Bebe Stores, Inc. v. MAY DEPT. STORES INTERN., 230 F. Supp. 2d 980 (E.D. Mo. 2002)

*981 Deirdre C. Gallagher, Andrew B. Mayfield, John H. Quinn, III, Jeffrey H. Kass, Armstrong Teasdale, LLP, St. Louis, MO, for Plaintiff.

Michael A. Clithero, Michael A. Kahn, Alan S. Nemes, Michael R. Annis, Geoffrey G. Gerber, Eric G. Enlow, Blackwell Sanders Peper Martin LLP, St. Louis, MO, Ronald J. Dolan, May Dept. Stores Co., St. Louis, MO, for Defendant.

PERRY, District Judge.

This motion for preliminary injunction pits two very different types of retailers against one another, and demonstrates the increasing competitiveness in the women's fashion industry. On the one side is May Department Stores, a well-established owner of traditional department stores, which has recently recognized that it is *982 failing to meet the challenge of providing fashionable clothes for younger women. On the other side is bebe stores, inc., an upscale fashion chain which began with one store in San Francisco in 1971 and has now grown to over 170 stores, and which has been extremely successful in providing fashionable clothes for younger women.

May's strategy to increase its market share and competitiveness in women's wear involved, among other things, introduction of a new line of clothes targeted to the 19 to 30 age rangethe same demographic targeted by bebe. After much study, May decided to call this new line of clothes "be." It began marketing its new line in ways May had never marketed before, including ads in fashion magazines and on bus kiosksthe very ways bebe has marketed itself for the last twenty-six years. May designed a logo for "be" that is very similar to the well-established and trademarked "bebe" logo.

May introduced its new clothing line into its department storesmany in the same malls where bebe has storesand launched its ad campaign in August of this year. After seeing the ad campaign, and after observing numerous instances of customer confusion, bebe sued for trademark infringement, unfair competition, and trademark dilution, and sought a preliminary injunction. The parties conducted extensive discovery on an expedited basis, and I held a two-day hearing within two months of the suit being filed.

May claims disbelief and surprise that anyone would confuse its staid image with that of the upscale bebe. May says bebe's clothes are "slutty" and would offend May customers, and that this proves there can be no customer confusion, and certainly no intentional confusion. May says the bebe trademark is weak and generic, because it means baby, which May says is a term meaning sexy young woman. May claims that bebe was not even on its "radar screen" when it developed this campaign. May argues that bebe waited too long and should have sued in June, when May's new line was announced in the business press, rather than in August when it hit the stores and the ads appeared. Finally, May says that requiring it to change its name or logo now would cost it millions of dollars, and that would be unfair.

None of May's defenses withstands scrutiny. Bebe has shown that it will suffer irreparable harm if May is not enjoined from continuing to infringe bebe's trademarks. Bebe has shown that its hard work and significant investment has resulted in a company that has a strong mark deserving of trademark-law protection, and that May's mark is confusingly similar to its. Bebe has shown: that it and May are targeting the same demographic population; that the "be" clothing line has many similarities to bebe and that many pieces would appeal to the same shoppers; that "be" is being marketed in a similar manner to bebe and is being sold in similar locations; that consumers have actually been confused and that confusion is likely to continue; that May knew about bebe and its mark before it launched this line, and could have changed the name before spending its millions; and that May's advertising strategy is likely to lead customers to think that bebe is now selling its clothes in May's department stores under the "be" name and logo. Accordingly, I will grant the preliminary injunction.

Defendant May Company owns and operates several hundred department stores throughout the country under the store names Famous Barr, Filene's, Foley's, Hecht's, Kaufman's, Robinson-May, Lord & Taylor, The May Company, and The Jones Store, among others.

Bebe stores, inc. is based in California and owns and now operates 171 stores *983 nationwide. Bebe is the owner of the federally registered trademark "bebe" and several other related marks. Manny Mashouf, President of bebe, began with one store in San Francisco in 1971. He expanded slowly at first, and in 1976 incorporated and changed the name of the four stores he then owned to "bebe." Bebe has licensing agreements for shoes and swim-wear, and those items are sold in some high-end department stores such as Nordstrom's. Except for the shoes and swim-wear items, all other bebe apparel and accessories are sold exclusively through bebe stores and on bebe's internet site, "www.bebe.com." Although at one point some bebe stores sold menswear items, the only menswear items bebe now sells are T-shirts and baseball caps with the bebe logo, and these are usually sold only in the summer.

Bebe's clothes are targeted to women in the 18 to 30 age range. The bebe witnesses described the clothes variously as: contemporary, cutting-edge, fashion-forward, body-conscious, feminine, sensuous, sexy, and versatile. The May witnesses described bebe's clothing variously as contemporary, cutting-edge, fashion-forward, body-conscious, sexy, and slutty. The clothing ranges from casual wear to business attire to evening wear. Neda Mashouf, co-owner of bebe and the wife of Manny Mashouf, testified that bebe's first major success was with wool gabardine suits, which were championed by celebrities such as TV star Heather Locklear. Bebe's garments have appeared in television shows (and were recently worn and mentioned by a character on the hit show The Sopranos), and various celebrities have worn them and spoken about them. Bebe was also a sponsor of this year's Miss America pageant.

Since its inception, bebe has spent millions of dollars on advertising. It advertises mainly through visual media, including extensive advertising in fashion magazines such as Vogue, Elle, Glamour, Marie Claire, Lucky's, and In Style. It also does "product placement" with these magazines, whereby bebe provides currently-available products to the magazines and then the magazines either write about the clothes or feature them as monthly "picks" or "favorites." From the evidence presented, it appears that one or more bebe item appears in such a feature, in one fashion magazine or another, almost every month. Bebe has also advertised extensively on mass transit kiosks over the years. In recent years, bebe has spent between $8 and $15 million annually on advertising.

