One in Three Olive Oil Samples: CFIA's 2024-25 Food Fraud Report and your supply chain
I look forward to the Canadian Food Inspection Agency's Food Fraud Annual Report each year. Active surveillance testing in Canada moved to an annual cycle around five years ago, and there’s always something illuminating in it. This year the illuminating thing was the inclusion of the words “67 percent.” Of the olive oil CFIA tested for authenticity in 2024-25, only 67 percent came back satisfactory. In other words, close to one in three samples did not.
I tend to read CFIA’s testing results carefully. CFIA aims this sampling at higher-risk products and businesses, using factors such as a history of non-compliance, unusual trading patterns and gaps in preventive controls, and the Agency expressly says the results are not representative of the Canadian marketplace. So, CFIA is not saying that one-third of the olive oil on Canadian shelves is inauthentic. It says the opposite, expressly, and that, among the products and regulated parties CFIA considered worth testing, olive oil produced the lowest satisfactory rate of any commodity in the report's authenticity testing.
Even with that context, if you import, bottle, or brand olive oil in Canada, that is still a serious signal.
Food fraud is a strange thing: it is hard to define, to say nothing of measuring it. Codex Alimentarius has fought over its definition for years, and whether geographical indications and provenance-based substitutions belong under the umbrella at all is still not settled among the people who work on it. For North American food regulators, this complex issue often sits distantly behind the core mission of human health.
CFIA helpfully gives an example of the economics it has considered when producing the report. Its inspection work identified fish labelled as Patagonian toothfish, which sells for approximately $41.70 per kilogram, that was in fact Antarctic toothfish, a lower-value species selling for about $23 per kilogram. So, a visually identical but economically inferior substitution to the tune of $19/kg moving quietly from the buyer to whoever made the substitution is the kind of thing CFIA is considering when developing and releasing this report. From an agency that is generally focused on the health and welfare of Canadians, this is helpful information for the consumer: they’re vulnerable to the ancient fraud of buying one thing and receiving another. And, from a policy standpoint, the report is evidence of competitive distortion: a business carrying the cost of the genuine article can be undercut by a product that is not what it says it is, wherever in the chain that mismatch arose.
The word fraud can make this sound as if every failed result proves dishonesty, which it doesn’t. I think the central thrust of CFIA’s food fraud program is to address food misrepresentation, whether or not the testing establishes who caused the mismatch or why. I think that distinction matters, because there’s a difference between a supplier-management problem and a problem you may not be able to manage your way out of.
Many in the food fraud community, and I am one of them, have been waiting to see what deeper surveillance and enforcement would look like ever since the Food Policy for Canada put $24.4 million behind tackling food fraud in 2019. Here it is, taking shape: an independent marketplace-monitoring stream, targeted testing aimed where the risk intelligence points, label and quantity verification inside the same program, and a seven-figure prosecution finally landing (we’ll get to that).
What CFIA measured
The report brings together three different kinds of verification: authenticity testing, basic label verification and net quantity verification. They sound similar but are not interchangeable. And alongside them sits an awareness section that is perhaps more important than any of the three.
CFIA tested 886 samples for authenticity or misrepresentation. Of those, 246 were marketplace-monitoring samples collected at retail by an independent third-party, and 640 were targeted inspectorate samples from up the supply chain, taken by CFIA inspectors at the regulated party responsible for the product in Canada. In short, marketplace monitoring is designed to gauge compliance in the marketplace whereas targeted inspectorate sampling is designed to find problems.
And where do the samples come from? I am always curious about how CFIA builds its surveillance. We’ve made access-to-information requests in past years about olive oil surveillance, in which CFIA demonstrated a surprisingly broad sampling base: small independent grocers, often serving particular cultural communities, alongside national and regional chain grocers. A retailer can be the regulated party for what it packages and labels in store, so that breadth fits the program design. This year, though, the olive oil sampling ran through one supply chain access point: every one of the 85 targeted samples was taken at import.
