COLE, Chief Judge.

ARGUED: C. Enrico Schaefer, TRAVERSE LEGAL, PLC, Traverse City, Michigan, for Appellant. David Brafman, AKERMAN LLP, West Palm Beach, Florida, for Hall Appellees. Michael B. Garfinkel, PERKINS COIE LLP, Los Angeles, California, for Appellee William Morris Endeavor. Eric J. Shimanoff, COWAN LIEBOWITZ & LATMAN, P.C., New York, New York, for Appellee UMG Recordings. ON BRIEF: C. Enrico Schaefer, Mark G. Clark, TRAVERSE LEGAL, PLC, Traverse City, Michigan, for Appellant. David Brafman, AKERMAN LLP, West Palm Beach, Florida, for Hall Appellees. Michael B. Garfinkel, PERKINS COIE LLP, Los Angeles, California, for Appellee William Morris Endeavor. Eric J. Shimanoff, COWAN LIEBOWITZ & LATMAN, P.C., New York, New York, for Appellee UMG Recordings.

Before: COLE, Chief Judge; DAUGHTREY and MOORE, Circuit Judges.

OPINION

COLE, Chief Judge.

Lee Jason Kibler, a disc jockey, brought federal trademark infringement, related state law, and federal trademark dilution claims against Robert Bryson Hall, II, a rapper, and professional entities supporting Hall's work. The district court granted summary judgment to defendants on all claims. Kibler has appealed that judgment, requiring us to answer two questions. First, has Kibler provided evidence sufficient to find that relevant consumers are likely to confuse the sources of his and Hall's products? Second, has Kibler provided evidence sufficient to find that Hall has diluted Kibler's mark? We conclude no and thus affirm the grant of summary judgment.

I. BACKGROUND

Kibler uses turntables and others' vocals to produce music containing jazz and funk elements, among others. He has performed and released several albums under the name "DJ LOGIC" since 1999 though he currently has no record deal. Kibler registered "DJ LOGIC" as a trademark in 2000, allowed the registration to lapse in 2003, and re-registered the name in 2013. He has also been known as just "LOGIC."

Hall has performed under the name "LOGIC" since 2009. He previously used the names "Young Sinatra" and "Psychological." Three Oh One Productions is Hall's personal company and Visionary Music Group his management company (with Hall, "the Hall defendants"). UMG Recording d/b/a Def Jam Recordings ("Def Jam") is Hall's record label and William Morris Endeavor Entertainment ("WME") his booking agent.

In September 2012, Kibler's attorney sent Visionary Music Group and WME an email ordering them to stop using the name "LOGIC" and to recall any product or advertisement that did. The attorney maintained that such use infringed on Kibler's mark. The next month, Three Oh One Productions applied to register "LOGIC" as a trademark.

In January 2014, Kibler filed suit against the defendants in the U.S. District Court for the Eastern District of Michigan. He alleged the following claims: 1) trademark infringement in violation of the Lanham Act, 15 U.S.C. § 1125(a) (2012) ; 2) breach of the Michigan Consumer Protection Act ("MCPA"), Mich. Comp. Laws § 445.901 –.922 (1977); 3) unfair competition under Michigan law; and 4) trademark dilution in violation of the Lanham Act, 15 U.S.C. § 1125(c) (2012).

In March 2014, defendants delayed Hall's tour and first album release due to ongoing settlement negotiations that ultimately collapsed. Def Jam proceeded to release the album in October of that year. It sold over 170,000 copies.

In May 2015, defendants moved for summary judgment on all of Kibler's claims. The parties fully briefed the matter and the district court held a hearing. In November 2015, the court granted defendants' motion in all respects.

II. ANALYSIS

A. Standard of Review

We review a district court's grant of summary judgment de novo. Med. Mut. of Ohio v. k. Amalia Enters. Inc. , 548 F.3d 383, 389 (6th Cir. 2008) ; Daddy's Junky Music Stores, Inc. v. Big Daddy's Family Music Ctr. , 109 F.3d 275, 280 (6th Cir. 1997) (clarifying that the rule holds in trademark infringement cases). Summary judgment is appropriate when the record shows "no genuine dispute as to any material fact." Fed. R. Civ. P. 56(c). In other words, we affirm summary judgment when there is no evidence that would allow a reasonable jury to find for the nonmoving party, entitling the moving party to judgment as a matter of law. See Anderson v. Liberty Lobby, Inc. , 477 U.S. 242, 248, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986) ; Fed. R. Civ. P. 56(c). We view all facts and inferences in the light most favorable to the nonmoving party. Matsushita Elec. Indus. Co., Ltd. v. Zenith Radio Corp. , 475 U.S. 574, 587, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986). B. Trademark Infringement

Kibler has made no separate arguments for his state law claims, and they rely on the same allegations as his federal trademark infringement claim. For these reasons, we address the state law claims along with the trademark infringement claim.