Bebe's logo is sometimes written with black ink on a white background, but on store signs, labels, and hang-tags it is printed white on black, as follows:

In addition to store signs, labels, and hang-tags, bebe features the logo on its web site, in all of its advertising, and on various "logo" items of clothing such as T-shirts. Bebe sells over 100,000 rhinestone logo T-shirts per month, mostly to "aspirational" customersteenaged girls younger than the target 18- to 30-year-olds.

In spring of 2001, May Company completed a study of its market share which concluded that it did not have enough market penetration in the category of 19- to 44-year-old females. As a result, May Company decided to launch three new lines of proprietary labeled clothing. Proprietary labels are clothing lines created *984 and sold exclusively in the department stores. May's proprietary brands for women's sportswear before introduction of these new lines were Karen Scott and Valerie Stevens, which both appeal to the over 44 age group. Historically May has promoted its store labels only through instore promotions and local advertising. May had never before advertised one of its store brands in national fashion magazines.

One of the three new lines was to be targeted to the 19- to 30-year-old customer, with the ideal target customer being 25 years old. This line was given code-name "blink" initially. Beginning sometime in the summer of 2001, May Company began the process of creating, developing, and arranging for production of this new line of clothing. A strategic plan was developed, which included a list of designer and store "inspirations and influences." May researched the competition, sending teams of people to other stores selling successfully to the target market to report on styles, fabrics, sizing, amount of stock, and proportions of different types of clothes. Teams also purchased clothes for research purposes. The competing stores that May studied intensively did not include bebe, nor did the strategic plan list of "inspirations and influences" include bebe, although at least two bebe garments were purchased at some point and used for design ideas.

As the clothing line was being designed and developed, teams of people were also trying to come up with a name. Numerous names were suggested and considered, but the name selection process did not move very quickly. One of the witnesses testified that the clothing had been designed and was beginning to be manufactured before the name was even selected. Most of the creative people liked the name "fyi," but May Chief Executive Officer Gene Kahn rejected that name because it stood for "for your information" in most people's minds and because he thought it was too similar to "INC," which was the Federated Department Stores proprietary brand targeted to the same customer group and sold in Macy's and other Federated stores.

After the search for a name had been going on for several months, CEO Kahn asked William Edkins, May's Senior Vice President of Strategy and New Business Development, to get involved in the name search to "bring clarity" to the process. Edkins, in turn, brought in a third-party consultant, Addison Whitney, to help with the name selection. Over the next several months, Addison Whitney came up with many proposed names, including "Bold Expressions," which could be shortened to the acronym "BE." Addison Whitney did tests of several of the proposed names.

The testing was done in May stores in Boston, St. Louis, and southern California. Participants were given a $25 gift certificate for answering questions. Part of the test involved putting cards with proposed names on a table, providing a list of attributes, and asking the participants, "Regardless of whether you have heard of them or not, just going by name, pull out for me the ones that say [the particular attribute being tested] to you." The name "be" ranked first among 25- to 35-year-old women for "youthful attitude." This was a major reason why it was selected.

The markets in which May ran this test all had a strong bebe presence, although there was not necessarily a bebe store in the same mall where the test was performed. Bebe argues that the test was flawed, and that it may well be that the reason young women shoppers related "be" to "youthful attitude" was that they already had heard of or even shopped at bebe stores and related those stores to "youthful attitude." There is no way to tell for sure if this happened, of course, but it certainly sounds like a realistic possibility. *985 There were no controls on the study that would reveal or correct for this kind of potential test bias.

On November 5, 2001, May's legal department ordered a trademark search for the name "be" through a company called Coresearch. Addison Whitney had also conducted preliminary trademark searches of names that it proposed, and informed May that "be" had over 800 hits for trademarks using the words "be" as part or all of a registered mark. More recently, a Coresearch bill shows that in addition to the earlier searches, on July 25, 2002, May requested a search for the bebe trademark, which Coresearch did for no additional charge. All of the May witnesses who were asked denied requesting the bebe search or having any knowledge of it.

Sometime in January of 2002, May decided on the name "be" with the following logo, designed in-house by the May Company's Packaging and Labeling Department:

May graphic artist Phillip Cason designed the logo that was selected. May's graphic designers testified about many differences between the "be" logo and the "bebe" logo, and I have no doubt that a graphic artist would understand and notice these distinctions, but even some of the May witnesses admitted that people other than graphic artists might consider the "be" and bebe logos to be similar.

The upper-level May executives who testified all stated that they never considered bebe at all in this process. "Not on the radar screen" was the term used by several. Most of the top executives testified that although they knew of the bebe stores, they did not consider bebe to be competitive with their stores or with the new line they were developing in any way.

May witnesses testified that the "be" clothing is much more conservative ("tailored" or "quiet") than the clothes sold at bebe. The other evidence presented did not support this distinction. It appears that bebe sells a higher mix of casual clothes, and that in general there are more "sexy" or "body-conscious" clothes in the bebe stores, but the evidence presented to me did not show a significant difference in "trashiness" or "sluttiness" of the clothes. The clothing shown in court appeared very similar in style. Both companies sell tailored skirts, pants, and jackets appropriate for wearing to court (bebe's were worn to court by the bebe employees). Both sell regular blouses as well as see-through black lace blouses; both sell midriff-baring tops and leather miniskirts. The clothing shown in court and in the photographs introduced into evidence would appeal to the same shopper, although I recognize that fashion professionals might see a difference in style and quality. The average non-professional consumer, however, is not likely to see any significant difference. The parties agreed that bebe is priced slightly higher than "be," although "be" is considered in the "better" price category at May (as opposed to the lower price category referred to as "moderate").

