Manufacturing

Shelf space is bought, not won. A new CPG brand pays slotting fees, funds trade promotions, carries the retailer's payment terms and absorbs returns before it learns whether the product sells, and co-manufacturing minimums put the first production run in the tens of thousands of units. The direct-to-consumer route that was supposed to bypass all of that now costs more in paid acquisition than the margin on a grocery-priced item supports. The product content software this industry buys is screened separately at the same code.

The pre-screen dismissed this as feed milling and plant capital. The better half of the code deserves a harder look. Freshpet, the listed pure play in refrigerated pet food, reported 2025 net sales of $1,102.0M, up 13.0%, at a 40.8% gross margin [A]. That is the attraction. The cut is what the growth cost. Freshpet's capital expenditure was $239.1M in 2023, $187.1M in 2024 and $148.2M in 2025 — $574M in three years — and 2025 was its first year of positive free cash flow, $12.4M [A]; it guides to roughly $150M of capital again in 2026 [A]. Over those three years it generated $390.8M from operations [A] and put half as much again into plant. A company with a billion dollars of sales and a differentiated product was still spending more on kitchens than the business threw off. Pet food is made on extrusion, retort or chilled lines in facilities inspected to a standard retailers and export markets will accept, and the product has to be on the shelf, in volume, every week. The Canadian shape says the same thing from the other side: the largest single size band among the 434 establishments is 20–49 employees (31%) [A] — these are plants, not workshops. The way around the plant is a co-packed treat or kibble brand, which is no longer manufacturing; it is the Consumer Packaged Goods Brand Operation already screened at 311, with that record's shelf-access problem intact. Livestock feed, the other half of the code, was not researched.

This is a growing industry, which is not what the pre-screen's "commodity margins" suggests. Statistics Canada reports that Canadian processors crushed a record 11.6 million tonnes of canola in 2025, up 1.6% and the third consecutive annual record, and that domestic crushers took 50.5% of all canola disposition in the 2024/25 crop year against 41.5% exported as seed [A]. More of the crop is now processed where it is grown than leaves the country as seed. The cut is the size of the ticket to take part. When Cargill announced its Regina crush plant it was reported as a $350M facility with one million tonnes of annual capacity [B] — at least $350 of plant for every tonne of yearly throughput, before a tonne of seed is bought. And the plant, once built, sets no price: seed is bought against a futures market, oil and meal are sold against theirs, and the crusher earns the spread the board leaves. That rewards the lowest-cost, best-located, largest plant with its own rail and origination, which is why the builders are global grain houses. Flour milling, wet corn milling and breakfast cereal share the shape: US establishments in this group average 63 employees [A]. The 84 Canadian establishments with fewer than ten employees [A] are stone mills and cold-press oil makers. What they sell is a premium bag of flour or bottle of oil, and that is a packaged-goods brand — screened at 311 — that happens to own a mill.

The pre-screen saw a craft-scale entry with the shape of the bakery record at 3118, and the counts support that there is one: 169 of the 330 Canadian establishments have fewer than ten employees [A]. The sugar end of the code is simply closed — Rogers Sugar is spending $280–300M to add 100,000 tonnes of refining capacity to a business that sold 781,454 tonnes in fiscal 2025 [A], and nobody enters refining beside that. Chocolate and candy are where an entrant can start, and the test is what making them earns. Rocky Mountain Chocolate Factory is a long-established listed brand with its own factory and a captive franchise network to sell through. In fiscal 2026 it reported total revenue of $27.5M, down from $29.6M, and product and retail gross profit of $0.7M — after $0.1M the year before — with a net loss from continuing operations of $4.6M [A]. With the brand, the plant and the channel already in hand, manufacturing confectionery earned it almost nothing. The mechanism differs from bread. Confectionery is bought a few times a year as a gift, against a shelf price set by multinational brands that buy cocoa and sugar at scale, while the small maker buys the same world-priced inputs in pallets. The premium a customer will pay is real but seasonal and capped, and it is paid at the counter — so what works is a shop with a kitchen behind it, which is specialty food retail rather than manufacturing, and is won or lost on the lease.

The pre-screen pointed at small-batch preserving reached through co-packers, and the small end exists: 163 of 376 Canadian establishments have fewer than ten employees [A]. But the group is unusually top-heavy — 60 establishments, 16%, employ a hundred or more [A], and the average US plant has 83 staff [A] — because freezing and canning are harvest-timed, high-throughput operations contracted to growers a season ahead. The question is whether the category rewards anyone for being in it, and the cleanest recent answer is Green Giant. B&G Foods reports a Frozen & Vegetables segment that is primarily that brand. In fiscal 2025 the segment's net sales fell 9.4% to $358.6M and its adjusted EBITDA was –$0.3M, down from +$9.5M [A], on lower volumes, more trade promotion and higher raw-material cost. B&G took $34.8M of impairments on Green Giant in the fourth quarter alone, sold Le Sueur, moved Green Giant Canada to assets held for sale, and closed the sale of the Green Giant US frozen line on 2 March 2026 [A]. One of the best-known names in the freezer aisle earned nothing on a third of a billion dollars of sales, and its owner's response was to leave. When the category is shrinking under private label, a new processor is not competing for growth. It is competing to be the lowest-cost supplier of a retailer's own brand, against plants that are already depreciated. The small-batch jam, pickle or sauce maker sits outside that fight, but it is a brand selling through shelf and market stall, and is screened as Consumer Packaged Goods Brand Operation at 311.

