Make vs Buy: The Manufacturing Decision Framework (2026)

An operations manager at a 45-person job shop is staring at two numbers: $8.40 per unit from a supplier in Monterrey, or an internal cost sheet that says $10.60 to run the same bracket on the shop floor. The plant manager wants to outsource. The owner, who bought the CNC mill that would make the part, wants to keep the work. Both are arguing from numbers that are wrong, because the $10.60 includes overhead that exists either way, and the $8.40 excludes freight, inspection, and the cost of carrying eight weeks of inventory. The make vs buy decision is one of the highest-leverage choices in manufacturing, and most shops make it with arithmetic that would not survive an accounting class. This guide builds the framework properly: full cost comparison, capacity opportunity cost, quality, supply risk, switching costs, and a worked example with real numbers.

The Full Cost Comparison Most Spreadsheets Get Wrong

The structural error in most make or buy analysis is asymmetry. The internal cost arrives fully burdened: direct labor, machine time, and a fixed overhead allocation covering rent, supervision, and depreciation. The supplier quote arrives naked: a piece price. Fixing the comparison means stripping the make side down to costs that actually change with the decision, and building the buy side up to costs that actually arrive with the part.

On the make side, the relevant costs are marginal: material, direct labor, machine operating cost and energy, consumable tooling, scrap, plus any new fixturing or equipment the part requires, amortized over realistic volume. Allocated fixed overhead stays out, because the rent does not go down when the part leaves. On the buy side, the landed cost stacks up fast: piece price, freight, duties and tariffs where applicable, incoming inspection, purchasing and expediting administration, and inventory carrying cost, since suppliers quote against order quantities that put weeks of stock on your shelf. A common rule of thumb in sourcing organizations is that landed cost runs 5 to 15% above the quoted piece price for domestic suppliers and considerably more for offshore sources once freight volatility and tariffs enter, which is exactly the trade space the Reshore, Nearshore, or Offshore Decision tool walks through input by input.

Capacity Opportunity Cost: The Invisible Line Item

Here is the question that decides more make vs buy cases than any other: what would the machine hours do otherwise? If the answer is "sit idle," then in-house production only needs to beat the supplier on marginal cost, and it usually does. If the answer is "run quoted work at $45 per hour contribution margin," then every hour the bracket consumes costs $45 on top of its marginal cost, and the supplier usually wins.

This is why utilization data has to enter the decision. A shop running 60% utilization has idle hours that make in-house work nearly free at the margin; a shop turning away work at 95% utilization should be buying everything that is not core. Most shops do not know their true utilization with confidence, which is where a Production Efficiency Grader earns its place in the sourcing conversation: bottleneck management and changeover discipline determine how many sellable hours the floor actually has, and therefore what those hours are worth when a make vs buy candidate wants them.

A Worked Example: 20,000 Brackets a Year

Take the bracket from the opening scene at 20,000 units per year. Assume the following inputs, which are illustrative but realistic for a US job shop:

At idle capacity, making wins by $0.10 per unit, $2,000 per year, a coin flip that the fixturing risk probably tips toward buy. Now add the capacity scenario: the bracket consumes 1,100 machine hours per year, and the shop is quoting work it cannot accept that would contribute $45 per hour. The opportunity cost is $49,500 per year, or about $2.48 per bracket. The true make cost becomes $11.58 against a $9.20 landed buy cost, and the decision is not close. Same part, same supplier, same spreadsheet, opposite answer, all driven by what the hours are worth.

Quality Control and the Cost of Poor Quality

Cost parity means nothing if the parts do not conform. The American Society for Quality (ASQ) places the cost of poor quality at 10 to 15% of revenue for many organizations, counting scrap, rework, containment, warranty, and the engineering hours that chase escapes. The make vs buy implication runs both directions. Keeping a part in-house keeps the process capability conversation inside your own walls, where a drifting dimension shows up at the machine, not at final assembly. Buying a part puts a supplier's process between you and the defect, and the detection delay is the expensive part: a nonconforming lot discovered at assembly has already multiplied its containment cost.

The practical screen: outsource candidates should be parts where conformance is verifiable at receiving with reasonable effort, where the controlling dimensions have generous capability margin, or where the supplier demonstrably owns better process technology than you do. Parts with tight tolerances on functionally critical features, or failure modes that hide until the field, deserve a strong bias toward make even at a cost penalty of several percent. And when a supplier does win the work, the quality cost does not drop to zero; it converts into supplier development, audits, and incoming inspection, all of which belong in the buy column at their honest annual cost rather than disappearing from the model the day the PO is cut.

Supply Risk Prices Itself Into the Decision

Every buy decision is a bet on a supply chain. McKinsey Global Institute research found that companies experience supply chain disruptions lasting a month or longer roughly every 3.7 years, which means a sourced part will, at some point during its production life, stop arriving. The relevant questions are concentration and recovery: is this a sole source, how deep is the safety stock, how long would requalifying an alternate take, and what does a stopped line cost per day while you find out?

None of this argues against buying; it argues for pricing the risk. Dual sourcing, safety stock, and contractual capacity commitments all cost money, and that money belongs in the buy column of the comparison. A Supply Chain Resilience Scorecard puts structure on the question, scoring supplier diversification, visibility, inventory buffers, and lead-time stability before the sourcing committee treats the quoted price as the whole story. The same logic extends downstream: shops weighing whether fulfillment and warehousing belong in-house can run the 3PL or In-House Logistics Decision through the identical marginal-cost-versus-landed-cost lens.

Switching Costs Cut Both Ways

The decision is reversible in principle and sticky in practice. Outsourcing a part starts a clock on your own capability: setup sheets go stale, the operator who knew the quirks moves on, and the fixtures get pushed to the back of the tool crib. Two years later, "bring it back in-house" is a re-learning project, not a scheduling change. In the other direction, leaving a supplier means requalification at the next source: in automotive and aerospace work, a new supplier typically needs full PPAP or first-article approval, a process measured in months, plus tooling transfer negotiations and dual-running stock through the transition. The Reshoring Initiative has tracked record levels of reshoring announcements in recent years, and the companies doing it well budget the switching costs explicitly rather than discovering them mid-transfer. Whichever way your make vs buy analysis points, add a line for the cost of changing your mind later, because volumes, tariffs, and labor markets will eventually force the question again.

A Five-Question Checklist Before You Decide

Run every make vs buy candidate through five questions. One: is the comparison symmetric, marginal make cost against fully landed buy cost? Two: what would the machine hours earn otherwise, and is that number in the model? Three: where would a defect be caught, at the machine, at receiving, or at the customer? Four: what happens in month two of a supplier disruption? Five: what does reversing this decision cost in three years? A decision that survives all five is one you can defend to the owner, the plant manager, and the auditor. For contract manufacturers and industrial suppliers, these are also the questions your prospects are working through before they send an RFQ; the lead generation tools for manufacturers page shows how embedding sourcing and readiness assessments captures those buyers while the decision is still open.