Despite the testimony of May's senior executives, there is evidence that bebe was *986 on the radar screen of at least some people at May, well before the roll-out and production of the line. Documents produced by May in discovery show that on December 21, 2001, someone at May who was working on the "be" line printed out a page from the bebe website. No one from May remembers doing this, and all who were asked are sure it was not them who searched the bebe website, although the printout is attached to a list of "be" slogans. Roger Stolle, May's former Director of Packaging and Labeling, who was extensively involved in the name and logo selection process, testified that after the label and hang tag had been designed, someone asked him, "Did you look at bebe?" He could not remember who asked him, but he simply said "no" and did no further checking. Stephanie Konczarek, May's Manager of Graphic Artists, who was also involved in the name selection and logo production, testified that in December of 2001, Mike Love, an assistant to one of the executives, mentioned to her that the logo was similar to bebe's in a short "in passing" conversation. She took no action, stating, "I just blew it off because I didn't think it was relevant at all." Sometime in January or February, 2002, May Company's in-house trademark lawyer, John Manos, was approached by a May employee who mentioned a potential problem between the "be" mark and the bebe mark. Manos testified that he had discussions about the bebe trademark with Senior VP Edkins and others, although May asserted the privilege for the content of this meeting. Edkins denies having had any such conversations, and denies ever knowing of any problem or issue regarding the "be" name and the bebe trademark.^[1] Additionally, as discussed above, Coresearch indicated on a bill that it was asked to do a trademark search on bebe in Julya month before the lawsuit was filedbut no one at May admits to any knowledge of this.

On June 13, 2002, May held a fashion show for the "be" line and for its other two new proprietary lines, "i.e." and "i.e. relaxed." May invited both the fashion and the business press to the fashion show, and also invited various financial analysts. May had never before held a fashion show for any of its store brands. Before the show May had a meeting for the high-level May employees who would be attending, and who would attend the party after the show. The purpose of the meeting was to provide information to the May employees to help them discuss the new lines with the press and with the financial analysts. Among the materials passed out at the meeting was a memorandum entitled "Possible Questions from Analysts or Reporters." The last question and answer on this sheet was as follows:

Q. Do you have any concern that your name be will be confused with bebe?

A. There is quite a difference. First of all, bebe is french for "baby." Secondly, we are using the word in its most American "verb" formyou'll see our advertising using it with descriptor lines like be original, let it be, be yourself, be bold, be noticed, be imaginative, be a star, and be dazzling.

As with the trademark searches and conversations about bebe before the launch, none of the executives can remember seeing this document. They all admit that it was passed out, but deny ever reading it, and continue to insist that bebe was "not on the radar screen" at all. The only May employee who acknowledges ever seeing *987 the document is the one who wrote it, Sharon Bateman, May's Vice President of Corporate Communications. She testified that she created it and made sure it was in the briefing packages passed out at the meeting. Bateman wrote the first draft andbefore the meetingcirculated it to her boss, Jan Kniffen, May's Senior Vice President and Treasurer, and to other "members of May's senior management." She testified that she included the question about bebe because, "In my mind I could see that someone might think there would be a similarity," between the "be" and bebe marks, and because, "The look appeared similar to me at first glance."

Following the fashion show, there were some articles written about May's new line, including one in Women's Wear Daily, although May chose not to introduce any of those articles into evidence at the hearing. The WWD article, however, was attached to May's memorandum in response to the motion for preliminary injunction as exhibit T, and it bears the headline "May Co.'s Formula: Build In-House Brands to Return to Growth." The "be" logo does not appear in that article, and the "be" name does not appear anywhere on the first page of the article, but can be found upon reading to the sixth paragraph of text.

At the end of July, the August issues of fashion magazines hit the stands and included advertisements for the "be" line. The May ads prominently featured the "be" logo. May ran the ads nationwide in Marie Claire, Glamour, Elle, Vogue, Cosmopolitan, and In Style. May had never run ads for its proprietary brands in these magazines before. During the month of August, May began running ads on bill-boards, on bus kiosks, and on large posters in malls, again featuring the "be" logo. In September it began running television ads for the "be" line.

May began selling the clothes in late July or early August. Its store displays feature large black signs with the "be" logo, which are very similar to the large black signs used at bebe stores. In some stores, May has placed two signs over two clothing displays next to one another, so that the signs literally read "be be."

Bebe first learned of the "be" line in early August, when its employees began seeing the ads in magazines and in malls. It filed this suit on August 16, 2002.

Bebe presented both live and affidavit testimony going to the issue of customer confusion. Customer Holly Shoemake, from Murray, Kentucky, testified that during a shopping trip to St. Louis in September she purchased "be" clothes from Famous Barr and a bebe item from a bebe store. She testified that she believed they were all bebe clothes when she bought them. She had been familiar with bebe clothes from magazines and from seeing them on other people, but had never purchased bebe before since there are no bebe stores in her area. She testified that when she showed her purchases to her husband, he argued with her assertion that they were the same make of clothes, and so she called the bebe store to ask.

Bebe presented live testimony of eight bebe employees from around the country. Each testified to several instances of customers coming into a bebe store and asking to purchase items they had seen on a "be" ad. One witness testified that in the mall where her bebe store is located the May department store has two large mall poster boards advertising "be" near the bebe store, and that a customer insisted bebe must have an item because it was on the ad right outside the store. This witness described walking with the customer out of the bebe store to the May ad where she was able to point out that it was not a bebe ad. Others testified to people bringing *988 "be" merchandise into bebe stores and attempting to return it there.^[2] Several employees testified that when they first saw the "be" ads they inquired of their supervisors whether bebe was now selling a line of clothes at the May stores. Several testified that customers or others had commented that they had noticed that bebe was now selling at May stores. Two witnesses described conversations where employees of May stores told them that May was now selling bebe clothes.^[3] Another testified that a customer asked why the bebe prices were lower at the May store.

