My impression is that one of the reasons is that Oracle has a full stack of applications, and once you have locked yourself in by using two or more applications from the stack, you can only see an "unnecessary" cost for moving to some other product, without gaining any functional benefits. And even if you manage to change one of the products, you will probably see little to no reducation in maintenance costs, because that's the way Oracle's SULS price model is constructed (for example the "SULS on all or nothing" clause).

The decision to move to another vendor is not done by technical experts alone. It is done by the company - which consists of the technical experts AND procurement AND legal AND business controllers. Each of these parties has to have a business case to move to another vendor - otherwise nothing will happen, unless something very disruptive happens.

We can only speculate what the triggers are for a software audit.

Some common causes are: when the customer stops paying maintenance, when there has been disagreements between the customer and the sales person about interpretation of license agreements, when the customer performs a larger hardware consolidation, or when the customer performs a larger migration. Oracle, as other software vendors, of course do not tell you the triggers, because the customers would try to avoid those.

It boils down to whether or not the software vendor believes they can get out more money out of it.

I worked at the IRS back in time, and one of the triggers was whether you had stains from a coffee cup on your papers, because that indicated that you had worked a long time on submitting the figures, and that indicated a high risk of errors/mistakes.