Every growing manufacturer eventually faces this question, usually when a regional customer base reaches the point where shipping everything from the plant starts feeling inefficient. The answer is genuinely situational, and getting it wrong in either direction is expensive.

The Case For Direct

Shipping from plant to customer means one movement, one handling, and no intermediate inventory. You avoid warehouse cost, avoid the capital tied up in distributed stock, and avoid the risk of stocking the wrong thing in the wrong place.

It also keeps things simple. One stock pool, one set of records, no allocation decisions, no transfers.

Direct works well with full or near-full loads, customers who accept your production lead time, and a wide product range where forward-stocking everything would be prohibitive.

The Case For Hubs

Hubs exist to solve two problems: delivery speed and freight efficiency on small orders.

Speed, because stock near the customer ships same-day or next-day in a way plant shipping can’t match if the plant is a week away.

Freight, because consolidating many small customer orders into full-load movements to a regional hub, then distributing locally, costs substantially less than shipping each order individually across the same distance. The economics improve as order size falls and customer density rises.

The Variables That Decide It

Order size relative to vehicle capacity. Large orders ship efficiently direct. Small orders don’t, and that’s what makes hubs pay.

Customer density in the region. A cluster of customers justifies a hub; scattered ones don’t, because you lose the local distribution efficiency that the model depends on.

Required delivery speed versus plant lead time. If customers accept a week and you can deliver in five days direct, a hub buys you nothing operationally.

Product range breadth. Stocking a wide range at multiple hubs multiplies inventory fast. Narrow ranges hub well; long tails don’t.

Demand predictability by region. Hub stocking is a forecasting bet at a more granular level than plant stocking, and forecast error is worse at regional level than national.

The Middle Options

It’s rarely all or nothing. Hub the fast-moving minority of your range and ship the tail direct — this captures most of the freight and service benefit with a fraction of the inventory.

Cross-docking gives some hub benefits without holding stock: consolidated inbound movements broken down and reshipped without storage. Useful where the freight consolidation matters more than delivery speed.

Third-party warehousing lets you test a region without committing capital, which is usually the sensible way to find out whether the economics work before building anything.

Modelling It

Run the comparison on your actual order history rather than on averages. Total cost including freight, warehousing, inventory carrying, and handling, against service outcomes. Order-level data usually shows the answer differs by product and by region, which is why single-model networks tend to leave money on the table.

Include the transition cost in any comparison. Opening a hub means building stock before it earns anything, and closing one means redistributing inventory that may not be needed elsewhere. Both are one-off costs that can swamp a marginal case.

ticktick.ai models network configurations against real order history, comparing total cost and service by product and region.