Bebe also introduced thirty-two affidavits from other bebe employees and three affidavits from other customers.^[4] The three customers testified that they believed the "be" ads they saw were for bebe clothes. The bebe employees testified to numerous other incidents of customers coming into stores asking for items they had seen in "be" ads, attempting to return "be" merchandise at bebe stores, and asking questions or making comments about bebe's selling its clothes in department stores. One employee described an encounter with an irate customer who did not believe that the item she wanted was not offered by bebe, even after the bebe employee explained that it was not in a bebe ad.^[5]

The parties disagree about how to pronounce bebe's name and what it means. Bebe's witnesses pronounce it with long e's: "bee-bee." May's witnesses pronounce it with long a's: "bay-bay." Bebe's founder, Manny Mashouf, testified that in picking the name he was looking for a name with "no specific meanings, or any relationship to fashion" and he thought of the Shakespearean expression "to be or not to be." He also testified that bebe (with the long e pronunciation) is the Persian word for the Queen of Hearts in a deck of cards, and that growing up in Iran he and boys he knew used bebe as slang for woman. At various times in the past bebe has stated that bebe meant woman in either Turkish or Persian, but it does not. Mashouf testified that an employee of *989 Turkish descent had once told him that bebe meant woman in Turkish, and he believed that, but now knows it not to be true. With different accents, bebe means baby in French, Spanish, Portugese, and Italian, and in most of those languages it is pronounced with the long a pronunciation, although placement of the accents changes which syllable is accented. May has argued, but did not present any evidence, that "baby" in English is slang for a sexy young woman. The court can take judicial notice that "hey baby" may be an expression sometimes used by males when a sexy young woman walks by, but also that "baby" is used as a term of endearment or affection for children as well as for grown men and women, whether sexy or not.

Bebe's witnesses testified that they make efforts to get people to pronounce their name properly, with the long e pronunciation. Mashouf described a humorous ad campaign from the 1980s which played on the french use of the word and whether the clothes in question were "bay-bay" or "bee-bee." Bebe presented evidence that its store associates are trained to "gently" correct customers when they mispronounce the name. Additionally, a magazine ad campaign from the mid-90s used photographs where the model's head separated the two parts of the word bebe, so that it would be natural to read it as "bee-bee." Finally, the bebe web site, in the FAQ (Frequently Asked Questions) section, contains the following:

Is it pronounced "bee-bee," "bay-bay" or "beh-beh"? What does "bebe" mean?

Although most people tend to pronounce it "bay-bay," it is really pronounced "bee-bee." bebe was founded as a San Francisco boutique in 1976 by Manny Mashouf, the company's chairman, CEO and president. The name bebe was selected because it represented the philosophy of the time, "to be or not to be." The name also has various multilingual meanings, including the Persian name for the queen in a deck of cards and the Turkish word for woman.

Most of the May witnesses testified that they knew of bebe from their years in retailing, and that they always heard it pronounced "bay-bay." There were two exceptions to this: Stephanie Konczarek, who had shopped at bebe but never purchased anything there, and Sharon Bateman, whose goddaughter had purchased clothes at bebe. Both of those witnesses both younger womenadmitted that they had heard it pronounced "bee-bee." May "senior management" witnesses were very insistent that they had never head the store pronounced any way but "bay-bay," and used that pronunciation throughout case. Aside from being high up in the May Company, these witnesses had another thing in common: they have spent their entire careers working for large retail companies such as May. None of them has worked as a retail entrepreneur or for a "specialty" retailer such as bebe, and none of them has worked for any stores that consider bebe a direct competitor. The only conclusion to be reached from all this evidence is: (1) many people who have seen bebe stores or ads may not knowor carewhat the correct pronunciation is, (2) customers who have shopped at bebe and members of the target market of young, fashion-conscious women know the name is pronounced "bee-bee," and (3) fashion executives who work for large department store companies incorrectly believe the name is "bay-bay," and never had any reason to find out differently.

May presented evidence that other vendors use the letter "B" (jewelry and logos, for example, Bellagio Casino in Las Vegas uses a stylized "B" on its T-shirts, and another manufacturer has items from the "B-Wild" collection), the word "Be" (a pair of pants from "Be Bop" has buttons with a cursive "Be," and a Bum Equipment football *990 type shirt has the logo "86-BE"), or the word "bebe" with various accents (baby items ranging from changing pads to hair ornaments to baby clothing and toys).

May's witness Edkins presented evidence regarding expenditures May had made on the "be" line and what it would cost if May were ordered to change the name or logo of the line. Edkins had great difficulty explaining how the figures to which he testified were obtained or what they meant, and at one point explained that he was not the proper witness to ask, but that a witness who had already testified would be better. May did not ask to recall that witness, so I am left with Edkins' testimony and his exhibit to try to understand the facts.^[6] I believe I can ascertain the following from this evidence: By the time this suit was filed, May had spent approximately $12 million in design, fixtures, start-up, and advertising expense. At that time it had inventory on hand or in shipment costing about $8 million, and it expected to make a 40% gross profit on that inventory. As of October 1 2002, May had inventory on hand or in shipment valued at approximately $11 million, and expected to make the same gross profit. May could replace the labels on garments that had not yet been shipped at a cost of about $750,000. Pulling print ads for future magazines would cost about $275,000.

Since the lawsuit was filed, but before the preliminary injunction hearing, May has spent over $3 million in television advertising, all of which could have been cancelled without cost at the time the suit was filed. It would cost May about $200,000 to remove the "be" fixtures in its stores, and those fixtures had cost about $4 million to design, build, and install. That expense had all been spent before the lawsuit was filed. Edkins and the exhibit indicated that damages to May of completely shutting down the "be" line now and throwing away all the clothes already manufactured would be approximately $34 million. Had May pulled the line immediately upon learning of the lawsuit, the damages would have been about $23 million.^[7]

The standards for issuing a preliminary injunction are well established. I *991 must consider the following factors: (1) the probability that the movant will succeed on the merits; (2) the threat of irreparable harm to the movant; (3) the balance between this harm and the injury that granting the injunction will inflict on other interested parties; and (4) the public interest. See United Indus. Corp. v. Clorox Co., 140 F.3d 1175, 1178-79 (8th Cir. 1998); Dataphase Sys. Inc. v. C L Sys., Inc., 640 F.2d 109, 114 (8th Cir.1981). No single factor is dispositive and all of them must be balanced. International Ass'n of Machinists and Aerospace Workers, AFLCIO v. Schimmel, 128 F.3d 689, 692 (8th Cir.1997). Bebe has the burden of showing it is entitled to the relief sought.

