Ghost kitchens in Canada 2026: 8 reports, savings misread — ghost kitchen vs full-service Canada

Stop treating this as a race to “cheap.” Operators frame ghost kitchen vs full-service Canada like a coin flip on costs. It isn’t. The winning choice depends on your concept’s true margins, what travels well, and the rules in your province. If your menu is built for speed, low holding loss, and already generates consistent delivery demand from dine‑in, a delivery‑only build can work. If your brand wins because of the room, the wine list, or ritual moments at the table, a second full‑service site can beat a ghost kitchen on lifetime value even with higher capex. Wrong framing, wrong math, wrong outcome.

Here’s the punchline most owners miss: big “savings” from delivery‑only vanish once you price in aggregator commissions, heavier promotional spend, category misfit, and Canadian licensing constraints on alcohol and commissary use. Choose the model that fits your category and local constraints, not the one with the loudest marketing. The sections below give you the benchmarks, category fit, and a Canada‑specific checklist to make the decision on paper before you sign anything.

Related: How Ghost Kitchens Went From $1 Trillion Hype To A Struggling Business Model — CNBC

Reframe the question: concept‑fit, not model superiority

The right question isn’t “Which model is cheaper?” It’s “Which model lets this specific menu earn a defensible contribution margin in my city under my licensing and labor constraints?” Some vendors pitch ghost kitchens as a universal 50% cost cut. That framing ignores order‑level economics and Canadian rules that affect revenue mix, like provincial limits on alcohol delivery and the practical availability and cost of commercial kitchen leases in Toronto, Vancouver, and Montreal. Those limits ripple through your break‑even math and your achievable gross margin per order. See how the premise changes once you look at restaurant unit economics instead of headlines?

Across 83 Canadian restaurant SMBs analyzed via the Aurevon Intelligence Service, one recurring threat labeled Platform and Channel Disintermediation (avg impact 3.4) appears when delivery platforms sit between you and your guest. In the same dataset, Low Review Volume and Visibility (avg impact 4.4) repeatedly shows that venues with near‑zero review counts are invisible in local search, which pushes up paid acquisition costs for delivery‑first brands. Those two forces explain why operators who chase low capex end up spending back the “savings” in commissions and marketing. What does this mean for you? Pick the model that preserves margin power for your menu, not the one that merely minimizes build costs. For a structured way to map real rivals you’ll fight for visibility, see how to identify your real competitors.

Ghost‑kitchen unit economics: realistic benchmarks

In Canada, a cloud kitchen or delivery‑only restaurant launch is usually faster and lighter on capex. Typical startup all‑in, including deposits, smallwares, packaging inventory, and brand build, lands around $30,000–$80,000. Operators often use shared commissaries or licensed pods, sometimes from providers such as CloudKitchens or Reef Technology, add a single hood line, limited seating or none, and skip many front‑of‑house fixtures. On paper, that looks like instant savings. In practice, the order P&L depends on aggregator costs and your paid acquisition burn‑in.

Commission structures in Canada matter more than rent on day one. Uber Eats publicly lists a 30% “Marketplace” fee for delivery orders and 10% for pickup orders on its Canadian merchant pricing pages, which is applied to the pre‑tax order value. That immediately changes your gross margin math on every incremental delivery order. (merchants.ubereats.com) DoorDash’s Canadian merchant site shows plan‑based delivery commissions, often in the 15–30% band depending on plan, plus separate pickup rates, which similarly compress contribution margin unless average order values and add‑ons keep pace. (merchants.doordash.com)

Here is how the order math plays out. After food cost and packaging, many operators model a 60% gross margin for dine‑in. In a delivery‑first setup, once you net out 20–30% commissions, payment processing, promo subsidies, and higher packaging, typical realized gross margins land closer to 35–45%, often lower in the first six months while you buy discovery. That range lines up with what many Canadian operators report once they reconcile payouts against their menu price, fees, and promos. See the difference?

Break‑even timelines look shorter for ghosts (often 3–6 months) because you can start selling fast. The catch is that paid acquisition is front‑loaded and the platforms own early discovery. In our dataset noted earlier, “review volume gap” and “review volume leadership” showed up as paired threat/opportunity patterns. Median Google rating was 4.7 across 80 businesses, but the median review count was only 45 and the p10–p90 range ran from 3 to 1166 reviews. New delivery‑first brands with near‑zero reviews have to over‑spend on promos to earn conversion, which pushes cash payback out unless menu‑fit and pricing are tight. Building review velocity is not optional for visibility. For fast, practical ways to watch rivals’ ad pushes while you build review volume, see how to track competitor pricing and marketing.

