Vertical integration is usually argued on control and quality. The decisive variable is simpler than either: it is whether the company can keep the new capacity busy.
Buying is variable and building is fixed
Paying a supplier converts a requirement into a per-unit cost that rises and falls with demand. In a slow quarter the cost falls with it, automatically.
Bringing the same activity in-house creates equipment, premises and salaries that continue regardless of how much work arrives. The cost per unit now depends on volume.
That single change explains most integration outcomes. Above a certain volume the in-house version is cheaper, and below it the arithmetic reverses sharply.
The supplier is running a portfolio and you are not
A specialist supplier serves many customers whose demand peaks at different times, so their capacity stays fuller than any single customer's would.
They also spread investment in tooling and expertise across that whole book, which lets them buy capability that no individual customer could justify alone.
Integrating removes both effects. The company inherits the capacity without inheriting the portfolio that kept it utilised or the scale that funded it.
Control is real but narrower than it sounds
Ownership does give genuine authority over sequencing, priority and quality standards, which matters most where a supplier's queue sits on a critical path.
What it does not give is competence. An in-house team starts behind a specialist and closes the gap only if the company invests continuously in it.
Where the activity is peripheral to what the company is actually good at, that investment competes for attention it will usually lose over a few years.
The commitment outlasts the conditions that justified it
Integration decisions are made when volume looks stable and rising. The assets acquired then persist through the periods when volume does neither.
Exiting is slow and expensive because it involves people and premises rather than a contract termination clause, so the position tends to be held longer than it should be.
This is why integration is more defensible for an activity central to the product than for one adopted mainly because a supplier relationship became irritating.
Partial integration keeps the option open
Some companies bring in a portion of demand and leave the remainder with suppliers, which caps fixed cost while preserving a fallback and a price benchmark.
The internal operation runs at high utilisation because it takes only the stable base of demand, and the supplier absorbs the peaks that would otherwise idle capacity.
The arrangement costs a little more per unit than full integration at high volume, and it removes most of the downside that makes the full version risky.