By enacting the Lanham Act, "Congress apparently intended to encourage competitors to seek injunctions as a method of combating false advertising, and, in such cases that ultimately prove to have merit, injunctive relief is not to be issued reluctantly." Clorox, 140 F.3d at 1179. Nonetheless, the burden of demonstrating that a preliminary injunction is warranted under the Lanham Act is heavy when granting the preliminary injunction will "give [the movant] substantially the relief it would obtain after a trial on the merits." Sanborn Mfg. Co., Inc. v. Campbell Hausfeld, 997 F.2d 484, 486 (8th Cir.1993).

To prevail on its claim under the Lanham Act, Bebe must establish that its trademark has priority over May's "be" mark and that May's continued use of "be" is likely to cause consumer confusion. Hubbard Feeds, Inc. v. Animal Feed Supplement, Inc., 182 F.3d 598, 601 (8th Cir. 1999). The ultimate inquiry always is whether, considering all the circumstances, the allegedly infringing mark "creates a likelihood of confusion, deception, or mistake among an appreciable number of ordinary buyers." Duluth News-Tribune, A Division of Northwest Publications, Inc. v. Mesabi Publ'g Co., 84 F.3d 1093, 1096 (8th Cir.1996).

In determining whether a likelihood of confusion exists, the court considers the following factors: (1) the strength of the trademark; (2) the similarity between the parties' marks; (3) the competitive proximity of the parties' products; (4) the alleged infringer's intent to confuse; (5) evidence of actual confusion; and (6) the degree of care reasonably expected of potential customers. Hubbard, 182 F.3d at 601. None of these factors alone is dispositive. Luigino's, Inc. v. Stouffer Corp., 170 F.3d 827, 830 (8th Cir.1999); see also Anheuser-Busch, Inc. v. Balducci Publications, 28 F.3d 769, 774 (8th Cir. 1994) (noting that these factors are not a distinct test, but represent the sort of considerations which a court should analyze in determining whether a likelihood of confusion exists).

To show that it is likely to prevail under Missouri's anti-dilution statute, Mo.Rev.Stat. § 417.061, bebe's burden is not as great, because the state law provides greater protection than the Lanham Act by expressly permitting claims "notwithstanding the absence of confusion as to the source of goods or services." Anheuser-Busch, 28 F.3d at 774. To prevail on its state anti-dilution claim, plaintiff must show: (1) its mark or trademark was valid at common law; (2) that its mark is distinctive; and (3) that defendant's use of its name created likelihood of dilution of distinctive quality of plaintiff's mark. Gilbert/Robinson, Inc. v. Carrie Beverage-Missouri, Inc., 758 F. Supp. 512, 527 (E.D.Mo.1991). Missouri's anti-dilution statute requires automatic imposition of injunctive where likelihood of injury to business reputation or dilution is found. See id. at 528; Anheuser-Busch, 28 F.3d at 777 n. 5.

As a preliminary matter, a court must determine whether plaintiff's mark is strong enough to merit trademark protection. A mark can be classified into one of four categories: (1) arbitrary; (2) suggestive; (3) descriptive; or (4) generic. Duluth News-Tribune, 84 F.3d at 1096. "A strong and distinctive trademark is entitled to greater protection than a weak or commonplace one." SquirtCo v. Seven-Up Co., 628 F.2d 1086, 1091 (8th Cir.1980). "Determining that a mark is weak means that consumer confusion has been found unlikely because the mark's components are so widely used that the public can easily distinguish slight differences in the mark, even if the goods are related." General Mills, Inc. v. Kellogg Co., 824 F.2d 622, 626 (8th Cir.1987). Evidence of third party usage of similar marks on similar goods is "relevant to show that the mark is relatively weak and entitled to a narrower scope of protection." Id. at 627.

Bebe's mark is arbitrary, in that it does not describe the goods being sold or anything related to women's apparel. May has argued that it is generic because it means "woman," but the evidence shows that it does not mean woman. It also does not mean "sexy young woman" or anything inherently related to the clothing being sold.

May also argues that the bebe mark is weak because other people use the word be or B or bebe to sell their goods. It is true that the word bebe is used to sell many baby-product items, but these are not similar goods. Additionally, the instances when May showed that the letter "B" or word "be" was used to sell goods that included a "B" in the name also does not show weakness, because in those cases neither the mark nor the goods were similar.

As bebe's market presence has grown over the years, the strength of the mark has increased. The evidence shows it to be a very strong mark, especially in its target market of 18- to 30-year-old women. Bebe's large sales of logo T-shirts also indicates the strength of the mark, since teenagers wear logo clothes as a status symbol. Bebe has expended significant money and effort over the years to expand both the stores and its name recognition among its customers and among these "aspirational" consumers, especially through its sustained advertising campaigns in fashion magazines.

May argues that the mark cannot be strong because in 1989 bebe (in one of its trademark applications) told the trademark office that it was a weak mark, since it was used for numerous baby items. I do not believe that this is an admission that bars bebe's claim, because the evidence showed that bebe has developed the mark into a very strong one since making that statement. Bebe has spent millions each year in advertising, has greatly increased the number of stores, and now has a strong, protectible mark.

The use of identical, even dominant, words in common does not automatically mean that the two marks are similar. Luigino's, 170 F.3d at 830 (citing General Mills, 824 F.2d at 627). "Rather, in analyzing the similarities of sight, sound, and meaning between the two marks, a court must look to the overall impression created by the marks and not merely compare individual features." General Mills, 824 F.2d at 627. Some courts have held that the controlling pronunciation for trademark purposes is the one that consumers commonly use. See, e.g., Smithkline Beckman Corp. v. Proctor & Gamble Co., 591 F. Supp. 1229, 1239 (N.D.N.Y.1984); Lebow Bros., Inc. v. Lebole Euroconf S.p.A., 503 F. Supp. 209, 212 (E.D.Pa.1980). "Where products are closely related, less similarity *993 in trademarks is necessary to support a finding of infringement." SquirtCo, 628 F.2d at 1091.