Dairy processing in Canada is a good business to already be in. Saputo's Canada sector reported fiscal 2026 revenues of $5.423B, up 5.0% on higher volumes, with adjusted EBITDA of $697M — a 12.9% margin [A]. The same company's consolidated margin across all its sectors, most of them outside supply management, was 9.5% [A]: its regulated home market is where it earns the most. That is how the system is built. Under supply management a processor does not negotiate for milk. It buys from a provincial marketing board at an administered price, by end-use class, in a volume the board allocates, and imports that would undercut the result are held out by tariff-rate quotas. Everyone pays the same for the input and no one can be undersold from abroad, so margin goes to whoever has the plant scale, the allocation history and the retail listings. For an entrant that stability is the wall. A new plant needs a licence, an inspected facility and a milk allocation before it sells a litre, and the volume available to a newcomer is what the board's new-entrant or artisan programme releases, not what the business plan calls for. The small end is real — 153 of 434 establishments have fewer than ten employees, and Quebec alone has 150 plants [A], most of them fromageries — but those are farm-linked cheesemakers selling a regional premium, often processing their own quota milk. Reaching that position means entering dairy farming, where the quota is the capital. The pre-screen said supply-managed and plant-heavy; the first half is the one that binds.

There is a real and repeatedly documented shortage of provincially inspected slaughter capacity, particularly in British Columbia, and that shortage persists precisely because the barrier is high: inspected facility construction, effluent handling, HACCP plans and inspector availability all precede revenue. The demand signal is strong and the capital and compliance ramp is what keeps clearing the field.

The pre-screen paired small-scale processing with the fishing records, and the pairing is the problem. Commercial Fishing Licence & Quota (114) found that the right to catch has been capitalised into an asset its owners rent out. The plant is the next link along, and it owns neither end of the chain. Upstream, fish comes from licence and quota holders, many tied to a buyer by financing, or from integrated companies that land their own — when Premium Brands and a Mi'kmaq coalition took Clearwater private in 2021 at $8.25 a share, about $1B including debt [A], they were buying a company built on its own offshore licences. Downstream sit grocery and foodservice buyers. High Liner Foods, the listed Canadian processor with the strongest brand on that side, shows what the middle earns. In fiscal 2025 sales rose 7.1% to US$1,026.9M while volume rose 0.9% — the growth was price and mix, passed through from raw material and tariffs — and gross margin fell from 22.7% to 20.7%, adjusted EBITDA margin from 10.8% to 8.9% [A]. If a national brand cannot hold its margin when input costs move, a custom processor with no brand and no quota will not either. It is paid a processing fee set by whoever controls the fish. Nor is this a small-shop industry. Only 29% of the 372 Canadian plants have fewer than ten employees, while 40% have fifty or more [A] — seasonal lines with large crews, concentrated in Nova Scotia and Newfoundland and Labrador, where a processing licence is itself a provincial decision. The value in seafood sits in the quota and on the shelf, and the plant in between rents its margin from both.

Retail price resistance on a staple sets the ceiling and ingredient and labour costs set the floor, and the gap has narrowed every year. Wholesale volume looks like the answer and brings a grocery buyer who dictates price and delivery windows. The successful version is a café with an oven, which is a different business with hospitality's cost structure.

This is where small food makers actually file: 1,717 Canadian establishments, 61% of them with fewer than ten employees [A], across snack foods, coffee and tea, flavouring syrups, seasonings and dressings, and an "all other" remainder. It is a residual, and this screen examined one niche properly — coffee roasting, the most commonly entered — and did not examine the rest. Roasting is easy to start and carries a good margin. Farmer Bros., a US roaster founded in 1912, reported a fiscal 2025 gross margin of 43.5% on net sales of $342.3M [A]. It also reported operating expenses of $150.4M against gross profit of $148.9M, and a net loss of $14.5M [A], and in March 2026 agreed to be bought by Royal Cup for $1.29 a share in cash [A], a deal reported at about $32M in total [C] — under a tenth of one year's sales. The roasting made money. The expense was the route: branches, trucks, drivers and brewing equipment placed in thousands of restaurants and offices, which is how wholesale coffee is sold and serviced. A roaster with a century's head start on that network could not make it pay, and sold for a fraction of revenue to a competitor seeking route density. The entrant's version is the same problem smaller. Beans, a roaster and a bag are not scarce. The account list is, and it is built one café, one grocer and one office at a time against whoever already delivers there. The 311 brand record cuts on the cost of shelf access, and this one cuts a step earlier, on getting the product to a customer at all.

Category volumes are falling, the shakeout among small producers is well under way, and the unit economics only work through a taproom — which makes this a hospitality business with a brewery attached, carrying both cost structures. Excise, packaging and distribution access all favour the incumbents that already own shelf space.

The pre-screen said excise and licensing close this to new entrants. That is right, though the licence is not the part that does it. There are 16 tobacco manufacturing establishments in Canada [A]. Statistics Canada's own monthly table shows what they sell into: cigarette sales fell from 22.9 billion in 2010 to 12.3 billion in 2025, down 46% in fifteen years [A], and the decline has not levelled off — June 2024 sales were 16.8% below June 2023 [A]. The three large incumbents had creditor-protection plans approved by the court in 2025, built on a $32.5B global settlement, payable over time out of their operating profits [B]. An entrant carries none of that liability, which on paper is the only interesting opening in the industry: the same excise, the same retail price, and no settlement levy on the margin. It does not survive contact with the law. A federal excise manufacturing licence, stamping and bonding are the entry cost, and they are manageable. What cannot be overcome is that Canadian law requires plain, standardised packaging and prohibits nearly all promotion, so there is no lawful way to tell a smoker that a new brand exists or why to choose it. Brands that predate the rules keep their customers by habit, and a brand created after them cannot acquire any. The ten establishments with fewer than ten employees [A] are not evidence of an open door; this screen did not establish what they are, and at that size none of them is a challenger to the three incumbents. With sales down by nearly half in fifteen years there is also no growth to come in on.