This court considers whether trademark infringement has occurred using a two-step test. First, we determine whether plaintiff's mark is protectable. Innovation Ventures, LLC v. N.V.E., Inc. , 694 F.3d 723, 728 (6th Cir. 2012). Then, we assess whether relevant consumers are likely to confuse the sources of the parties' products. Id. ; Homeowners Grp., Inc. v. Home Mktg. Specialists, Inc. , 931 F.2d 1100, 1107 (6th Cir. 1991). The relevant consumers are potential buyers of defendant's products. Maker's Mark Distillery, Inc. v. Diageo N. Am., Inc. , 679 F.3d 410, 419 (6th Cir. 2012). Here, the parties agree Kibler's mark is protectable. So we focus on the likelihood that potential buyers of rap would believe Kibler's music is Hall's or vice-versa.

In assessing the likelihood of confusion, we take into account the following eight " Frisch " factors: 1) strength of the plaintiff's mark, 2) relatedness of the products, 3) similarity of the marks, 4) evidence of actual confusion, 5) parties' marketing channels, 6) likely degree of purchaser care, 7) defendant's intent in selecting the mark, and 8) probability that the product lines will expand. CFE Racing Prods., Inc. v. BMF Wheels, Inc. , 793 F.3d 571, 592 (6th Cir. 2015) (citing Frisch's Rest., Inc. v. Shoney's Inc. , 759 F.2d 1261, 1264 (6th Cir. 1985) ).

Plaintiff need not establish each factor to prevail. Id. Each case is unique, so not all of the factors will be helpful. Homeowners , 931 F.2d at 1107. Further, there is no designated balancing formula for the factors. CFE Racing , 793 F.3d at 592. "The[ir] enumeration is meant 'merely to indicate the need for weighted evaluation of the pertinent facts in arriving at the legal conclusion of confusion.' " Id. (quoting Frisch , 759 F.2d at 1264 ).

1. Strength of Plaintiff's Mark

The first Frisch factor favors defendants. While Kibler has shown that "DJ LOGIC" is moderately strong conceptually, he has failed to provide evidence of the mark's commercial strength.

The stronger a mark is, the greater the risk of confusion. Homeowners , 931 F.2d at 1107. A mark cannot be strong unless it is both conceptually and commercially strong. Maker's Mark , 679 F.3d at 419. And it cannot be conceptually strong unless it is inherently distinctive. Id. Arbitrary marks, which convey something unrelated to the product they announce, e.g. , the "Apple" in "Apple computers," are distinctive. Therma–Scan, Inc. v. Thermoscan, Inc. , 295 F.3d 623, 631 (6th Cir. 2002) ; see also Maker's Mark , 679 F.3d at 420 (finding red dripping wax seal announcing bourbon inherently distinctive, and hence conceptually strong). Descriptive marks, which describe the product they announce, are usually indistinctive. See, e.g. , Therma–Scan , 295 F.3d at 632 (finding "Therma–scan," which describes the services plaintiff performs, indistinctive, and hence conceptually weak).

Further, courts presume that an incontestable mark is conceptually strong. Daddy's , 109 F.3d at 282. A mark is incontestable when it has not been successfully challenged within five years of its registration. Id. ; see Dieter v. B & H Indus. of Sw. Fla., Inc. , 880 F.2d 322, 328 (11th Cir. 1989) (explaining that merely descriptive marks cannot be registered as trademarks unless they have acquired a secondary meaning).In this case, the district court found that "DJ LOGIC" is moderately strong conceptually. The court reasoned that while "DJ" describes Kibler's craft, "LOGIC" is not even "suggestive of the characteristics of [his] music." Ki bler v. Hall , No. 14–10017, 2015 WL 6865928, at *2 (E.D. Mich. Nov. 9, 2015). Defendants concede this. Kibler contends only that the court erred in not considering the mark's incontestability. We need not address this argument because we agree with the district court's assessment, which renders "DJ LOGIC" at least as conceptually strong as a finding of incontestability would.

But a mark can be conceptually strong without being commercially strong, and thus weak under Frisch . Maker's Mark , 679 F.3d at 419–20. A mark's commercial strength depends on public recognition, the extent to which people associate the mark with the product it announces. Id. at 419.

Survey evidence is not a prerequisite for establishing public recognition, but it is the most persuasive evidence of it. See, e.g. , id. at 421 (characterizing proof of "extensive marketing" and "widespread publicity" around a mark as abundant evidence of public recognition); Frisch , 759 F.2d at 1265 (relying on evidence that around thirty percent of respondents identify non-plaintiff restaurants with plaintiff's mark to conclude mark is commercially weak). Proof of marketing is not a prerequisite either. Therma–Scan , 295 F.3d at 632. But plaintiffs lacking such proof must provide other evidence of "broad public recognition." Id.