The two logos look very similar to me, and when bebe is pronounced correctly, they sound very similar.

The evidence showed that shoppers of bebe products normally pronounce the word "bee-bee," and that bebe itself takes considerable effort to get the correct pronunciation across to its customers. The fact that May executives or even other people with similar large-company retail experience may not know the correct pronunciation does not change this. The marks are very similar both in sight and sound.

May argues that the marks are not similar because its ads use "be" as a verb, and include, in addition to the large logo, text statements like "be creative" and "be yourself." This additional use of the word "be" does not change the fact that the signs, ads, labels, and hang tags all prominently feature the logo, and the name and logo are very similar to bebe and its logo.

May is selling its "be" lines in many of the same malls where bebe has stores. In areas where they are not in the same malls, they are being sold in the same towns, and May has placed billboards close to malls where bebe stores are located. Additionally, the clothing lines are very similar and appeal to the same shoppers.^[8] While the two companies may not consider themselves to be direct competitors, they are targeting the same customers with very similar goods, in many of the same locations. Closer competitive proximity could not be shown.

Bebe has not shown that May started out intending to confuse customers. The evidence shows that May was attempting to deal with a business problem, and did not have bebe in mind initially when it began designing the clothing line and the name. I believe that the most reasonable inference to be drawn from the credible evidence presented is that when May began the process bebe was truly "not on the radar screen" of most of the people involved, but that at some point it did show up as a concern. In the initial design process, the list of stores and lines that were considered "influences and inspirations" did not include bebe, but did include many of the same designers from whom bebe gets its inspiration, and did include many of the specialty retailers that bebe considers its competitors. At some point when designers and others were going to the competing stores to study them and purchasing items for "inspiration," someone from May purchased some bebe items, and the testimony was that one stitch from one garment was incorporated into a "be" *994 design. In that creative process, bebe had to get on the radar screen.

More importantly, inferences that bebe definitely showed up on May's executive radar screen can be drawn from the May search of bebe's web site in December, from the two "in passing" questions "what about bebe"asked when the logo was designed, from the lawyer's testimony that he was asked about bebe and had meetings with upper management about the trademark issue, from the Q & A passed out at the fashion show specifically anticipating that people might think "be" and bebe were confusingly similar, and from the Coresearch bill indicating that May ordered a bebe trademark search just weeks before the "be" line was launched.

My conclusion from these facts is that while May did not set out with the intent to infringe bebe's trademarks and generate confusion at the beginning, it learned there was a problem later on and intentionally chose to ignore it. I believe that at some point Edkins, and most likely others at May, learned about bebe's trademark and was concerned that it might be confused with the "be" name, logo, and label that had been selected. These people decided to go forward anyway because it had taken so long already to get the line and its advertising strategy developed, and because CEO Kahn liked the name. May's dismissive attitude about bebe demonstrated throughout this lawsuit by numerous witnessesindicates that it simply thought bebe was too small, too "specialty," too "slutty," and certainly too unimportant for it to worry about at that late date, and so it just rolled on. Or, as one witness said, "I just blew it off."

Proof of actual confusion is relevant to an ultimate finding of likelihood of confusion. See General Mills, 824 F.2d at 628. Although evidence of actual confusion is not necessary, "it is perhaps the most effective way to prove a likelihood of confusion." Hubbard, 182 F.3d at 603. "[W]hen determining whether there exists a likelihood of confusion, weight is given to the number and extent of instances of actual confusion." Duluth News-Tribune, 84 F.3d at 1098 (quoting Life Technologies, Inc. v. Gibbco Scientific, Inc., 826 F.2d 775, 777 (8th Cir.1987)).

Bebe's evidence of actual confusion is extremely strong. Bebe presented evidence, in one form or another, from over forty witnesses showing that all over the country, in areas where May is advertising and selling "be," bebe's customers are confused. This is very strong direct evidence, unlike the speculative hearsay evidence rejected in Vitek Sys., Inc. v. Abbott Labs., 675 F.2d 190 (8th Cir.1982), and similar cases.

It is also reasonable that customers are confused. As the May witnesses testified, in today's fashion marketplace, everyone competes with everyone else, and the same customer might buy apparel from Target, Famous Barr, Nieman Marcus, and in moderate or expensive specialty stores in malls. That consumer is likely to know that there are different versions of the same designer labels or brands in different stores, or even in different areas of the same store. For example, "Polo by Ralph Lauren" is sold in May stores and at other high-end department stores, and the brand "Ralph"referred to by a May witness as "Little Ralph"is a lower-priced line made by the same company but sold in a different part of the May stores. If a reasonable consumer knows that "Jones New York" is sold in Famous Barr while the lower-priced "Jones Wear" made by the same company is sold at JC Penney's, such a consumer would not be surprised at all to find that higher-priced specialty bebe might be selling a lower-priced version in *995 department stores. As the "Polo by Ralph Lauren" and "Ralph" example shows, shortening a name to differentiate between versions of the same company's clothes is not unusual in this industry.

Bebe has shown significant actual confusion.

May presented evidence that women are discerning shoppers, and argues that the target customer can be expected to ascertain the difference in the clothes and the makers. While women may be more discerning shoppers than men when they buy clothing, this does not mean that they can be expected to understand the difference here, or that they will be bothered to find out the difference. The confusion evidence presented by bebe shows that some customers did ask, but most did not. Given the high degree of similarity in the mark, and the reasonable customer expectation that the same company might sell different versions of its clothes in different stores, this factor also weighs in bebe's favor.

