 2006).
The district court concluded that the JOULES mark and the composite MAI-SON JULES mark were more dissimilar than similar. This finding was not clearly erroneous. Although the marks share similarly stylized scripts, use similar spellings for the words “Jules” and “Joules,” and are not facially distinguishable in meaning, consumers can nevertheless distinguish the composite MAISON JULES mark from the JOULES mark based on (1) sound because the dissimilar word “mai-son” appears first in the MAISON JULES mark, (2) the block-text typeface for the word “maison” ⅛ the MAISON JULES mark, and (3) the differences in overall appearances when the marks are considered in their full contexts. See Streetwise, 159 F.3d at 744-45 (finding dissimilarity because “[w]hile the two names sound similar, the trademarks themselves are not confusingly similar, given the context in which a purchaser sees them,” and identifying differences in product features, logos, dress colors, typefaces, font sizes, and the number of words in each mark). Accordingly, the court did not clearly err in finding the similarity factor favored MMG.
E.Balancing of factors
To summarize, four Polaroid factors (strength, similarity, actual confusion, and good faith) favor MMG, two factors (competitive proximity and quality) favor *639Joules, and two factors (likelihood of bridging the gap and purchaser sophistication) are neutral. As discussed, the competitive proximity factor does not strongly favor Joules due to differences in product nature, but the actual confusion factor clearly favors MMG. Upon our de novo review of the balancing of factors, we agree with the district court that the overall balance favors MMG and that Joules has therefore not demonstrated a likelihood of consumer confusion.
We have considered Joules’s remaining arguments and find them to be without merit. Accordingly, we AFFIRM the judgment of the district court.