Restaurants Canada’s recent indices also show soft, value‑seeking demand and fragile conditions, a backdrop that punishes thin contribution margins. Their late‑2025 outlook flagged minimal real sales growth into 2026 as operators face persistent cost pressures, which means ghost‑only models must assume cautious volume. (restaurantscanada.org) Statistics Canada reports that 2025 food‑service sales rose year over year, but volatility remained by segment; this complicates single‑channel bets when demand wobbles. (www150.statcan.gc.ca) And the Bank of Canada highlighted that food inflation dynamics in 2025 remained sticky for some inputs like meat, so assuming rapid margin relief is risky. (bankofcanada.ca)

So the risk is real. What can you do about it? Model each order like you would a SKU in retail: net revenue after platform fees, promo burn, processing, packaging, and refunds/adjustments. Then run a realistic ramp curve for the first 90 days with a separate marketing line. If your margin pool depends on alcohol, remember that alcohol delivery is narrower than on‑premise service and license‑specific in Canada; more on that below.

💡 Pro Tip
When you model a delivery‑only launch, add a 10–15% buffer for customer acquisition and menu testing in months 1–3. Operators routinely exceed initial marketing budgets while building review volume and tuning packaging.

Now, put the numbers side by side.

Metric	Ghost Kitchen (typical)	Full‑Service (typical)	Notes / What operators often miss
Startup capex	$30k–$80k	$300k–$2M	Ghosts skip FOH build, but deposits, hoods, and smallwares add up; full‑service pays for experience and capacity.
Time to open	4–12 weeks	6–12 months	Permitting and landlord work letters still apply to ghosts; supply of compliant kitchens varies by city.
Gross margin after fees	35–45%	55–65%	Delivery commissions, promos, and packaging compress margins; dine‑in keeps more dollar‑margin and adds alcohol. (merchants.ubereats.com)
Break‑even timeline	3–6 months	12–24 months	Ghosts launch faster, but early promo spend can push out cash payback; full‑service builds loyalty and check growth. (restaurantscanada.org)
Alcohol revenue	Limited, license‑specific	Core profit lever	Provinces restrict off‑premise alcohol; on‑premise programs drive mix and margin. (agco.ca)
Discovery channel	Aggregators + ads	Street presence + owned channels	Low review counts throttle aggregator conversion; FOH creates repeatable experiences that compound reviews.

Want a sharper handle on who you’re actually battling for discovery during ramp? Use a quick competitor SWOT to surface local leaders whose review volume and pricing set guest expectations.

Full‑service unit economics and the value of owning the customer

A full‑service second site asks for real capital and patience: $300,000 to $2 million in build, equipment, and working capital, and a 12–24 month runway to break even. That capital buys levers delivery‑only can’t match. You own the guest experience, from pacing to plating, which directly drives average check and review velocity. You also own the most profitable part of the P&L: alcohol mix. In Quebec and Ontario, delivery is allowed only under license‑specific conditions and typically must accompany food, with added restrictions on spirits, formats, and who can deliver. That pushes more margin to dine‑in programs where you set the mix and the moment. (racj.gouv.qc.ca)

Costs shift in a way that matters for lifetime value. A ghost kitchen spends on aggregator visibility and promo slots to acquire a customer it doesn’t fully own. A full‑service room spends on service design, sound, glassware, and energy that yield repeat visits without a toll‑booth in the middle. In a fragile demand year, that compounding loyalty turns into an insurance policy. Bank of Canada analyses in 2025–2026 repeatedly flagged sticky cost pressures around food away from home, which means units that control mix and experience have more pricing power than those arbitraged by platform discovery. (bankofcanada.ca)

Here’s a compact comparison.

Metric	Ghost Kitchen (typical)	Full‑Service (typical)	Notes / Where full‑service wins
Customer data	Mostly platform‑controlled	Owned	Dine‑in yields emails, birthdays, preferences; delivery marketplaces throttle access.
Contribution per order	Lower after fees	Higher with alcohol	Beer, wine, and spirits lift gross profit dollars on each check.
Marketing	Ongoing promos to maintain rank	Experience and community	Room, events, and staff create organic reviews and PR.
Capacity utilization	Flexible by daypart	Fixed seating, but upsellable	Slow periods can host events, tastings, or classes that generate reviews and cash.

Concept‑fit checklist: which menus travel (findings from 8 reports)

The biggest error isn’t cost modeling. It’s menu mismatch. In eight reports from the dataset referenced earlier that focused on delivery‑native entrants, the categories that sustained healthy net revenue per order shared three traits: fast assembly, low hold‑time decay, and packaging that preserves texture. Categories that bled margin did the opposite.

What travels well most of the time? Bowls, burritos and wraps, pizzas under 16 inches with bready crusts, fried chicken sandwiches in breathable packaging, sauced proteins over rice or grains, and mid‑price comfort food built for heat retention. These items tolerate a 20–30 minute delivery window with limited sensory loss when packed correctly. What fails? Fine‑dining plates that rely on plating theatrics, delicate fish cooked to a narrow band, multi‑item share plates where heat loss and sogginess compound, and desserts that depend on glassware or last‑second assembly.