The marketplace-monitoring stream, which covered fresh meat and tea this year, came back fully satisfactory. The commodity results that attract attention, including the 67 percent olive oil result, all came from the targeted inspectorate program.
CFIA also performed 362 label verifications: 141 basic label verifications, of which 77 percent were compliant, and 221 net quantity verifications. Basic verification examines the label itself against the mandatory requirements (the report sets out the full element list), but the finding of seventy-seven percent compliance on basic labelling in the report is quietly likely to describe most food businesses’ own products: not fraud, just ordinary mandatory label content, checked properly
Net quantity verification asks whether the amount of food in the package meets, on average, at least the declared quantity, subject to the applicable tolerances (the Canada Agricultural Review Tribunal decided exactly this question in June: Metro Ontario Inc. v Canadian Food Inspection Agency, 2026 CART 29, where a seven-gram shortfall on a package of striploin, one gram outside the tolerance, was confirmed as a violation).
Putting this all together, a product can be authentic and still have a non-compliant label. It can be correctly identified and still be underweight. Each is a separate problem and each calls for a different control. In law, though, they converge.
The category results, and what they mean
Within the 640 targeted inspectorate samples, CFIA reported the following satisfactory rates:
Olive oil: 67 percent (57 of 85 samples)
Honey: 81 percent (66 of 81)
Grated hard cheese: 84 percent (31 of 37)
Other expensive oils: 86 percent (49 of 57)
Fruit juice: 88 percent (151 of 172 satisfactory; 12 unsatisfactory and nine investigative)
Fish: 90 percent (63 of 70)
Meat: 94 percent (89 of 95)
Maple syrup: 100 percent (43 of 43)
Remember when you’re reading this: these numbers do not tell you about the proportion of fraudulent food on Canadian shelves. CFIA presumably re-aims this program each year at wherever its risk-intelligence points, so the results are risk-weighted, not representative. They don’t even really work as a trend-line report. So, a perfect result in 43 maple syrup samples does not establish that the marketplace is fraud-free, just as the olive oil result does not establish that one-third of the market is inauthentic. Case in point, see this article I wrote earlier this year on Québec-based maple syrup fraud for the Food Authenticity Network.
What the numbers will tell you is what CFIA found when it went looking in the places it judged most likely to have a problem. That is category-specific risk intelligence, drawn from international authorities on food fraud, and it is enough for food-sector stakeholders to ask whether the controls for the products they handle are proportionate to what the Agency is finding.
There is also a dissonance worth naming between CFIA's results and what independent testers find. Oceana Canada's targeted testing of high-value seafood at retail and in restaurants, using DNA barcoding by Dr. Robert Hanner's lab at the University of Guelph, has repeatedly reported mislabelling in the mid-forty percent range, whereas the fish sampling in this report, and previous CFIA reports, found roughly 10 percent of the fish tested to be unsatisfactory, and previous CFIA reports have landed in the same band. That is not to say something is wrong with this year's fish number, necessarily. The two programs sample different species, at different points in the chain, for different purposes. But the gap is a reminder that a sampling program's results describe the program as much as the marketplace.
Maple syrup makes the same point from the clean end of the sampling table: all 43 samples CFIA drew in 2024-25 came back satisfactory. A year after that sampling window closed, the Bourdeau affair arrived, and it surfaced through a Radio-Canada investigation and the Québec producers' federation rather than through any sampling program. The proposed class action reaches purchasers back to April 2023, which takes in the very year those 43 samples all passed. A clean sampling year and a major adulteration story can describe the same twelve months without contradicting each other.
What an unsatisfactory result proves
For the purposes of the report, CFIA assesses a sample as satisfactory when it meets the regulatory criteria or parameters being tested. An unsatisfactory result establishes a mismatch against those criteria. It does not, by itself, establish where in the supply chain the problem arose or whether it was deliberate.
For olive oil, that may be adulteration with a lower-value oil, or an extra virgin, cold pressed or unrefined claim the product does not meet. For fish or meat, a species substitution. For honey or maple syrup, foreign sugars. The testing identifies the product problem. Inspection and follow-up work identify its source and the appropriate response.
