Warehouse network design is the process of deciding how many distribution nodes a supply chain needs and where to put them, based on where demand sits and how fast it has to be served. Every node you add buys service level and parcel savings on one side of the ledger, and costs fixed facility overhead plus additional safety stock on the other, so the real question is never whether to add a warehouse, it’s whether that node pays for itself. This guide works through that math for one, two, and three-node Canadian networks using population distribution and real ground transit times drawn from a network of 150+ pre-vetted facilities across 25+ markets.
The Population Math That Drives Everything
Canadian network design looks different from US network design for one structural reason: population here is extremely concentrated. Nearly everyone lives inside a narrow band running along the US border, and within that band, two provinces do most of the work. Ontario holds roughly 39 percent of the national population and Quebec adds another 22 percent, meaning about 61 percent of Canadians live in two adjacent provinces before you’ve placed a single warehouse anywhere else.
| Province / Region | Share of Canadian Population |
|---|---|
| Ontario | 39% |
| Quebec | 22% |
| British Columbia | 14% |
| Alberta | 12% |
| Manitoba | 3.6% |
| Saskatchewan | 3% |
| Atlantic provinces | 6.6% |
(Source: Statistics Canada population estimates)
The consequence of this concentration is that the first node in almost every Canadian network is eastern, usually Ontario. There isn’t much of a real debate on node one for a general consumer network: no other single location reaches as much of the country as fast. The actual strategic decision, the one that determines your total network cost and your service map, is what node two and node three should be and whether you need them at all.
One Node: Which City
For most Canadian networks, Toronto is the default single-node location. It sits inside the country’s densest population corridor, it has the deepest carrier and highway infrastructure in Canada, and it reaches Montreal in about one day and the Maritimes in one to two days by ground. A Toronto node alone covers roughly half the country within two-day ground service, simply because of how much of the population sits within Ontario and Quebec.
The exception is a network with demand weighted west. If your customer base skews toward Alberta, Saskatchewan, British Columbia, and Manitoba, a single node in Calgary covers that region far better than Toronto does, reaching Edmonton in about 3 hours, Saskatoon in about 6, Regina in about 7, and Winnipeg in about 13. Calgary serves fewer total Canadians than Toronto, but it serves the ones it does serve faster, and for a west-heavy brand that’s the number that matters.
The other exception is Quebec-weighted demand. A Montreal node reaches Ottawa in about 2 hours, Quebec City in about 2.5, Toronto in about 5 (same day), and the Maritimes in one to two days. For a brand whose revenue concentration sits in Quebec, running Montreal as the single node avoids retrofitting French-language compliance onto an Ontario operation later.
| Single node | Reaches next-day | Reaches 2-day | Weak spot |
|---|---|---|---|
| Toronto | Montreal, most of Ontario | Quebec, Maritimes | British Columbia (4-6 days by ground) |
| Calgary | Edmonton | Saskatoon, Regina, Winnipeg | Quebec and Atlantic Canada, multi-day ground |
| Montreal | Ottawa, Quebec City, Toronto | Maritimes | Western Canada, multi-day ground |
| Vancouver | Lower Mainland and nearby BC | Rest of British Columbia | Rest of Canada (4-6 days to Toronto) |
Two Nodes: The Real Decision
Once a brand outgrows a single node, three pairings come up more than any others, and each one unlocks a different kind of coverage.
Toronto + Vancouver is the default national pair. It adds British Columbia’s 14 percent of the population as a fast-served region and gives import-heavy brands a port-fed inbound option on the west coast. The tradeoff is that Toronto and Vancouver are Canada’s two highest-cost warehouse markets, so this pair maximizes coverage at the top of the cost curve.
Toronto + Calgary is the cheaper second node. It doesn’t reach quite as far into British Columbia as a Vancouver node would, but it covers Alberta, Saskatchewan, and Manitoba inside two days and does it at a meaningfully lower facility and labour cost. For a domestic-inbound brand without a strong reason to sit at the port, this pair usually wins on total cost.
