Likelihood of confusion (LOC) surveys assess if the reasonable consumer will be confused, mistaken, or deceived as to the source or affiliation of goods or services.[i] Although surveys are not a requirement in trademark and trade dress litigation, they are frequently used as evidentiary support, beginning in the pretrial phase and continuing through settlement or verdict.[ii]
This article provides an overview of two commonly utilized LOC survey methodologies: the Eveready and Squirt formats. We will focus on surveys measuring forward confusion, which can occur when consumers mistakenly believe that the products or services of the junior user, the party that adopted the mark later, come from, or are in some way affiliated with, those of the senior user, the party that first used the mark. The principles described below generally apply to surveys measuring reverse confusion, which can occur when the parties’ roles are reversed, though certain details may differ.
Comparing the Eveready and Squirt Survey Formats
Courts have historically accepted the Eveready and Squirt survey formats. Both typically ask questions that assess consumer perceptions about the source of the marks, as well as whether respondents perceive the marks as connected to, affiliated with, sponsored by, or approved by a particular company or brand. The answers to these questions can provide evidence about likelihood of confusion for trademark matters. The appropriate LOC survey format depends on the circumstances of the case.
Eveready Survey Format: Testing Confusion Without Side-by-Side Marks
The Eveready survey format originated in the 1976 Union Carbide Corp. v. Ever-Ready Inc. case.[iii] The plaintiff, Union Carbide, had been selling batteries, flashlights, and miniature bulbs under its trademark EVEREADY when the defendant began to sell similar products, including lamps and miniature light bulbs, under the name Ever-Ready. Union Carbide requested “an injunction against Ever-Ready’s use of the term Ever-Ready on or in connection with the advertising or sale of electrical products.”[iv]
What the Original Survey Included
In order to assess likelihood of confusion between the two brands’ products, the plaintiff conducted a survey in which the respondents were shown a picture of the defendant’s lamp.”[v] Respondents were then asked who they thought produced the product, why they thought so, and to name any other products from the same source as the lamp. This last question was included to help determine whether respondents who answered “EVEREADY” were referring to the senior or the junior user, since the two names and their spellings were nearly identical. (This question remains helpful for modern matters involving marks with similar or identical spellings.)
The survey questions were open-ended, with respondents choosing their own phrasing. The results indicated that more than half of the respondents believed that batteries (a Union Carbide product) were produced by the lamp’s maker, and the court determined there was an appreciable likelihood of confusion between EVEREADY and Ever-Ready.[vi]
How Eveready Surveys Are Used Today
The Eveready format is now widely used to measure likelihood of confusion and has been called a “standard and widely accepted survey format.”[vii] In a forward confusion case, an Eveready survey presents consumers with only the junior user’s mark and asks respondents who they believe uses the mark, is affiliated with, or endorsed by, the user of the mark. In this type of survey, prospective consumers of the junior user’s product are interviewed.[viii] (The format is modified somewhat in reverse confusion cases. In a reverse confusion Eveready survey, only the senior user’s mark is shown, and respondents are asked the same series of questions. In this case, the survey interviews prospective consumers of the senior user’s product.[ix])
Since its introduction 50 years ago, the Eveready survey format has evolved in several respects. Most notably, Eveready surveys today typically include controls, a survey design that allows researchers to remove the effects of biases and other extraneous influences from their data. For more on using controls in litigation surveys, please see the discussion of controls below or read our recent article.
A Newer Example: MetaBirkins and Birkin Handbags
Mason Rothschild created and sold non-fungible tokens (“NFTs”) called “MetaBirkins.” The MetaBirkins consisted of digital images of Birkin handbags covered in colorful faux fur. Hermes filed a lawsuit in January 2022 against Rothschild, alleging confusion between MetaBirkins and its Birkin bag. Rothschild argued that the project was protected artistic commentary on fashion’s fur-free movement.