Conversely, proof that third parties have extensively used a trademark or similar trademarks in the relevant market indicates the trademark is commercially weak. Homeowners , 931 F.2d at 1108. The presumption is that the third parties have muddled the mark's source. In Homeowners , for example, defendant offered evidence that many "real estate related firms" were using trademarks identical or similar to plaintiff's. Id. We found that a reasonable jury could determine that the use had weakened plaintiff's mark because plaintiff sold to real estate brokers and defendant sold to real estate sellers. Id. at 1103, 1108. Compare id. at 1108with Maker's Mark , 679 F.3d at 420–21 (denying third-party use of similar marks has weakened plaintiff's mark because it occurred among all distilled spirits rather than the relevant market of tequila).

Here, the district court concluded that "DJ LOGIC" is commercially weak. The court cited Kibler's lack of survey or marketing evidence and limited commercial success. Kibler , 2015 WL 6865928, at *3 (noting sale of fewer than 300 albums in past three years and fewer than 60,000 albums in past sixteen years; current lack of a recording contract; and inability ever to secure a recording contract with a major label). The court found that third parties have weakened the mark even further by marketing music under nearly ninety variations of "logic."

Kibler admits he offered no survey evidence, but claims that the district court treated it as a prerequisite. He further argues he provided marketing evidence. First is a sworn declaration that he advertises in print and online, including on MySpace, Twitter, and Facebook. Second are a 2006 Downbeat article featuring him, a 2001 New York Times review mentioning him, and a 1999 Gig article featuring him. Third is a sworn declaration that he has appeared on television shows such as The Tonight Show Starring Jimmy Fallon, The Today Show, and Good Morning America. Kibler also points to his tours and online music sales as proof of marketing. Additionally, Kibler insists he is commercially successful, noting that there is no fixed number of album sales establishing commercial success. Kibler denies that third parties have weakened his mark.

Defendants reinforce the district court's findings. Def Jam and WME argue that Kibler's failure to provide the number of his Facebook "likes" or Twitter followers creates an adverse inference, dismiss the publications as obscure and out-of-print, and question the number of people who have attended Kibler's concerts. All of the defendants highlight Kibler's deposition testimony that he appeared on the television shows to support other, headlining artists.

The district court properly found that Kibler's evidence would prevent a reasonable jury from concluding that "DJ LOGIC" is commercially strong. But its analysis was incomplete and at times flawed. The court did not treat survey evidence as a prerequisite for establishing commercial strength. Rather, it also considered whether Kibler had provided marketing evidence. The court erred, however, in finding that he had not. Promotion on platforms such as Twitter and Facebook not only constitutes marketing, but is among the most popular and effective advertising strategies today. And whether publicity like magazine interviews and television appearances constitutes marketing or a separate form of evidence, it speaks to commercial strength. See Maker's Mark , 679 F.3d at 421.

But some proof is not enough. Kibler must offer evidence that would permit a reasonable jury to determine that wide segments of the public recognize "DJ LOGIC" as an emblem of his music. This means "extensive" marketing and "widespread" publicity around the music and mark. Id. Kibler's evidence may not create an adverse inference of broad recognition, but it lacks the information jurors would need to find such awareness. For instance, how many and what kind of Twitter followers does Kibler have? A large number of followers, or celebrities likely to re-tweet Kibler's messages to their large number of followers, for example, would suggest that many types of people know his work and mark. We can say the same of the number and kind of Kibler's Facebook fans, likes, posts, and re-posts.

Similarly, Kibler fails to provide the circulations or target audiences of Downbeat and Gig , which appear to be niche publications. Further, the New York Times review focuses on two other artists, placing "DJ Logic" in a series of supporting musicians. This leaves a slim chance that readers noticed and recalled Kibler. In any event, both the Gig article and New York Times review are over fifteen years old. Even if they suggested broad recognition, Kibler would have to show continuing awareness of his mark to justify a likelihood of confusion.

Kibler has neither refuted nor explained his deposition testimony that he appeared on television shows to support other, headlining artists. For instance, he testified that Carly Simon, "the main act," introduced "the guests she had playing with her" on the Fallon show. (Kibler Dep., R. 92–3, PageID 2930.) We do not know how many guests there were, if Simon introduced them individually, if she said anything other than their names, etc. Kibler did not need to address each of these considerations. But they indicate the sort of information a jury would need to assess the extent to which the public affiliates "DJ LOGIC" with Kibler's music.

Finally, Kibler's performances and songs are his products, not advertisements or publicity. Artists may attract consumers directly through their work, as when someone enjoys a musician's concert enough to then buy the music online. But treating the products that advertisements are meant to sell as advertisements themselves would mean finding marketing proof in virtually every infringement action, making the consideration superfluous. This is not what Frisch intended. Viewing the evidence in the light most favorable to Kibler, we take his tours and online sales as proof of his commercial success, discussed below.

