Be easy to leave
Every vendor is good at arrivals. Importers for the competition’s format, migration guides, a real person on a call for the first ninety days. That effort is genuine and the instinct is correct. The first week decides a great deal.
Now ask what the same company has built for the way out.
Importing from a competitor and exporting to one are the same engineering problem pointed in opposite directions. One gets a team and a place on the roadmap. The other gets a support article.
Both deserve the same care and almost never get it, because nobody inside a software company is measured on how easily customers leave. Retention has an owner. So does growth. The door has nobody, so it gets built once, early, to answer a question in a sales call, and then it rots while the product grows around it.
Which is why an export usually stops at a file. There is nothing wrong with a file. The question is whether it loads anywhere else, and whether anybody has ever checked. Mostly nobody has, and the first to find out is you, in the week you can least afford it.
It matters more than it sounds. How hard it is to leave sets how badly you can be treated while you stay. A company that knows you cannot go loses its best reason to stay good, and eventually it stops, however much it means well today.
Leaving should be one action, not a project. What comes out has to load into something the vendor does not own, and keep working after you stop paying, because that is exactly when people need it. Then the vendor should leave regularly themselves, often enough to know it all still holds.
We would rather be left easily than kept accidentally. Someone who stays because going is expensive is not a customer. They are a hostage with an invoice.
So look at both doors on whoever holds your work now. You already know how good they were at letting you in. Ask for everything back, put it somewhere they do not control, and see what survives the trip.
Do it while you are happy with them. That is the only time the answer is worth anything.