Hermes retained a survey expert,[x] who conducted an Eveready survey to measure likelihood of confusion between the MetaBirkin web page and the Hermes Birkin bags. The MetaBirkin web page advertised MetaBirkins NFTs, showed pictures and text related to the NFTs, and provided links to online marketplaces where consumers could purchase the NFTs. The survey showed this web page to NFT purchasers and then asked questions to measure likelihood of confusion with Hermes and Birkin handbags. Respondents were shown one of two images: either MetaBirkin’s web page as it existed in the real world, called the “test version” in surveys, or a web page image that replaced “NOT YOUR MOTHER’S BIRKIN” with “NOT YOUR MOTHER’S HANDBAG,” the “control version.” Respondents were never shown any website or images from Hermes.
Respondents first answered questions designed to measure likelihood of confusion using the Eveready survey format. They identified the company, companies, person, or people they believed made or provided the items on the web page and explained their reasoning in response to an open-ended question. Next, respondents were asked whether they were aware of any other brands or products made or provided by whoever makes or provides the items shown on the web page. For those who answered yes, open-ended questions followed, asking which other brands or products they believed were made or provided by whoever makes or provides the items shown on the web page, and an additional open-ended question that probed for any other brands.
Finally, respondents were asked whether they believed whoever made or provided the items was sponsored, authorized, or approved by any other company. Those who indicated that they believed such an affiliation or connection existed were asked to identify the company, person, or brand they believed sponsored, authorized, or approved the website operator, and to explain what led them to that belief.
Taken together, NFT purchasers’ answers to the questions showed that a notable percentage provided a response that referred to either Hermes or Birkin. These results indicated a likelihood of confusion between the junior user, MetaBirkin, and the senior user, Hermes. As a result of this likelihood of confusion Eveready survey and other evidence, the court sided with the plaintiff, Hermes.[xi]
Squirt Survey Format: Testing Confusion in Marketplace Proximity
Because an Eveready survey does not present respondents with both marks simultaneously, it may not fully capture confusion that could arise when consumers encounter competing marks in close marketplace proximity.[xii] Accordingly, a different survey format may be appropriate when marketplace conditions make direct comparison of the marks more likely.
When the Squirt Method Was Developed
The Squirt survey format originated from the 1980 SquirtCo v. Seven-Up Co. case.[xiii] SQUIRT was a trademark of SquirtCo. When the Seven-Up Company introduced QUIRST, SquirtCo argued that the similarity of the names was confusing to customers.
SquirtCo commissioned surveys to measure the likelihood of confusion between the two names. In the surveys, respondents were exposed to radio ads for both SQUIRT and QUIRST, as well as radio ads for two other products, and then asked whether they believed that SQUIRT and QUIRST were made or put out by the same company or by different companies.[xiv] Like in the original case, Squirt surveys today typically show both the plaintiff’s and the defendant’s products to survey respondents and ask closed-ended questions to assess likelihood of confusion.
How the Approach Has Evolved
Squirt surveys may be conducted in a variety of ways, depending on the specific market context. However, Squirt surveys should always include an array of products. This means that other distractor brands from the same product category are shown in the survey in addition to the products from the plaintiff and defendant.[xv] This use of an array is why the Squirt format is sometimes referred to as a “lineup” survey.
Like Eveready surveys, Squirt surveys now routinely include controls. Controls for Squirt surveys are created in the same way as those for Eveready surveys. However, because Squirt surveys often include a lineup of competing marks that are not involved in the dispute, the control group sees the control mark as part of the lineup, in the place where respondents in the test group would see the allegedly infringing mark.
The original Squirt survey was what is now referred to as a one-room survey. That means the plaintiff’s and defendant’s products are shown to respondents in an array at the same time, along with other similar products. However, many Squirt surveys use a two-room format, where the allegedly infringed mark is shown first, followed by a lineup of other marks in the category and the allegedly infringing mark.[xvi] The sequence described here is that of a forward confusion survey; a survey measuring reverse confusion would reverse the order of presentation.
