A vendor with your data needs an exit you have tested

The claim

Before you commit your business's data to a software vendor, the question that matters most is not what the service does but how you get your data back out if you leave — and "we have an export button" is not an answer until you have pressed it and confirmed that what comes out is complete, usable, and enough to run your business on. Most businesses discover the true cost of leaving a vendor at the exact moment they most want to leave, which is the worst possible time to find out the export is a partial CSV missing half of what you need.

Why the exit is the decision

Every vendor relationship ends eventually — the vendor raises prices, degrades the product, gets acquired and pivots, or simply stops being the right fit — and your leverage in every negotiation with them is proportional to how credibly you could leave. A vendor who knows you cannot practically extract your data holds a position over you that has nothing to do with the quality of their service and everything to do with your captivity. The exit is not a pessimistic afterthought; it is the thing that keeps the relationship honest for its entire duration, because a vendor you can leave is a vendor who has to keep earning you.

What "getting your data out" actually requires

A real exit is more than a file download, and the gaps are where businesses get trapped. Evaluate four things before you commit, not after:

• Completeness. Does the export include everything, or just the obvious records? Often the primary data exports cleanly but the history, the relationships between records, the attachments, and the metadata do not — and those are frequently what make the data useful.
• Format. Is it a documented, standard format you can import elsewhere, or a proprietary dump that only the vendor's own system understands? An export you cannot load into anything else is a hostage with extra steps.
• Completeness of access. Can you get it yourself, on demand, via an API or a self-serve export — or must you file a request and wait on the vendor's goodwill and timeline, which evaporates the moment you tell them you are leaving?
• The pieces that never export. Configuration, workflows, integrations, and the accumulated customisation that made the tool fit your business usually do not come out at all, and rebuilding them elsewhere is the real switching cost.

Test the export before you depend on the vendor

The single most valuable thing you can do is run the full export early — during evaluation or in the first weeks, not years later when you are trying to leave — and actually attempt to use what comes out. Load it into a spreadsheet, into a competing tool, into a test database. The questions you are answering are concrete: is every record here, can you tell how the records relate to each other, are the dates and amounts intact, could you actually reconstruct your operation from this file alone?

Exit test, run during evaluation:
- [ ] export every data type the vendor holds
- [ ] confirm record counts match what is in the live system
- [ ] load the export into a different tool or a database
- [ ] verify relationships, history, and attachments survived
- [ ] time how long the whole process took

The time it took is itself a number worth knowing, because an export that takes a week of manual work is a very different exit from one that takes an afternoon, and you want that figure before it is on the critical path of an urgent migration.

Put the exit in the contract

Verbal assurances about data portability are worth what they cost. The contract should specify that your data is yours, that you can export it in a usable format on demand for the life of the agreement and for a defined window after termination, and that the vendor will not hold it hostage over a billing dispute. Pay particular attention to the termination clause: some vendors cut off access immediately on cancellation, which means the moment you give notice is the moment you lose the ability to extract your own data — so your export must be complete before notice is given, not after.

Weigh the switching cost honestly when you choose

The switching cost is a real part of the price of a vendor, and it belongs in the decision at the start, not as a surprise at the end. A cheaper tool that traps your data may be more expensive over its full life than a dearer one you can leave freely, because the trapped tool can raise its price toward the cost of the painful migration and you will pay rather than undertake it. When you compare vendors, compare not just their fees and features but the credibility of your exit from each — the one you can leave in an afternoon is worth a premium over the one that would take a quarter to escape.

What to do this quarter

For each vendor that holds data your business depends on, run the exit test once: export everything, try to use it, and record what you found and how long it took. The results sort your vendors into two groups — the ones you could genuinely leave, and the ones you are effectively captive to — and the second group is where your real risk sits. You do not have to leave any of them. You have to know that you could, because the ability to leave is the only thing that guarantees you are staying by choice rather than by capture.