Toronto + Montreal isn’t really a coverage play at all, since both cities already sit inside each other’s two-day radius. It’s a compliance and depth play: Montreal gives Quebec native French-language handling and a second node close enough to Toronto that it also functions as redundancy, without doing much to extend national reach.
The way to actually decide between these pairs is to map the last twelve months of orders by postal region, model the parcel zone cost from each candidate node against your actual order geography, and compare that savings against the node’s fixed cost plus the incremental safety stock it adds. The pairing that looks best on a map is not always the one that wins on total landed cost once inventory carrying cost is in the model.
The Cost of Every Node You Add
Every node added to a network increases total inventory, and it does so faster than most people expect. Aggregate safety stock roughly follows the square root law of inventory: it scales with the square root of the number of stocking locations, not linearly with node count. Going from one node to two increases safety stock by about 41 percent for the same service level. Going from one node to three increases it by about 73 percent. That’s before a single unit has shipped faster, it’s simply the cost of holding buffer stock in more places at once.
Safety stock isn’t the only per-node cost. Each additional facility typically carries its own monthly minimums, its own account management fee, its own WMS setup and integration cost, and its own share of inbound freight that now has to be split and routed across multiple destinations instead of one. There’s also the operational overhead of managing allocation logic between nodes: deciding which orders route where, keeping inventory balanced so one node doesn’t stock out while another sits overstocked, and reconciling two or three operations instead of one.
Modelling a network change? Get market-benchmarked rates for every city you are considering, from pre-vetted operators, normalized to one specification.
See enterprise programs →The honest way to evaluate a new node is to add up parcel savings plus the incremental revenue from improved service, and compare that against carrying cost plus fixed cost. Run that math seriously and a pattern shows up in most networks: the second node tends to get added too early, before volume or margin can absorb the safety stock increase, and the third node tends to get added too late, well after the transit-time complaints from underserved regions have already cost the brand customers.
Three Nodes and Beyond
The most common three-node Canadian network is Toronto, Vancouver, and Calgary, which covers the country’s population and its major cost and speed tradeoffs in one setup. Beyond three nodes, Canadian networks generally stop adding capacity for general coverage and start adding it for a specific, defensible reason.
A Montreal node past that point is usually about Quebec compliance and next-day service depth, not incremental reach. A Halifax node is usually about Atlantic service, turning a three to five day delivery promise into a next-day one for Nova Scotia, New Brunswick, and Prince Edward Island; see 3PL fulfillment in Halifax for what that market actually offers. A Winnipeg node is usually about cost, not speed: Winnipeg reaches both Toronto and Calgary in about two days and virtually all of Canada in two to four, which makes it a strong low-cost single-node option for B2B replenishment programs where parcel speed doesn’t drive the sale the way it does in direct-to-consumer.
The Hybrid Most Brands Actually Run
Most enterprise networks that get this right don’t commit to a fixed node count up front. They run a dedicated or committed core node sized to steady-state volume, and they pair it with flexible overflow storage for peak season and for testing new regions before a permanent commitment is justified. That structure lets you validate a market’s actual order volume and transit-time sensitivity before signing a multi-year lease on a node that may or may not earn its keep. It’s the same dedicated-versus-flexible tradeoff covered in our guide to contract logistics in Canada, applied at the network level instead of the single-facility level.
Where to Start
Before any of this coverage math means anything for your business, pull the last twelve months of orders by postal region. Everything else in this guide, which pair to run, whether a third node pays for itself, whether Calgary or Vancouver is the better western node, follows directly from that dataset rather than from population share alone. Once you know where your orders actually originate, the Canadian Warehouse Market Report 2026 gives you rate bands by market to sanity-check facility cost, and the warehouse cost calculator lets you model the fixed and carrying cost of each candidate node before you commit to it.