Bebe has shown that all of the factors regarding likelihood of confusion save oneMay's intenttend to show confusion. And the intent factor is mixed, because May's intent at the beginning of the process is not necessarily the same as May's intent when the line was finally launched. This is a clear-cut case of trademark infringement. Bebe has shown a strong likelihood that it will prevail on the merits.

Where likelihood of success on the merits has been shown in a trademark infringement action, the threat of irreparable harm may be presumed. See, e.g., Clorox, 140 F.3d at 1183; Calvin Klein Cosmetics Corp. v. Lenox Laboratories, Inc., 815 F.2d 500, 505 (8th Cir.1987). Here, however, it is not necessary to rely on that presumption because bebe has shown that it is suffering irreparable harm, and that legal remedies of money damages will not be adequate to protect it from this harm. The customer confusion shown here cannot be corrected by damages paid after the fact, and if allowed to continue, could ultimately force bebe out of existence. May is a much larger company, with many more stores and much more square footage in malls around the country.

Although no evidence of its advertising budget or its sales or profits was presented, in two months May has already spent more promoting this one line than bebe spends annually on advertising for its entire company. Retailers spend money on advertising because it works. It is very likely that bebe will lose customers to May's department stores and never be able to get those customers back, nor will it ever be able to measure, for money damages purposes, the lost sales. The harm will be irreparable.

May argues that it would be inequitable to grant the injunctive relief sought because it will cost May somewhere in the range of $30 million. Although May did not present very good evidence to support this claim (and its briefs use different numbers), I agree that granting the injunction will be expensive, although the actual amount is probably much less than the $30 million range May suggests. Of course, May could have saved some of this loss even after the lawsuit was filed, for example, by not going forward with its planned television advertising, or by modifying its holiday orders. It chose not to do so, instead deciding, as May's lawyer told me at the initial scheduling conference, that "May is willing to take that chance."

*996 In considering the balance of the equities, I must consider both the harm to May if the injunction is granted, and the harm to bebe if the injunction is denied. Whatever I do will necessarily cause hardship to one company or the other. This is especially unfortunate where it appears that both of these are well-run companies with carefully planned strategies for reaching their consumers.

This line of clothing is important to May's financial growth and continued success, and May had good business reasons for deciding to target this market. It also appears that May used good business practices in designing and developing the clothing line. May's mistake, however, was in choosing a name and logo that were confusingly similar to an established store, which had been very successful in selling similar goods to the same target market. May went forward with this even after it learned of the problem, and it did so at its peril. By infringing on bebe's trademark, May gained an unfair advantage, and its "be" line is undoubtedly more successful than it would have been had it chosen an unknown name.

Bebe has only its stores and its name. It does not sell hundreds of other categories of items, as May does. All bebe does is sell contemporary fashions to this target market, and if the injunction is not granted, bebe stands to lose the goodwill and customer base it has grown through investment in advertising and through savvy marketing and the provision of a good quality product to its customers over the last twenty-six years. Bebe does just this one thing, and it has done it very well over the last twenty-six years, and it stands to lose it all if May is not enjoined.

The balance of harms weighs in bebe's favor. May could have stopped when it learned of the problem, picked a new name (or picked one of the thousands it had considered already and rejected), and still introduced its new line with little, if any, delay. Even had it stopped when the lawsuit was filed, or at least pulled some of its advertising, it could have reduced its damages significantly. It chose not to do any of these things, and any harm to May from granting the injunction is harm of May's own making.

As is almost always the case, the public interest factor favors neither party strongly, when considered in isolation. This is because the public interest favors healthy competition, but does not favor unfair competition. Thus, the public interest in any case is likely to favor which ever side is likely to succeed on the merits of showing unfair competition or lack thereof.

Here the public interest favors granting the injunction, because the public interest favors the protection of holders of valid trademarks against misuse by competitors. May has obtained an unfair competitive advantage by trading on the good name and reputation that bebe has built up over the years, and this is not in the public interest.

Bebe has shown that it is entitled to the preliminary injunction it seeks, and so I must determine the appropriate scope of that relief. May argues that a mandatory injunction is not proper, because injunctive relief should only preserve the status quo. This argument is somewhat circular, however, because since bebe has shown that it is likely to prevail on its claim that May is infringing its valid trademark, preserving the status quo would preserve the illegality and allow May to continue to violate bebe's rights, when bebe is suffering irreparable harm. Under May's reasoning this type of injunctive relief could never be granted.

*997 I will, however, use a phased approach, and allow May to continue to sell at least some of the clothing that it currently has on hand or in shipment. I believe May's evidence that re-labeling goods already in the stores or in shipments is impractical, and I believe that bebe's rights can adequately be protected by more limited relief. May's witnesses testified that it could with reasonable ease re-label clothes that are still at the factory, and I will require it to do so.

May's evidence about what merchandise is on hand, in process or being shipped, or on order for future delivery was very confusing, and was supported mainly by the exhibit that Edkins could not explain. That exhibit seems to show that as of October 1, 2002, May had merchandise on hand or in shipment carrying the "be" label that would be delivered as of December 1. The document indicates that merchandise scheduled to be delivered December through April could be re-labeled at a cost of less than $650,000. It also states that the first re-labeled merchandise could be in the stores by December 14, assuming the "decision date" was October 1. Another witnesses estimated that re-labeled goods could be available by late January, but May did not attempt to explain the difference in this evidence. I will select December 31, 2002 as the cut off date for selling "be" labeled goods, as it seems reasonable based on the conflicting evidence.

I will enjoin May from violating the bebe trademarks, and from using "be" as the name of its new line, with the following exceptions: May can sell the merchandise it now has in stock, whether in its stores or in stock or in transit from the factories, but only through December 31, 2002. May cannot sell any "be" labeled goods that are currently being manufacturedonly those that have already been shipped.^[9] As of January 1, 2003, no merchandise carrying a "be" label can be sold in any May store, and May cannot resell any "be" goods to discounters unless all labels are removed (i.e., if the goods are resold to discounters the labels must be completely removed, not just sliced or cut in parts).

