Promotion in Motion has lost its bid to overturn the refusal to register SWISSKISS chocolate trademarks, although the Federal Court of Appeal cautioned against excluding online consumer surveys merely because no interviewer supervises the responses.

In Promotion in Motion, Inc. (PIM Brands, Inc.) v. Hershey Chocolate & Confectionery LLC, 2026 FCA 40, the court dismissed the appeal with costs. The February 25 decision left intact the refusal of the SWISSKISS word and design marks for chocolate of Swiss origin.

The appellate court reviewed the challenged factual and mixed findings deferentially, asking whether there was a palpable and overriding error. New evidence allowed fresh analysis only on the issues it materially affected, not a wholesale rehearing of confusion. The court also upheld the discretion to admit the late-filed certification-mark affidavit, given its documentary nature and the absence of prejudice.

Geographic certification did not identify PIM as the source

Hershey opposed the applications based on its KISS, KISSES and other marks. The Trademarks Opposition Board found that PIM had not established, on a balance of probabilities, that there was no reasonable likelihood of confusion. The Federal Court upheld that result.

PIM relied on additional evidence that SWISS, SUISSE and SWITZERLAND were registered certification marks owned by a Swiss chocolate manufacturers’ association. It argued that incorporating SWISS into its marks contributed distinctiveness that the Federal Court had undervalued.

The appeal court explained that certification marks fall within the statutory definition of trademarks but perform a different function from ordinary source-identifying marks. Geographic certification indicates compliance with a standard concerning where goods originate; it does not necessarily identify one trader.

The special registration rule for descriptive certification marks allowed SWISS to be registered despite its geographic descriptiveness. That did not give the word inherent distinctiveness as part of PIM’s proposed marks. The court also found no direct evidence in the record establishing that the association had licensed or permitted PIM’s use, as distinct from withdrawing an earlier opposition in United States proceedings.

In the confusion analysis, SWISS remained largely descriptive. KISS was the most striking component, creating a meaningful resemblance to Hershey’s registered marks. The evidence did not establish a separate market or trade channel insulating Swiss chocolate from other chocolate products.

Survey design, not just survey format, mattered

PIM submitted online surveys reporting low levels of association between SWISSKISS and Hershey. The Federal Court rejected the survey evidence, including a survey submitted by Hershey, for validity and reliability concerns.

The appeal court found no reversible error in the concerns about repeatedly directing respondents toward Swiss chocolate and narrowing the respondent population. Those features could orient answers away from non-Swiss manufacturers rather than reproduce how consumers encounter chocolate in stores.

It also upheld concern about whether participants could use a browser’s back button to view a mark again after it disappeared. The relevant test sought to simulate a hurried consumer’s imperfect recollection, and the expert had not tested that possibility in the live survey.

But the appeal court disagreed with treating the absence of a supervising person as enough, without further analysis, to exclude Internet-based panel surveys. Properly designed online surveys may supply useful evidence. Their safeguards, sampling, verification and particular circumstances must be assessed case by case; some concerns may affect weight rather than admissibility.

That error did not change the outcome because the remaining grounds for rejecting the surveys rested on supported factual findings.

No universal percentage resolves confusion

The court also rejected PIM’s proposed premise that Canadian law generally requires a confusion rate above 10 per cent. The lower figure discussed in the Federal Court depended on assumptions and adjustments accepted for argument’s sake.

More fundamentally, a survey asking consumers to name a product’s source measures identified confusion under its design conditions. A low result does not by itself prove an absence of likely confusion. The registration dispute therefore could not be resolved simply by selecting a numerical cutoff.

The ruling preserves the refusal of these applications while leaving online survey evidence available for careful, case-specific evaluation in future disputes.