A strategy offsite ends with a slide: "We are integrating our seat-mechanism supply." The room hears security and control. Six quarters later the line is built, the volume guarantee is signed, and the spot market that used to keep the price honest is gone. Nobody in that room was from procurement, and the cost of getting out was never on the slide.
Integration does not remove your dependence on a supplier. It moves that dependence onto your own balance sheet.
The word "integration" gets used for two opposite things in direct procurement, and the confusion is expensive. One is vertical integration — you bring the part or process in-house, a make-versus-buy call where you now own the asset, the labor, and the fixed cost. The other is supplier integration — you weave a key supplier deep into your planning, your systems, and your shared roadmaps. Both get called "integration," both get sold as "control," and only one of them is strategy. The rest is a liability wearing a strategy badge.
McKinsey's make-or-buy writing frames the whole trade in a single breath: integration is "of making fixed costs variable, or losing knowhow? Of gaining flexibility, or plunging into dependence?" That sentence is the argument. When you integrate, you are usually making fixed costs out of variable ones and trading flexibility for dependence. The only real question is whether you did it on purpose — and priced what you gave up.
The tell that it was not on purpose is who was in the room. Deloitte's Global CPO Survey finds that involvement in make-versus-buy decisions is the area where procurement's seat at the table has slipped the most — down about 40 percent in a single survey cycle. Integration is the single most expensive make-or-buy call a company makes, and it is the one procurement is least often asked to price. The capital gets committed by strategy or engineering, the option to walk disappears, and procurement is handed the bill.
What integration actually does to your cost structure — the part that lands on procurement:
1.It converts a variable cost into a fixed one. Buying lets you flex volume with demand. Making commits you to capacity you pay for whether or not the program sells. Procurement prices the buy every cycle; the make is a capital decision made once, by people who will not live with the utilization rate two years out.
2.It removes your switching option. A bought supplier you can re-source. An integrated one you cannot leave without writing off the asset. BCG's automotive battery work makes the mechanism concrete: OEMs that take ownership stakes or build cells in-house gain control but "can limit an OEM's ability to react quickly to technological advances," while the traditional buy model "reduces their up-front costs and the potential cost of switching to an alternative technology." Integration is a bet that today's choice stays right longer than the asset lasts — and assets last a long time.
3.It concentrates the risk you used to spread. This is the mirror image of consolidation: dropping suppliers removes the competition; absorbing them removes your escape. Either way, the risk moves from the market into the room you sit in, and procurement is the function left holding it.
Where the word is actually earned — genuine integration:
The label only fits when integration builds a capability you could not buy. Academic work on 251 manufacturers finds product codevelopment with suppliers directly improves product performance, and the authors urge managers to involve suppliers early in design (Lau, Tang & Yam, Journal of Product Innovation Management, 2010). Note what the research actually rewards: codevelopment, information sharing, early design involvement. Every one of those happens with a supplier you still buy from. None of them require owning the supplier. The gain comes from weaving the relationship, not from absorbing the balance sheet. That is integration as capability: shared roadmaps, shared technology, a supplier woven into how you win — not a supplier you swallowed to close a cost gap. The first is strategy. The second is a liability with a strategy badge pinned to it.
One question before you integrate:
Before any integration decision is signed, procurement should answer one question out loud: "If the program halves in two years, can we get out, and what does that cost?" If the honest answer is "we have not priced that," the deal is not integration. It is an unpriced liability that procurement will be asked to explain at the next renewal.