As a law firm, this is where we start: Section 6(1) of the Safe Food for Canadians Act prohibits manufacturing, preparing, packaging, labelling, selling, importing or advertising a food commodity in a manner that is false, misleading or deceptive, or that is likely to create an erroneous impression regarding its character, quality, value, quantity, composition, merit, safety or origin, or the method of its manufacture or preparation. Section 5(1) of the Food and Drugs Act carries a similar prohibition with a slightly narrower list. CFIA does not first have to prove criminal fraud before it can act on a product that is not what it is represented to be. And in the food space, we’re typically working with strict liability: the act matters, not the intent. How much legal work that file does depends on which route the Agency takes.
That’s not to say that intent is irrelevant: under CFIA's Standard Regulatory Response Process, harm, compliance history and intent, including negligence, drive which enforcement response the Agency selects, and genuinely intentional deception lives further up the ladder, in Criminal Code fraud and the Competition Act's representation provisions. On the prosecution track, due diligence is a statutory defence under the SFCA (s. 39(2)) and under the FDA (s. 31.3). As a point of practice, it’s worth noting where this defence disappears: if a designated SFCA or SFCR contravention proceeds by notice of violation under the administrative monetary penalty regime, section 18 of the Agriculture and Agri-Food Administrative Monetary Penalties Act expressly removes both due diligence and reasonable mistake of fact as defences.
Practically, this means the due diligence file has to be built before the test fails, not after. The same supplier controls and verification work prevent the mismatch, help locate where it occurred, and support your management of the issue when responding to CFIA. How much legal work that file does depends on which route the Agency takes. If you read the section 6(1) list again, you’ll see character, quality, value, quantity, composition, merit, safety or origin. For products on the market, authenticity testing, net quantity verification and origin-claim work are separate ways of checking three different elements in the same subsection, which is why one annual report covers all three, and why one blog post has to.
What CFIA did with the failures
Where CFIA found problems, it removed products from Canada, detained them, destroyed them or required relabelling. The Agency described the result as more than 150,000 kilograms of misrepresented food prevented from being sold. Almost all of that weight was honey, at 133,420 kilograms; the rest runs from 7,245 litres of adulterated olive oil all the way down to 156 misrepresented oat cakes. Sum the itemized figures, though, and you get about 134,700 kilograms plus roughly 9,900 litres, which does not quite reach the headline. The report does not explain the gap.
CFIA issued 13 letters of non-compliance, two notices of violation with warning and six notices of violation with penalties. The eight notices of violation are the eight administrative monetary penalty cases the report counts, and they totalled $60,000.
The report also points to the MPY Trading Ltd. prosecution. MPY falsely described crab imported from the United States as Product of Canada in export certificates to China, and failed to keep adequate records of the source of the crab it exported. The court imposed fines totalling $1,155,685. We check the CFIA prosecution bulletins as they post, and as of July 2026, MPY is the largest food-misrepresentation fine listed. The conduct ran from December 2019 to January 2020 and the fines landed in December 2024, which is what CFIA means when it warns that prosecutions and administrative proceedings may span several years.
Origin claims have moved from complaints to penalties
The part of the report I would flag for clients is not in the testing tables. It is in the awareness section, and it is about Canadian-origin claims.
When the Buy Canadian movement gained momentum in February 2025, CFIA saw complaints about origin claims increase from one or two per month to 25 in February and 33 in March. The Agency followed up on all of them and issued an industry notice reminding businesses to use Product of Canada, Made in Canada and other origin claims accurately.
In March 2026, CFIA reported five financial penalties totalling $47,000 for inaccurate or misleading origin claims since April 2025, and told media it had identified more than a hundred origin-claim non-compliances over the same period, and that inspectors were reviewing claims on labels and in advertising, including in-store signage. That activity falls in the year after the period this report covers, which tells you where enforcement is headed.
