The licence is a tax you stopped noticing 

Every large enterprise carries a handful of platforms that arrived as a solution and stayed as a condition. The licence renews, the number goes up, and nobody re-opens the decision because re-opening it would mean admitting how much has been built on top.

I have displaced several of these with products built in house. The savings were substantial, and the savings were not the point.

What you are actually buying

A commercial platform sells you capability today and control tomorrow. The capability is real. The control is the part that gets under-priced in the business case, because it is not a line item. It shows up later, as a series of small refusals.

• The integration you need is on their roadmap for next year, or it is not.
• The performance ceiling is theirs, and raising it is a commercial negotiation rather than an engineering task.
• Your differentiation has to be expressed within their extension model, which was designed for the average customer.
• The migration cost compounds quietly, because every year of adoption raises the price of leaving.

None of that is dishonest on the vendor's part. It is simply what a product built for many customers has to be. The question is whether the thing you are buying is genuinely undifferentiated for your business, because if it is not, you are outsourcing your roadmap to someone optimising for a different set of priorities.

When building is the right call

Build is not the default. It is the right call under a narrow set of conditions, and it is worth being honest about them:

• The capability is close to how you actually compete, not adjacent to it.
• You already employ people who understand the domain deeply, so you are not hiring into the unknown.
• The licence is large enough that the engineering investment is recovered inside a couple of years, not a decade.
• You are prepared to own it forever, including the unglamorous parts: on-call, upgrades, documentation, the second engineer who has to understand it.

That last condition is the one organisations skip. An in-house platform without a durable owner becomes a worse version of the problem you were escaping, and it becomes it quietly, over about three years.

The argument to make internally

Displacing a licence returns two things: margin, and the ability to decide what gets built next. Only one of them shows up in the first year.

Present it as an architecture decision that happens to have a financial consequence, not as a cost-cutting exercise that happens to involve engineering. The distinction matters, because cost-cutting exercises get handed to procurement and negotiated down to a discount, which leaves the roadmap exactly where it was.