The district court rightly found that Kibler has enjoyed limited commercial success and that this implies that "DJ LOGIC" is not broadly familiar. But the court's analysis was incomplete. Album sales and even recording contracts are less critical markers of success than before because of widespread internet use. As a result, a plaintiff with low album sales or no representation could nevertheless show commercial success suggesting broad recognition of his mark using web-based indicators of popularity, e.g. , YouTube views. Because Kibler has not done that, we have only his low album sales, current lack of a recording contract, and inability ever to secure a recording contract with a major label. Kibler declares that he has participated "in hundreds of live performances held in at least 46 states," but he does not indicate the number of people who attended, the number of other artists involved, and whether he ever received top billing. (Kibler Decl., R. 91–1, PageID 2702.) Kibler's silence on his popularity online and general statement about his performances do not allow for a finding that "most people will be familiar" with "DJ LOGIC." Therma–Scan , 295 F.3d at 632.

"DJ LOGIC" lacks commercial strength though we find no proof that third parties have weakened it. Defendants identify the parties' marks as trademarks in their brief, but do not show they are registered. See AutoZone, Inc. v. Tandy Corp. , 373 F.3d 786, 794 (6th Cir. 2004) (involving evidence of 745 trademarks using "ZONE"); Induct–O–Matic Corp. v. Inductotherm Corp. , 747 F.2d 358, 362 (6th Cir. 1984) (involving evidence of numerous trademarks using "induct").

Nor do defendants show that the third parties use the marks in the relevant market. Defendants imply that the market is music sold in the U.S. over Amazon and iTunes. But that is far too broad a market to assume that marks similar or even identical to "DJ LOGIC" weaken it simply by inhabiting the same space. In Maker's Mark , we denied that all distilled spirits was narrow enough of a field to conclude that similar marks in that industry had weakened the mark of a tequila brand. 679 F.3d at 420–21. We found that tequila itself was the relevant market. Id. Likewise, the relevant market here is not countless types of music or even hip-hop, but DJ music sold in the U.S. over Amazon and iTunes. Because defendants have not shown which, if any of the marks, operate in that market, no reasonable jury could find the marks have weakened "DJ LOGIC."

Because the record reflects that "DJ LOGIC" is moderately strong conceptually, but weak commercially, the first Frisch factor favors defendants.

2. Relatedness of Products

This court uses the following test to decide whether relatedness favors either party: 1) if the parties' products compete directly with each other, consumer confusion is likely if the parties' marks are sufficiently similar; 2) if the products are somewhat related, but do not compete directly, the likelihood of confusion will depend on other factors; 3) if the products are completely unrelated, confusion is unlikely. Daddy's , 109 F.3d at 282.

Products belonging to the same industry are not necessarily related. Homeowners , 931 F.2d at 1109. To be related, they must be marketed and consumed in ways that lead buyers to believe they come from the same source. Id. Take Therma–Scan , in which we found two thermology services unrelated enough that confusion was unlikely. 295 F.3d at 633. We noted that the parties marketed the services to different populations. Id. Compare id. with Maker's Mark , 679 F.3d at 423 (concentrating on other Frisch factors after finding both products high-end distilled spirits, but not directly competitive given their price differential); Little Caesar Enters., Inc. v. Pizza Caesar, Inc. , 834 F.2d 568, 571 (6th Cir. 1987) (finding Italian food services related enough to cause confusion because both concentrate on pizza).

The district court found the relatedness factor neutral insofar as the parties' products are somewhat related, but not directly competitive. The court reasoned that while both Kibler and Hall perform and sell music, only Hall uses his vocals. Kibler maintains that the factor favors him based on proof that he and Hall both sell hip-hop incorporating turntables and rap. Kibler refers to print and online media about Hall, much of which affiliates him with hip-hop and all of which describes him as a rapper.

The district court correctly found this factor neutral because the record supports that the parties' products are somewhat related, but not directly competitive. The most relevant evidence is a booking notice describing Hall as a "hot upcoming rapper" and two online ads featuring Hall holding a microphone. (Booking Notice, R. 91–4, PageID 127.) They indicate that while both are musicians and perhaps hip-hop artists, Hall markets himself as a rapper and Kibler a disc jockey. The parties' products are comparable to the bourbon and tequila goods we found only to belong to the same broad category of high-end distilled spirits in Maker's Mark , 679 F.3d at 423. Incidental overlap of their customers could not sustain a finding of direct competition at trial. Id. at 421 (affirming district court's conclusion that products only somewhat related despite district court's finding that indeterminate number of defendant's customers likely patronize plaintiff given drinkers' habits); see Maker's Mark Distillery, Inc. v. Diageo N. Am., Inc. , 703 F.Supp.2d 671, 692 (W.D. Ky. 2010). Accordingly, the factor is neutral.