None of this is separable from the trade climate Canada is experiencing in 2025-2026: origin claims have become political objects in Canada, and consumers and regulators are both reading them more closely as a result. The two claims do different work: Product of Canada means all or virtually all major ingredients, processing and labour are Canadian; CFIA generally treats non-Canadian material under two percent as minor, though the percentage is not the whole test, since the significant ingredients, processing and labour must also support the claim. Made in Canada means the last substantial transformation happened here, and it requires a qualifier saying the product was made from imported ingredients, or from domestic and imported ingredients. Both claims are voluntary. The exposure is not: an origin representation that creates an erroneous impression is what section 6(1) reaches.
We have been auditing these claims for clients because of increased attention, and are finding that complaints are coming from product pages, digital advertising, retailer-supplied copy, shelf signs and maple leaf imagery, all of which can be considered to create an origin representation.
Net quantity already belongs in the food fraud conversation
CFIA completed 221 net quantity verifications in 2024-25 and found 88 percent compliant overall: 92 percent for meat (122 of 133), 84 percent for fish (61 of 73) and 73 percent for the small "other foods" group (11 of 15). Most of the meat number comes from a separate dedicated project CFIA ran between January 25 and February 28, 2025, on meat packaged, labelled and weighed at retail. Where it found inaccurate net quantity declarations, it issued notices of violation with warning to the stores involved.
The legal requirement is not new - the more interesting point is methodological: CFIA now treats quantity the way it treats identity, inside the food fraud program, with formal warnings when a project finds problems. I wrote about the three-regulator split at the meat counter earlier this year, but this point is narrower: an inaccurate quantity declaration is a misrepresentation issue, not a production variance to be left with operations.
What to do with this
So, if you’re a regulated party, what should you be doing with this tome of a blog post? There are three things to review, I think, in order of how much they tend to get neglected:
Treat authenticity as a supplier-qualification problem if you handle a high-risk category. In CFIA’s 2026-27 Inspection Campaign, we are seeing increased attention from the regulator on supplier verification programs within preventive control plans. If you don’t have a third-party auditor doing this work, then take some time to define what authenticity means for the product and what evidence will establish it. We find that a generic certificate of analysis is useful only if it addresses the characteristic in question using an appropriate method. For higher-risk oils, honey, premium fish and other particularly valuable ingredients, especially those that rely on provenance, that may mean lot-level documentation, periodic independent testing, supplier audits and a response protocol for results that do not fit. Contract terms for bulk goods are difficult to vary, but orders for specialty goods can often provide for testing costs, audit rights and specific recourse against a supplier. While you can work through contract, these terms do not transfer your regulatory responsibility, so due diligence is important.
Audit your Canadian-origin claims now. We’re still seeing increased complaint activity on origin claims, with the Elbows Up / Buy Canadian movement showing no sign of fading. So, aggregate your claims from labels, packaging, websites, retailer listings, advertisements and in-store material. Test each one against the current ingredients, processing and labour, and confirm that the claims are substantiable. If a Made in Canada claim lacks its qualifier, or a Product of Canada claim depends on imported content that is more than negligible, fix it before it becomes a complaint or an inspection finding. Given complaint and enforcement volumes, and media interest, these are easy issues to resolve. If unattended to, they carry some not-insignificant exposure for many consumer brands.
Treat net quantity as a representation, because CFIA is now testing it as though it is. So average-quantity controls, fill settings, check-weigh data and change-control records should carry more weight: they are what substantiates the declaration on the label. Where catch-weight products are involved, include tare programming and packaging changes in the review.
Ultimately, for food-sector businesses, the Agency's overall numbers will not tell you the prevalence of fraud in your category or the condition of the product on your shelf, but this report helps tell you where CFIA is looking and what it is finding when it looks. And for olive oil this year, 67 percent is a very good reason to check whether the controls behind your label are doing the work you assume they are.
---
Glenford Jameson is the principal of GSJ&Co., a Canadian food law and regulatory affairs practice. This post is for informational purposes only and does not constitute legal advice.