Although May can sell the clothing it has on hand, it must remove all store signs showing the "be" name or logo, and it must cease all advertising using the "be" name and logo. All television ads must be pulled immediately, as must all print ads, billboards, mall signs, and kiosk ads. In addition, in each store where "be" clothing is sold, May must prominently post, at locations near the "be" clothing and at the nearby cash registers, a sign stating, in large letters noticeable to an average shopper: "`Be' products are designed and manufactured by May Department Stores Company, and are not affiliated with or approved by bebe stores, inc. in any way," or other similar language acceptable to bebe.

I realize that some of these things are more easily accomplished than others. May should be able to remove the mall signs immediately. May should be able to remove its in-store signs within a day or two. May indicated that it will have to pay for its television ads for the next thirty days, but it can pull them off the air immediately, whether it has to pay for the time spot or not. Similarly, May may incur penalties to cancel print ads, but it must do so, and must pull any print ads that can be pulled. If a magazine has already gone to press and the ad cannot be pulled, May will be expected to present evidence showing that this is beyond its control. Similarly, billboard and kiosk ads *998 must come down, but I realize that May may not be able to control when a billboard company or transit company removes a sign. May will be expected to provide evidence of its compliance with this provision as well, and must show that it has taken all means possible to remove the advertisements from the public view.

These requirements will become effective upon bebe's posting a bond in the amount of $3 million.

May will be required to provide a status report ten business days after bebe's posting of the bond, indicating what it has done to comply with the preliminary injunction, and to provide a status report every ten days thereafter until it completely complies with the preliminary injunction and is no longer selling any "be" labeled merchandise. The status reports must be supported by affidavits of May employees or officers with knowledge of the facts.

As indicated at the hearing, I will order this case referred to mediation immediately, and will set a scheduling conference in the near future in order to establish a schedule for the remainder of the case.

Accordingly,

IT IS HEREBY ORDERED that plaintiff's motion for preliminary injunction [# 6] is granted.

IT IS FURTHER ORDERED that plaintiff shall post a bond in the amount of $3 million before the preliminary injunction shall be effective.

IT IS FURTHER ORDERED that defendant shall file status reports, attaching any necessary affidavits, showing compliance with the preliminary injunction order, within ten business days of bebe's posting of the bond, and every ten business days thereafter, until such time as defendant has indicated that it has fully complied with the order and is no longer selling any "be" labeled merchandise in the stores.

A separate preliminary injunction in accordance with this order is entered this same date.

Separate orders referring this case to mediation and setting it for a Rule 16 conference are also entered this same date.

[1] For reasons discussed more fully in footnote 8 and in the discussion section below, I do not believe Edkins' testimony on this point.

[2] May insists that all of this evidence is inadmissible hearsay, but it is not being introduced to prove the truth of the matter asserted by the customer.

[3] Again, May objects that this is hearsay, but bebe is obviously not introducing this evidence to prove that bebe is selling its clothes at May's stores.

[4] May correctly asserts that these affidavits are hearsay, because the witnesses did not testify live at the hearing and were not subject to cross-examination. Because this is a preliminary injunction hearing, however, and was held on short notice and on an expedited schedule, I believe that receipt of these hearsay affidavits is appropriate under the residual exception contained in Fed.R.Evid. 807. The issue here is likelihood of success on the merits, and bebe's production of nine live witnesses and sworn statements of thirty-five morewho certainly can all be called live or by deposition when we ultimately have trial on the meritsis probative on the issue of whether bebe is likely to succeed on the issue of confusion. Requiring bebe to call all these witnesses live under the expedited circumstances of this motion hearing would not be reasonable. Additionally, May's cross-examination of the nine live witnesses was very limited, and did not seriously challenge the credibility of any of the witnesses. The lack of persuasive cross-examination of the live witnesses adds to the circumstantial guarantees of trustworthiness of these affidavits, and so I have considered them.

[5] Again, May asserts that this is hearsay within the already hearsay affidavit. I have described above why I believe the affidavits are admissible. The customer comments contained in the affidavits are not hearsay because they are not being introduced to prove the truth of the matter asserted by the customer, e.g., that bebe was now selling its goods in May's department stores, or that bebe was offering for sale an item advertised by May.

[6] I overruled plaintiff's foundational objection to this exhibit, but clearly the foundation was inadequate. In a normal trial this evidence would not have been allowed on this foundation. If this testimony had been presented by a "business loss" expert rather than a high-ranking executive of May it would not have passed even a cursory review under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993), but since this is a preliminary injunction matter heard on an expedited basis, I received it under the same spirit under which I received plaintiff's consumer confusion affidavits discussed in footnote 4.

[7] Edkins was a very incredible witness. His demeanor was hostile and dismissive, both in the video deposition played and in his live testimony. On direct he testified that May had studied the demographics of persons who had purchased "be" clothes and found that 85% of them were over the age of 35. But questioning by the Court showed that this testimony was false, because the exhibit itself shows that it was based only on "household level demographics" of addresses, and Edkins eventually admitted it used only some of the "be" purchases made during the month of August. The study actually showed only that some of the credit cards used to purchase some of the clothes were associated with addresses where some of the residents were over 35 years old. This evidence would never have passed a Daubert review at trial, and Edkins' reliance on it casts additional doubt on his credibility. Edkins also demonstrated his bias in his deposition by repeatedly answering questions about "be" advertising as if they had been asked about "bebe." He was so obviously programmed to "stick to his story" and make sure his testimony supported May's case that he failed to even listen to the questions.

[8] May says its target market is 19- to 30-year olds, and that the line is especially targeted to a 25-year-old shopper. Bebe says its target market is 18- to 30-year-olds. This distinction is not significant enough to make a difference in the analysis.

[9] May's compliance with this requirement should be easily verifiable, as it should have records showing what has been shipped as of the date the injunction becomes effective. I will expect it to provide proof of compliance.