3. Similarity of Marks

The more similar the marks are, the more likely it is that relevant consumers will confuse their sources. We determine the similarity of marks by considering whether either mark would confuse a consumer who did not have both marks before her and had only a vague impression of the other mark. Daddy's , 109 F.3d at 283. We consider the marks' pronunciation, appearance, and verbal translation. Id. ; see, e.g. , Maker's Mark , 679 F.3d at 421–22 (finding factor favors plaintiff because marks are facially similar and some companies offer several kinds of distilled spirits).

The anti-dissection rule requires us not to dwell on the prominent features of a mark and instead consider it as a whole. Little Caesar , 834 F.2d at 571–72 ; see, e.g. , Jet, Inc. v. Sewage Aeration Sys. , 165 F.3d 419, 423–24 (6th Cir. 1999) (finding distinctions in appearance, syllables, language, and pronunciation prevent "JET" and "AEROB–A–JET" from being confusingly similar despite their common word); Little Caesar , 834 F.2d at 572 (finding differences in sound, appearance, and syllables distinguish "Little Caesar" from "Pizza Caesar USA" despite the prominent word they share).

The district court concluded this factor favors defendants based on the anti-dissection rule. The court acknowledged that both marks include the prominent word "logic." Then it noted that the " 'DJ' portion not only changes the look and sound of the mark but also describes or suggests certain characteristics of [Kibler's] music." Kibler , 2015 WL 6865928, at *3. Kibler claims that the district court misapplied the anti-dissection rule in two ways. First, it overvalued "DJ," which is merely descriptive. Second, it neglected to compare just "LOGIC," which Kibler has also gone by, to Hall's "LOGIC." Kibler argues that both approaches conflict with Daddy's . Defendants accept the district court's findings.

The district court properly found this factor favors defendants by correctly applying the anti-dissection rule. This meant examining "DJ LOGIC" as a whole, including its appearance, sound, language, and impression. Kibler's call for this court to "focus on the dominant features of each mark and disregard the non-dominant features" is precisely what the anti-dissection rule forbids. Kibler Opening Br. 19. Further, the "DJ" in "DJ Logic" is more distinctive than the "Family Music Store" in "Big Daddy's Family Music Store," which did raise an issue of fact as to whether "Daddy's" was sufficiently similar. Daddy's , 109 F.3d at 284.

The district court also correctly declined to compare Kibler's "LOGIC" to Hall's "LOGIC." Kibler's reliance on Daddy's is again misplaced. There, we faulted the district court for not comparing just the "Daddy's" in "Daddy's Junky Music Stores" with defendant's mark because "Daddy's" itself was a separate trademark. Id. at 278, 284. Here the parties agree that Kibler has not registered "LOGIC" alone as a trademark. Thus, the anti-dissection rule requires the similarity of marks factor to favor defendants here.

4. Evidence of Actual Confusion

Evidence of actual confusion is the strongest proof of likely confusion. Frisch , 759 F.2d at 1267. So any such evidence favors the non-movant. Therma–Scan , 295 F.3d at 635. But the weight we give that evidence depends on the amount and type of confusion. Id. at 634. On one end of the spectrum are persistent mistakes and confusion by actual customers. Homeowners , 931 F.2d at 1110. On the other are relatively few instances of confusion and inquiries rather than purchases. The analysis is, above all, contextual.

In Therma–Scan , for example, the court found that six email inquiries implying that plaintiff manufactured the defendant's products provided only weak support for the conclusion that relevant consumers were likely to confuse the two. 295 F.3d at 635–36. The court considered the number of emails against the scale of defendant's operations. Id. (noting defendants sold 3,200,000 products and received 11,000 calls per month around the time of the emails). Further, it found that the evidence implied carelessness rather than confusion. Id. at 636 (noting that a mistaken internet search could easily yield the wrong email address).

Kibler offers evidence of at most ten instances of actual confusion. These include tweets and webpages advertising a performance by "DJ Logic," but meaning Hall; an email offering to book "DJ Logic," but meaning Hall; and inquiries about whether Kibler would be performing somewhere advertising "logic" and referring to Hall.

The district court concluded that the evidence of actual confusion favors Kibler only slightly. The court suggested that the ten instances paled in comparison to Hall's 170,000 album sales and popularity on YouTube, Facebook, and Twitter. The court also indicated that computer rather than human error caused the confusion on the webpages. Kibler argues that the court neglected to consider the evidence in the light most favorable to him. The Hall defendants claim that the district court erred in finding that this factor favors Kibler at all. They add that the record shows no mistaken purchases.

Because past confusion is the best proof of future confusion, any evidence at all favors the plaintiff. Kibler has offered some proof, but it is scant. If "LOGIC" really threatened to confuse consumers about the distinctions between Hall and Kibler, one would see much more than ten incidents throughout 170,000 album sales, 1.7 million album downloads, and 58 million YouTube views. The fact that none of the incidents were purchases would further prevent a jury from finding that this factor significantly helps Kibler. Homeowners , 931 F.2d at 1110.