As marks from both parties are shown to respondents in a Squirt survey, this format replicates conditions under which marks have marketplace proximity. As one author has stated, this format is used when marks “are simultaneously or sequentially accessible in the marketplace for comparison. It relies on the ‘proximity’ factors (e.g., overlapping customers, channels of trade and advertising) in a likelihood of confusion analysis.”[xvii]
Why Proximity Affects the Format
When using a Squirt-format survey, measuring competitive proximity between the two trademarks at issue is required to provide reliable results. To be competitively proximate, products may exist near each other in brick-and-mortar stores (seen side by side or in quick succession) or may appear near each other in the “broader internet marketplace.”[xviii] Although it can be simple to assess proximity in brick-and-mortar buildings, doing so in online environments can be a more complex process: “The mere fact that a senior user’s and a junior user’s marks both appear somewhere on the internet is, by itself, basely insufficient to justify use of the Squirt format.”[xix] As a result, authorities have explained that the products should be commonly explored online and that common search terms should produce both products “within a manageable number of hits.”[xx]
According to a 2019 journal article, “What constitutes a manageable number of hits is marketplace dependent.”[xxi] The authors outline four scenarios in which there may be sufficient online proximity to justify using a Squirt survey format:
- First, a single, commonly used search query leads to websites from both parties appearing on the first page of results.
- Second, among multiple commonly used search queries, both websites appear on the first page at least once. This scenario assumes that most consumers search for the product more than once, and that at least one of those searches will return products from both parties.
- Third, a consumer sees a mark online and soon after comes across the other mark in a brick-and-mortar store.
- Fourth, the consumer is likely to go directly to the website of one party and is likely to encounter the other party while searching the internet.[xxii]
Ultimately, what constitutes proximity on the internet varies widely, depending on product categories and whether the products are high-involvement goods. As such, there is no single definition of internet proximity that justifies using a Squirt format. Instead, researchers need to assess “whether the two brands with appreciable frequency are likely to exist simultaneously in [the] cognitive workspace.”[xxiii]
A Newer Example: Crystal Head Vodka and KAH Tequila
In Globefill Inc. v. Elements Spirits, Inc., Globefill’s Crystal Head Vodka was sold in a distinctive skull-shaped bottle. When Elements Spirits, Inc. began offering KAH Tequila in skull-shaped bottles, Globefill alleged that consumers would be confused into believing that the tequila bottles were made or authorized by Crystal Head.
A Squirt survey was conducted to measure the likelihood of confusion between KAH Tequila and Crystal Head vodka. The survey was conducted in person with tequila drinkers (the customers of the junior user, Elements Spirits). Respondents first saw a bottle of Crystal Head vodka and then were shown a lineup of bottles of tequila, including KAH Tequila in a skull-shaped bottle or one of two control bottles. This replicates a scenario in which a consumer in a retail location, such as a liquor store, sees Crystal Head and then sees KAH among other tequila bottles.
Several questions common to the Squirt format were used in this survey. For example, respondents were asked whether any of the products in the lineup were made by the company or person who made the first product they saw. If they replied affirmatively, respondents could choose from among the bottles of tequila in the lineup. This question measured confusion as to source, and another series of questions measured confusion as to sponsorship or approval. The results indicated likelihood of confusion between the accused bottle of KAH tequila and Crystal Head vodka.
The seven-year case concluded in 2017; the jury found in favor of Crystal Head, and the court issued a permanent injunction and final judgment consistent with the jury’s verdict.[xxiv]
Three Key Considerations for Survey Design
There are many important factors to consider when designing Eveready and Squirt surveys. Each of the elements below could be discussed in much more detail than space allows in this article. However, we would be remiss not to provide an overview of these topics and their importance to these commonly used LOC survey formats.
1) Selecting the Proper Survey Universe
As a leading authority has noted, “Courts consider the selection of the proper universe as one of the most important factors in assessing the validity of a survey as well as the weight that it should receive… ‘Selection of a proper universe is so critical that even if the proper questions are asked, the results are likely to be irrelevant.’”[xxv]
Regardless of the approach, one of the first steps in designing a litigation survey is to define the appropriate survey universe, which is the group of people whose thoughts, beliefs, or opinions are relevant to the legal question.[xxvi] From this survey universe, a sample can be drawn. A forward confusion case typically interviews the junior user’s prospective consumers, whereas a reverse confusion case interviews the senior user’s prospective consumers.[xxvii] A survey that reaches a population reasonably similar to the relevant consumer group may still be considered admissible, although a mismatch in the universe may diminish the weight afforded to the results. By contrast, when the surveyed population differs substantially from the consumers whose perceptions are relevant to the dispute, the survey may be deemed so unreliable that it is excluded altogether.[xxviii]
Overinclusive and Underinclusive Survey Universes
Common ways that a survey universe can be selected incorrectly include selecting the wrong survey universe entirely, selecting a partly correct universe that may be overinclusive, and selecting a partly correct universe that may be underinclusive. In each case, the survey may be afforded little weight.[xxix] A survey universe is overinclusive when it includes consumers whose state of mind is not relevant to the legal issue at hand, as well as those whose state of mind is relevant. For example, in a forward likelihood of confusion case, an overinclusive universe might incorrectly include prospective purchasers of the senior user’s product who are not also prospective purchasers of the junior user’s product.