Operationally, winners share short critical path timing and parts that can be held warm without collapse. Losers demand à‑la‑minute firing, last‑second garnish, or air‑sensitive textures like shattering crisp skins. In that same dataset, Brand Confusion and Local Identity Risks and Review Volume Gap were frequent themes for delivery‑only multi‑brand stacks run from one address; customers get confused, reviews split across brand names, and discovery lags. It is like sending two salespeople to pitch the same client at the same time. Messages collide, and nobody closes. If you pursue a multi‑brand strategy for concept expansion inside one facility, keep brand architecture simple and packaging distinct.

Use this quick‑reference to pressure‑test your menu:

Menu Category	Delivery Suitability	Key Constraints	Recommended Model
Bowls, burritos, wraps	High	Rice hydration, tortilla steam; vented packaging	Ghost or Hybrid
Pizza (12–16")	High	Box ventilation, cheese oiling, hold time	Ghost or Hybrid
Fried chicken sandwiches	Medium–High	Condensation, bun texture; use breathable wraps	Ghost or Hybrid
Noodle soups	Medium	Broth separation, spillage; send components	Hybrid (ghost risk if assembly slow)
BBQ and sauced proteins	Medium–High	Sauce separation, reheating fatigue	Ghost or Hybrid
Sushi/nigiri	Medium	Temperature control, rice texture	Hybrid (local delivery radius only)
Fine‑dining mains	Low	Plating, narrow doneness bands	Full‑Service
Share plates/tastings	Low	Sequencing, temperature decay	Full‑Service
Delicate plated desserts	Low	Assembly, melting, structure	Full‑Service

Before: choosing a format first and forcing your menu into it. After: scoring each core item on demand x margin x delivery‑compatibility, then letting the score pick the model. If you expand a delivery‑fit menu, a ghost or hybrid can shine. If your menu sells theater and timing, the room is the product.

Two practical guardrails from our eight focused reports: first, price point ceilings matter. Delivery‑only checks that exceed your local casual‑premium ceiling struggle unless bundle value is obvious; customers will trade down under cost pressure. Second, assembly time is destiny. If a hot dish spends more than 90 seconds between bag‑seal and handoff during a rush, your texture decay will show up in reviews within hours, which raises the promo tax you’ll pay to keep conversion steady. Want a competitive sanity check before you commit? Start by mapping the actual rival set you’ll share search real estate with using how to identify your real competitors, then stress‑test strengths and blind spots with a quick competitor SWOT.

Canadian constraints and the practical hybrid option

Rules shape revenue. In Ontario, the Food Premises Regulation (O. Reg. 493/17) defines standards for commercial kitchens and sources of inspected food, requirements that apply whether you operate dine‑in or rent a commissary. Any shared or commissary arrangement you use must still meet those public‑health requirements and be inspected as a food premise. (ontario.ca) British Columbia’s Food Premises Regulation also requires approved food safety and sanitation plans, which becomes a gating item when multiple brands share lines and storage. (bclaws.gov.bc.ca) For alcohol, provinces differ. Ontario and Quebec allow takeout and delivery of alcohol only under specific license conditions, often with food and with format limits; British Columbia requires Serving It Right certification for staff who deliver liquor for employers. You’ll need to verify license class, delivery rules, and who can legally deliver in your province before assuming alcohol adds to delivery margin. (agco.ca)

Leasing markets are another Canadian‑specific constraint. In late‑2025 CBRE reported low retail vacancy, rising rents for fixtured space, and operators gravitating to second‑generation units to counter construction costs, a pattern visible across Toronto, Vancouver, and Montreal. That supply pressure pushes some teams toward ghosts when full‑service rents feel out of reach, but it also means well‑located second‑gen rooms can open faster than you think with less capex. (cbre.com) Statistics Canada’s 2025 sales data confirm there’s still a market for food away from home, but the mix is jittery; locking into a long lease without a clear category advantage is a bigger bet than it looks. (www150.statcan.gc.ca)

When does a hybrid work? If your existing kitchen has slack capacity during off‑hours, a hybrid lets you test a delivery‑native sub‑brand without new rent. It acts like a wind tunnel for menu fit and packaging. It fails when labor or equipment conflicts degrade service for the core brand, or when your liquor license and food‑safety plans don’t cover the second brand’s SKUs and delivery process. The “Service Capacity and Operational Constraints” theme from the earlier dataset shows up here: restricted service hours, parking, or line bottlenecks cap revenue when you try to run two concepts through one set of hands. Treat hybrid as a test bed with strict guardrails.