In sum, Kibler has not presented the quantity or type of proof that would tilt the actual confusion factor substantially in his favor.

5. Marketing Channels

The marketing channels factor requires us to compare both how the parties market their products and their main customers. Homeowners , 931 F.2d at 1110. The more channels and buyers overlap, the greater the likelihood that relevant consumers will confuse the sources of the parties' products. The reverse is true too. In Homeowners , for instance, the court found little overlap where one party marketed to real estate brokers through telemarketing, brochures, and conventions, and the other marketed to real estate owners through newspaper and direct ads. Id. at 1111. That is, the methods used and consumers targeted lessened the chance that a buyer would encounter both products, let alone confuse their sources. See id. at 1110–11.

Today, most parties advertise online. Network Automation, Inc. v. Advanced Sys. Concepts, Inc. , 638 F.3d 1137, 1151 (9th Cir. 2011) (finding shared use alone of a ubiquitous marketing channel like the internet does not clarify the likelihood of confusion). We consider the following in deciding whether certain online marketing could support a finding of likely confusion. First, do the parties use the internet as a substantial marketing channel? Therma–Scan , 295 F.3d at 637 (citing Entrepreneur Media, Inc. v. Smith , 279 F.3d 1135, 1151 (9th Cir. 2002) ). Second, are the parties' marks used with web-based products? Id. See, e.g. , Brookfield Commc'n., Inc. v. W. Coast Entm't. Corp. , 174 F.3d 1036, 1042, 1057 (9th Cir. 1999) (identifying "MovieBuff," which denoted software, and "moviebuff.com," which denoted a website, as marks used with web-based products). Third, do the parties' marketing channels overlap in any other way? Therma–Scan, 295 F.3d at 637. Only one case has conducted this analysis and there it was straightforward: the plaintiff did not produce web-based products or market them online. Id.

Here, the district court found the marketing channels factor favors neither party because both Kibler and Hall failed to produce any evidence on it. The court rejected Kibler's promotion on social media and proof of online sales because it "does not support an affirmative answer to any of [the] three questions." Kibler , 2015 WL 6865928, at *4.

Kibler maintains that he has offered proof that would allow a reasonable jury to find the factor favorable to him. This includes deposition testimony 1) that he advertises on a personal website, MySpace, Twitter, and Facebook; 2) that he sells his music on Amazon and iTunes; and 3) that the parties have played fifteen of the same venues. It also includes tweets promoting Hall's album and performances and screenshots of Hall's Facebook page. Kibler counts his press clippings, use of a booking agent, and sheer length of his career as supporting evidence too.

Defendants reinforce the district court's findings, highlighting Kibler's deposition testimony that thousands of artists have played two of the fifteen venues and noting that Hall has never appeared in Downbeat or Gig . WME urges us to discount the parties' online advertising, reasoning that such a pervasive channel as the internet cannot clarify the likelihood of confusion.

The district court correctly concluded that this factor is neutral, but underestimated the impact of widespread internet use on the Therma–Scan framework. Kibler has shown that the parties market their products on the same websites, Twitter and Facebook, and target the same customers, users of Amazon or iTunes. At first glance, this overlap is compelling. But we must assess the likelihood of confusion in the real-life circumstances of the market. Under that rubric, most musical artists use those websites to advertise and sell their products today. As a result, most plaintiffs belonging to that group will meet parts one and two of the Therma–Scan test. At the same time, the popularity of these channels makes it that much less likely that consumers will confuse the sources of the parties' products. There are just too many other contenders. For these reasons, shared use of the above websites does not help us determine the likelihood of confusion.

Though evidence of the parties' common venues comes closest, it would not permit a reasonable jury to find that Kibler's and Hall's customers substantially overlap. Kibler himself admitted that thousands of artists have played two of the fifteen venues. The more artists there are, the fewer the chances of any one attendee encountering both Kibler's and Hall's songs, let alone confusing their sources. Proof of the remaining venues carries minimal weight without information about their traditional line-ups or patrons, for example.

Kibler's press clippings, moreover, are not probative at all. He has not shown that any of the same publications have featured Hall. Further, shared use of the press cannot support a finding of significantly overlapping marketing channels. The vast majority of artists seek publicity and the medium itself has infinite variations, with everything from amateur zines to well-established newspapers. Similarly, mere use of a booking agent and the length of Kibler's career do not tell us anything about how he has advertised or whom he has targeted.

The marketing channels factor is neutral because there is minimal evidence that the parties' advertising methods or targeted customers substantially overlap beyond shared use of congested websites like Facebook and iTunes.

6. Likely Degree of Purchaser Care

When consumers are more likely to exercise caution in purchasing items, they are less likely to confuse their origins. Champions , 78 F.3d at 1120. This happens when consumers have expertise in the items and when the items are particularly expensive. Id. ; see, e.g. , Homeowners , 931 F.2d at 1111 (finding factor weighs against likelihood of confusion where plaintiff's customers, real estate brokers, are savvy commercial buyers, and defendant's customers, people seeking to sell their home, are engaging in one of the most consequential transactions of their lives).