On the other hand, the universe may be underinclusive when it excludes an entire group of consumers whose opinions are important in the matter. For example, an LOC survey may interview respondents in only one geographic location, as opposed to all locations relevant to the matter, neglecting to assess the beliefs of a significant proportion of relevant consumers.
2) Using Controls to Measure Confusion Reliably
Controls are commonly used in trademark surveys because the survey results can be affected by extraneous elements, such as pre-existing knowledge and attitudes among respondents, guessing or inattentiveness by respondents, or other factors unrelated to the real issues of interest. A control measures the effect of these extraneous influences. To remove those effects, researchers subtract the control measures from the test measures, creating a net measure. As discussed below, the net confusion measure is the one courts often rely on to assess whether there is a likelihood of confusion.
It has been said that any discussion about likelihood of confusion surveys “would be incomplete without an acknowledgment of the crucial importance of controls.”[xxx] Both major likelihood of confusion survey methodologies should incorporate effective controls to adjust for survey noise or unforeseen biases in survey responses. In an Eveready survey, a control is particularly critical to filter for brand dominance. In the original Eveready case, the lack of a control was a primary criticism.[xxxi] In Squirt surveys, controls are critical to adjust for guessing between marks that are shown side by side.[xxxii]
Control stimuli are typically shown to a different group of respondents from those shown the at-issue trademark or trade dress. The control is normally designed to be reasonably similar to the test item, except that it does not include the elements under dispute. A control should share “as many characteristics with the experimental stimulus as possible, with the key exception of the characteristic whose influence is being assessed.”[xxxiii] Therefore, respondents whose answers indicate confusion after looking at the control stimuli are likely responding based on preconceived biases, a lack of attention, or factors other than the at-issue elements.
3) Meeting the Minimum to Indicate Confusion
With both the Squirt format and the Eveready format, there is no set numeric threshold indicating confusion; rather, courts look for an “appreciable” amount. Historically, a net measure of around 15% has generally been accepted as indicative of a likelihood of confusion, but accepted percentages vary depending on the matter and the court..[xxxiv] As one authority explains, “numerous courts have found likelihood-of-confusion surveys with percentage results between 10% and 20% probative of likely confusion.”[xxxv] However, courts tend to rule against measures under 10%.[xxxvi] Furthermore, the survey must be well-designed for percentages indicating a likelihood of confusion to be afforded weight.
Expertise Is Critical in Likelihood of Confusion Surveys
Survey evidence indicating a likelihood of confusion, or a lack thereof, can greatly aid in trademark and trade dress proceedings by providing relevant consumer perception evidence of an allegedly confusing mark. Likelihood of confusion surveys are commonly used in trademark litigation, and a well-designed survey is generally accepted as additional evidence in a case. However, while LOC surveys can provide structured, empirical insight into consumer beliefs and perceptions, the weight they may be afforded depends on the methodological choices underlying them.
As a result, assessing likelihood of confusion through survey evidence requires a disciplined approach. The Eveready and Squirt formats are two reliable, court-accepted methods to structure a LOC survey, but each has case-specific considerations. A survey expert can help you choose which methodology is appropriate to the unique facts and issues in your matter.
Trusted Experts for Every Stage of Trademark Litigation
From initial case assessment through survey design, implementation, analysis, and testimony, IMS survey experts deliver reliable evidence for trademark and trade dress disputes, including likelihood of confusion surveys tailored to the circumstances of each matter. Contact us today to discuss your trademark survey needs with our team.
