Plaintiff CollegeSource, Inc. (“CollegeSource”) offers access to subscription-based and free databases providing information on college and university course curriculums, equivalencies, and transferability. CollegeSource hosts course catalogs and updates its databases annually by collecting digital course catalogs and digitizing paper course catalogs. It sells access to its database in addition to tools that facilitate student academic-credit transfers and offers a free service, CataLink, that allows schools to link directly to CollegeSource’s digitized course catalogs. Defendant AcademyOne, Inc. (“AcademyOne”) builds systems that allow faculties to evaluate academic courses for credit equivalency. To build its database of course descriptions, AcademyOne hired a developer to download course catalogs and convert them into a text database. The developer downloaded some course catalogs from schools using CollegeSource’s CataLink database and “scraped” the PDF files for text. AcademyOne allowed free access to its course-description database and hosted 4,000 courses on its website. AcademyOne purchased the terms “college source” and “career guidance foundation,” both trademarks of CollegeSource, as search-engine keywords. CollegeSource sued AcademyOne for trademark infringement and unfair competition, among other claims. AcademyOne moved for summary judgment and the district court granted the motion.


The court dismissed CollegeSource’s infringement and unfair-competition claims because CollegeSource did not provide sufficient evidence that AcademyOne’s use of CollegeSource’s trademarks was likely to cause confusion. The court noted the four-factor likelihood-of-confusion test for keyword cases adopted by the Ninth Circuit in Network Automation, Inc. v. Advanced Systems Concepts, Inc., 638 F.3d 1137, 1154 (9th Cir. 2011): (1) strength of the mark, (2) evidence of actual confusion, (3) types of goods and degree of care likely to be exercised by the typical purchaser, and (4) the labeling and appearance of the advertisements triggered by the keywords. In applying the traditional Third Circuit multifactor likelihoodof- confusion test known as the Lapp factors, the district court stated it would “place emphasis” on the four Network Automation factors.

The strength-of-the-mark factor favored CollegeSource because it used COLLEGE SOURCE for eighteen years and CAREER GUIDANCE FOUNDATION for over thirty years, and both marks were suggestive and thus inherently protectable. The court found the strength-of-the-mark factor particularly relevant in the keyword-advertising-infringement context. Specifically, it stated that consumers searching with a generic term are more likely to be searching for a category “instead of knowing exactly what [they are] looking for from the outset,” whereas consumers searching for a suggestive or arbitrary term are most likely searching for the actual product.

The actual-confusion factor favored AcademyOne because there was “little evidence” showing actual confusion stemming from AcademyOne’s purchase of CollegeSource’s marks as keywords. CollegeSource did provide evidence that sixty-five Internet users who searched for CollegeSource clicked on an AcademyOne advertisement, but the court held that many of these individuals could have been employees of either company or those curious about the litigation. Furthermore, even if all sixty-five users were actually confused, the court deemed this evidence de minimis.

The care and attention expected of consumers when making a purchase also favored AcademyOne. First, quoting Network Automation, the court held that “the degree of care exercised by Internet users is becoming more heightened as the novelty of the Internet evaporates and online commerce becomes commonplace.” According to the court, the “modern Internet user’s increasing level of experience with search sites decreases the likelihood that they would be confused by the advertisements at issue in this case.” In addition, given the importance of consumers’ education-related inquiries, they were “likely to practice diligence in their research” and “exercise prudence.” The court concluded that “[m]odern Internet users, particularly ones who are interested in credit distribution in higher education institutions, are not likely to be confused by Internet advertising.”

The final Network Automation factor, labeling and appearance of the advertisements, also favored AcademyOne and turned out to be the critical factor in this case. Even though AcademyOne did not identify itself in its ads, the court found that the labeling and partitioning of the paid ads “decrease[d] the likelihood of confusion.” Specifically, the ads were separate from the organic listings, the ads were labeled as “sponsored links,” and some ads were even “differentiated by a shaded text box.” In addition, AcademyOne did not use CollegeSource’s marks in its advertisements, and instead used the marks only to trigger the ads. In short, the court held that the “entire context of the advertisement’s appearance, especially the fact that CollegeSource’s name does not appear [in] the advertisement,” favored AcademyOne.

The court next weighed the rest of the Lapp factors. It initially agreed with Network Automation that the similarity-of-the-marks factor was less relevant in keyword cases than in situations where consumers confronted two different marks and were unable to distinguish between them. The court thus “place[d] little weight” on this factor and found it neutral. The intent factor favored AcademyOne because there was no evidence of an intent to confuse. “The only evidence CollegeSource produce[d] to support its assertion [was] that AcademyOne knew of CollegeSource’s marks at the time it purchased its AdWords,” but the court held that mere knowledge of the marks was “insufficient to infer an intent to capitalize on CollegeSource’s goodwill.” The court also weighed the similarities in marketing channels in favor of AcademyOne, finding that CollegeSource did not participate in online advertising of any kind and did not compete with AcademyOne in keyword advertising. Finally, because both companies produced similar products and solicited business from similar clients, the court weighed both the similarity in targets of the parties’ sales efforts and similarity of the goods’ function in favor of CollegeSource.

In conclusion, the court held that the remaining Lapp likelihood-of-confusion factors “did not seriously weigh in favor of CollegeSource” and thus did not outweigh the three of four Network Automation factors that favored AcademyOne, including the critical “appearance of the advertisements” factor. The court thus granted AcademyOne’s motion for summary judgment.


This case is of interest because it is one of only a few district courts, and the first one in the Third Circuit, that has adopted the Network Automation factors for determining likelihood of confusion in the keyword context.