The Real Cost of Adding a Second Supplier

Most explanations of the economics of dual sourcing stop at the point where it starts to matter. This one carries on.

Why the obvious answer is not free

Adding a second supplier is universally recommended and much less universally implemented, which suggests the costs are not trivial. Qualification requires engineering time, sample testing, process audits and often production trials before any commercial order is placed. In regulated sectors it may additionally require documentation submitted to an authority and a waiting period before approval.

Those costs land immediately while the benefit is contingent on a disruption that may never happen during the product's life. That timing mismatch, not complacency, is the main reason single sourcing survives every resilience review.

The scale you give up

Splitting volume between two suppliers gives each less scale, and unit prices reflect scale in almost every manufacturing process. Tooling must often be duplicated, and tooling is a capital cost that produces no additional output when idle.

Setup and changeover costs are incurred twice, and learning effects accumulate more slowly at each site than at one. The premium is genuine and measurable, which is precisely why it is easy to cut when budgets tighten. Framing the premium as an insurance payment rather than a purchasing failure changes how it survives budget review.

Keeping the second source alive

A supplier qualified once and never ordered from is not a second source; it is a document describing a former capability. Processes drift, staff change, tooling is repurposed and materials are requalified, so dormant approval decays quietly. Meaningful dual sourcing means placing a real share of volume with the alternate, which is where the scale cost is actually incurred.

Some firms rotate volume periodically instead, keeping both lines current without splitting every order. Whatever the pattern, the second source must be producing occasionally to be there when it is needed.

Correlated sources are not two sources

Two suppliers using the same sub-component, the same specialised process or the same raw material source fail together. Geographic separation addresses some correlations, such as regional infrastructure or weather, but not shared upstream dependence.

Verifying independence means asking each supplier about its own critical inputs, which returns to the visibility problem. A dual-source arrangement that has never been checked for correlation may deliver no protection at all.

The question to ask is not how many suppliers you have but how many independent ways the part can arrive.

Alternatives when a second source is impractical

Where technical scarcity is real, buffer inventory buys time and is often far cheaper than a qualification programme. Contractual capacity reservations and priority terms improve position during allocation without duplicating production. Design changes that allow an alternative specification, qualified in advance but not routinely ordered, sit between the two approaches.

Some firms fund capacity expansion at their sole supplier directly, which addresses volume risk while leaving site risk untouched. Each option manages a different failure mode, and naming which failure you are protecting against prevents spending on the wrong one.

Deciding where to spend

The sensible approach applies the full programme only to parts where absence stops production or where a single site is exposed. For the rest, cheaper measures such as modest buffers and standardised specifications carry most of the benefit. Scoring parts on consequence and on concentration produces a short list that is usually far smaller than the catalogue.

At port, reviewing that list annually against design changes keeps it current without a continuous programme. Resilience spending, like any other, works better concentrated where it matters than spread thinly everywhere.

The takeaway

Count independent routes to the part, not supplier names on a purchase order.

Capacity takes a decade to build and one quarter to look like a mistake.