There is no single best route, there is a route that fits your volume and your use case. Below about ten units a local dealer is usually faster and cheaper once freight is counted. From ten to fifty units, factory procurement with first-trip delivery becomes competitive. Above that, and on a recurring schedule, direct factory relationships start to pay for the overhead they require.
Buy locally. A dealer in your country holds stock, delivers this week, and the margin they take is smaller than the freight you would pay to bring one or two boxes across an ocean. We would rather tell you that than quote you something you should not buy.
The exception is a specification your local market does not stock at all, which in practice means Door Units, unusual heights or a specific color and logo.
This is the range where factory procurement starts to work. Factories will engage at this volume through a partner who already holds the relationship, the per-unit freight cost falls, and the specification becomes yours to set rather than whatever the depot happens to hold.
The route only works if the freight problem is solved. A first trip, where the container carries paying cargo toward your destination, is what keeps the landed number competitive at this volume.
If you buy every month, the pricing model matters as much as the price. A fixed fee per unit that you agree in advance is easier to plan against than a percentage of order value, and it makes the quotation legible: container price, fee, logistics, three lines. The volume program page describes how that works.
At this volume it is also worth mapping both sources every month. Depot inventory and factory production move independently, and the cheaper one is not always the same one.
Open-side units, 20FT High Cubes and other thin-production types do not accumulate in depot surplus, because carriers do not reposition enough of them to create one. For these, factory production is not the cheaper route, it is the only route with a supply behind it. Guides 11 and 13 cover why.