In this case, the district court found this factor unhelpful because the degree of care exercised by music consumers varies greatly by consumer and transaction. The court compared buying a song on iTunes to purchasing an expensive concert ticket, and a "turntabling aficionado" to a "casual fan of rap." Kibler , 2015 WL 6865928, at *4.

Kibler does not address this factor and the Hall defendants agree with the district court. Def Jam and WME, on the other hand, argue that the factor favors the defendants. They reason that "fans of each artist know their music" and tend to exercise substantial care in buying recordings. Def Jam Br. 27 (emphasis added); WME Br. 2 n.1 (incorporating the other defendants' arguments).

The district court was right to disregard this factor. Def Jam and WME artificially narrow the pool of consumers of Hall's music. These consumers range from people seeking a variety of recordings for use in their cars to fans following Hall on tour. Thus, the district court's analysis was sound and the factor is insignificant here. See Homeowners , 931 F.2d at 1107 ("not all of the[ ] factors may be particularly helpful in any given case").

7. Intent in Selecting the Mark

This court may infer a likelihood of confusion from evidence that defendant chose its mark to confuse consumers about the source of the parties' products. Therma–Scan , 295 F.3d at 638. The standard assumes that defendant itself believed that using the mark would divert business from plaintiff. Daddy's , 109 F.3d at 286. Circumstantial evidence of intent is sufficient when direct evidence is unavailable (as it often is). Therma–Scan , 295 F.3d at 638–39. And evidence that defendant knew of plaintiff's trademark while using its mark constitutes such circumstantial evidence. Daddy's , 109 F.3d at 286–87. In Champions , the court treated testimony that defendant learned of plaintiff's trademark before using it as slight evidence that could support a finding of intent at trial. 78 F.3d at 1121. The testimony trumped the defendant's identification of independent reasons for choosing the mark, including the "championship" caliber of a local basketball team and horses. Id. Conversely, a lack of intent has no effect on the determination of likelihood of confusion.

Having found no evidence of intent, the district court concluded that the factor is neutral in this case. See Daddy's , 109 F.3d at 287. On appeal, Kibler asserts that two pieces of evidence create a triable issue here. One is his sworn declaration that a Google or YouTube search for "logic music" or "logic musician" yielded "DJ LOGIC" and Kibler's picture or music before Hall adopted "LOGIC." The other is Hall's deposition testimony that he ran Google, Facebook, and Twitter searches for "any other rappers" using "LOGIC" before adopting it. (Hall Dep., R. 92–2, PageID 2878.) Hall testified that he ran the search "[t]o see if [any rapper] with this name was already at a level where it wouldn't make sense for two people to coexist with the same name." (Hall Dep., R. 92–2, PageID 2879.) Def Jam and WME argue that Kibler must show that defendants intended to "usurp [his] goodwill." Def Jam Br. 22; WME Br. 2 n.1 (incorporating the other defendants' arguments).

The district court properly found the factor neutral because the record prevents a reasonable jury from inferring intent. As an initial matter, Def Jam and WME cite the wrong legal standard. Evidence that defendants knew of "DJ LOGIC" while using "LOGIC" would be sufficient circumstantial proof of intent. Daddy's , 109 F.3d at 286. Def Jam and WME's reliance on a non-binding and distinguishable case is puzzling given the clear and applicable law. See Chrysler Grp. LLC v. Moda Grp. LLC , 796 F.Supp.2d 866, 871 (E.D. Mich. 2011) ("In this case, Plaintiff's mark has not been deemed protectable. Therefore, it would be unreasonable to infer intent here.").Here, we have no proof that Hall searched for "logic music" or "logic musician," no reason to believe he had to, and thus no evidence he knew of "DJ LOGIC" before adopting "LOGIC." Hall's testimony shows, to the contrary, that he avoided choosing a mark that might lead consumers to confuse his product with that of another musician. The factor is therefore neutral.

8. Likelihood of Expansion

A strong possibility that either party will expand its business to compete with the other's increases the likelihood of consumers confusing the sources of the parties' products. Daddy's , 109 F.3d at 287–88 (finding evidence of preliminary negotiations by plaintiff to buy stores in state where defendant operates could support a finding of likelihood of confusion). As with intent, a finding that neither party will expand its business is irrelevant in determining the likelihood of confusion. Champions , 78 F.3d at 1122.

The district court concluded that this factor is neutral after finding it "unlikely that the parties will expand their markets to put them in competition." Kibler , 2015 WL 6865928, at *4. Kibler identifies book excerpts, press clippings, and deposition testimony describing his experimentation with different musical genres as proof he will expand his reach. He adds there is "no evidence that [Hall] will not continue to expand his musical reach as well." Kibler Opening Br. 26 (emphasis added). Kibler stresses that the parties' mutual use of hip-hop predisposes them to expansion.

