It is not easy to find a stable solution. In the efforts to get things right, many processes swing too far one way then the other. In the theoretical world of control theory from college days, there are mathematical equations for this, and consequently the optimal solutions can be calculated. In the real world, we tend to swing well past the ideal point before we realise that we have passed it.
When choosing the England football manager after the honest long-ball bluntness of Graham Taylor, we got the urbane christmas trees of Terry Venables. After the enthusing Kevin Keegan, we got the suave Sven Goran Ericsson. After the smooth Swede, back to anyone English, even if under-qualified. Then back to experienced foreigner.
It is the same when developing any major new projects. Wherever you start, you will not spiral in towards perfection, instead you will oscillate. If you start with an outdated homegrown development that is becoming too complex, you will buy in a simple standard product. When that product becomes too inflexible for business requirements, then you will go back to big global development. When the global solution becomes too remote for individual market needs, then you will develop something local. When you have too many locally developed solutions, then you will go back to a central universal product.
Those principles apply everywhere. Not because any of the global corporations where I have worked have taken fundamentally wrong decisions, but because there is an inevitable evolution of business requirement and technological capability.
Wait long enough, your best decision will look odd enough.