Two “do this today” actions before you commit: first, pull the last 90 days of delivery payout statements and compute true net revenue per order after commissions, promos, processing, packaging, and refunds. If your net dollar margin per order is under $6 on average without alcohol, a ghost‑only build will be tight. Second, ask a broker for second‑generation options in your target corridor and cross‑check landlord work letters and venting language; CBRE’s 2025 surveys show second‑gen space is how many operators are beating cost inflation on buildouts. (cbre.com) If you do move toward delivery‑first, set up basic competitor monitoring so you aren’t flying blind on promos and price experimentation. A lightweight start is here: how to track competitor pricing and marketing.

Common questions about choosing a ghost kitchen or full‑service location

Are ghost kitchens profitable in Canada, and will they always be cheaper to open and run?

Not always. Profitability depends on menu fit, average order value, and your local fee structure. While initial capex can be much lower, recurring aggregator fees on each order, higher early‑stage marketing to buy visibility, and limited alcohol revenue often narrow or erase the expected savings. Accurate comparison requires modeling expected paid and organic order volume, average order value after platform fees, and marketing spend over the first 12 months. Uber Eats’ Canadian merchant pricing shows 30% for delivery and 10% for pickup plans, and DoorDash’s Canadian plans advertise delivery commissions in the teens‑to‑30% range, which materially change per‑order economics. (merchants.ubereats.com) (merchants.doordash.com)

Which menu categories from the eight reports should I avoid putting into a ghost kitchen?

Avoid delicate, high‑touch or theatrical categories: fine‑dining mains, share plates that rely on sequencing, complex plated desserts, and items that require rapid à‑la‑minute assembly or very narrow doneness windows. In the focused reports from the dataset referenced earlier, categories with low hold‑time tolerance and high packaging sensitivity performed worst on review velocity and promo spend required to maintain conversion. If your core items fall in those buckets, let a room carry the experience and keep delivery as a convenience channel.

What permits does a ghost kitchen need in Canada, and how do provincial rules change the decision?

They change it at the margins and sometimes at the core. Provincial health codes govern how you can use commissary or shared kitchens and what approvals your operation needs; for example, Ontario’s Food Premises Regulation and B.C.’s Food Premises Regulation set requirements for inspected food sources and sanitation plans. Local business licensing, fire and building inspections, and zoning approvals may also apply depending on municipality. Alcohol licensing is license‑ and province‑specific: Ontario and Quebec allow alcohol takeout/delivery only under defined conditions and often only with food, while B.C. requires Serving It Right certification for employees delivering liquor. These rules affect your ability to lift contribution margin with alcohol and how fast you can go live in a shared facility. (ontario.ca) (bclaws.gov.bc.ca) (agco.ca)

Can you run a ghost kitchen out of an existing restaurant, and when is the hybrid model the best first step?

Yes, if your existing kitchen has unused capacity, proper approvals, and a menu that won’t compromise service for the core brand. A hybrid works when your delivery menu can be operationally decoupled from dine‑in service and your food‑safety and liquor plans cover the second brand’s process. It’s a lower‑cost way to test demand, packaging, and pricing without a separate lease. It is risky if staff scheduling, storage, or hood time will bottleneck your main service, or if your licensing doesn’t permit the added activities. Think of hybrid as a structured experiment with a fixed end date and kill‑criteria. To keep your experiment honest, track nearby rivals’ promos and reviews with a simple, weekly ritual based on this field guide: how to identify your real competitors.

Five questions to answer before you sign anything:
1) What is my net dollar margin per delivery order after all fees and promo subsidies at month 1, month 3, and month 6?
2) Which 5 menu items drive 70% of my contribution margin, and how do they hold up at 20 and 40 minutes in the bag?
3) Under my province’s rules, what alcohol can I legally sell via delivery, who can deliver it, and what documentation is required? (agco.ca)
4) In my target corridor, are second‑generation sites available at rents that let me open faster with less capex? (cbre.ca)
5) Do I have a 90‑day plan to generate review volume leadership, or will I rent visibility from platforms indefinitely?

Final step today: open your latest 30 delivery payouts, calculate the realized take‑rate (commissions plus promo subsidies plus processing as a percent of pre‑tax sales), and run the result through your top five menu items. If your realized take‑rate exceeds 28% and your top items don’t travel into four‑plus‑star reviews, a second room might be the better bet this cycle.

To go deeper on decision math for this choice, bookmark how to identify your real competitors, pair it with this competitor SWOT template, and audit promo pressure using tracking competitor pricing and marketing. Build your model, then pick the format your numbers can actually carry.

Aurevon’s Ecosystem Dynamics Report distills category fit, local constraints, and competitor patterns into a decision‑ready model for Canadian operators weighing delivery‑only versus a second room. If you want one concise brief that reflects Canadian rules and unit economics without hiring consultants, start here: https://aurevon.ca/.