The district court rightly concluded that this factor is neutral. But the basis on which it inferred that expansion was affirmatively unlikely is unclear. All we can conclude is that Kibler offered no proof that the parties will expand their businesses. Kibler's supposed evidence says nothing of the potential for competition with Hall, whether Kibler anticipates rapping or working closely with a rapper, for example. Further, Kibler inverts the burden of proof under the factor, which requires plaintiff to present evidence of expansion, not the other way around. With no sign of any future overlap in the market, the parties' mutual use of hip-hop is irrelevant. Thus, the factor is neutral.

9. Balance of Factors

Def Jam and WME claim that Kibler has waived "any challenge" to "the district court's ultimate balancing of the Frisch factors." Def Jam Br. 16 n.1; WME Br. 2 n.1 (incorporating the other defendants' arguments). We disagree. As part of de novo review, we have a duty to consider and weigh the relevant facts in light of the Frisch factors, which Kibler has amply addressed. CFE Racing , 793 F.3d at 592. Surely, Def Jam and WME would not have us consider all of their evidence and contentions only to cede the ultimate determination to the district court.

We note then that evidence of actual confusion favors Kibler only marginally and both the strength of plaintiff's mark and similarity of the marks favor defendants. Though the Frisch inquiry is flexible and contextual, these are the "most important factors." Maker's Mark , 679 F.3d at 424. Further, the remaining factors are either neutral or insignificant here. Because no reasonable jury could find a likelihood of confusion based solely on a few instances of actual confusion, defendants are entitled to judgment as a matter of law on Kibler's federal trademark infringement and related state law claims.

C. Trademark Dilution

Kibler also alleges trademark dilution in violation of the Lanham Act, 15 U.S.C. § 1125(c). The Act entitles "the owner of a famous mark that is distinctive" to an injunction against someone who "commences use of a mark ... in commerce that is likely to cause dilution ... of the famous mark" "any time after the owner's mark has become famous." § 1125(c)(1).

The Act specifies that a mark is famous when it is "widely recognized by the general consuming public of the United States as a designation of source of the goods or services of the mark's owner." § 1125(c)(2)(A). In evaluating whether a mark is sufficiently recognized, courts may consider the duration, extent, and reach of advertising and publicity around the mark; amount, volume, and extent of product sales; and actual recognition of the mark. Id.

Courts have interpreted the Act to require the mark to be a "household name." Coach Servs., Inc. v. Triumph Learning LLC , 668 F.3d 1356, 1373 (Fed. Cir. 2012). That is, "when the general public encounters the mark in almost any context, it associates the term, at least initially, with the mark's owner." Id. (internal quotation marks omitted). See, e.g. , Audi AG v. D'Amato , 469 F.3d 534, 547 (6th Cir. 2006) (finding "AUDI" marks famous under Lanham Act because Audi had spent millions of dollars on them and they are known globally); see also Starbucks Corp. v. Wolfe's Borough Coffee, Inc. , 588 F.3d 97, 105 (2d Cir. 2009) (noting that parties agree that "Starbucks" marks are famous under the Lanham Act); Louis Vuitton Malletier S.A. v. Haute Diggity Dog, LLC , 507 F.3d 252, 257, 265 (4th Cir. 2007) (noting that parties agree that "LOUIS VUITTON" marks are famous under the Lanham Act). It is difficult to establish fame under the Act sufficient to show trademark dilution. Coach , 668 F.3d at 1373.

The district court concluded that summary judgment was appropriate because no reasonable jury could find "DJ LOGIC" is famous under the Lanham Act. The court cited its finding that Kibler failed to show the mark is commercially strong for trademark infringement purposes. Indeed, it is easier to show public recognition under Frisch than it is under the Lanham Act. Id. ("While fame for dilution is an either/or proposition ... fame for likelihood of confusion is a matter of degree") (internal quotation marks omitted).

Kibler contends that the district court erred in discounting proof of his fame. Kibler cites his sworn declaration describing his experience in the music industry and his deposition testimony that he was a guest contributor on a Grammy-winning album.

Kibler's evidence clearly falls short of the high threshold for fame under the Lanham Act. "DJ LOGIC" is simply in a different league from the marks that have met this threshold. Indeed, having failed to show that his mark is commercially strong for even trademark infringement purposes, Kibler cannot point to a triable issue here. Thus, we do not address Kibler's remaining arguments on his trade dilution claim.

III. CONCLUSION

Kibler has not provided evidence that would allow a reasonable jury to find relevant consumers are likely to confuse the sources of his and Hall's products, or that Hall's mark has diluted his. For these reasons, we affirm the grant of summary judgment to defendants on Kibler's federal trademark infringement, related state law, and federal trademark dilution claims.