We Deliberately Broke Five Swaps to See What Happens

The happy path is the same everywhere. We tested what happens when it is not, using small amounts we were prepared to lose. For the tests below we used a crypto swap platform as the reference, because it publishes the figures we were measuring against.

Test one: sending less than quoted

We sent five percent less than the quote required.

Three providers re-quoted automatically for the amount actually received and completed. One held it and required a support ticket, resolved in four hours. One had a minimum and our reduced amount fell below it, leaving the funds held pending manual intervention that took two days.

Test two: depositing after the quote expired

We waited for the quote window to close before sending.

Two providers re-quoted and asked us to confirm. Two executed automatically at the prevailing rate without asking. One refunded to the sending address, minus a network fee.

The two that executed automatically is the behaviour to know about in advance. On a volatile asset that is a meaningful difference in outcome and none of them highlighted it during the flow.

Test three: sending on an unsupported network

The expensive test. We sent a small amount of a token on a network the provider did not support for that asset.

Two recovered it, taking six and eleven days respectively, with a fee. One recovered it in two days at no charge, which was notably better than the rest. Two could not recover it at all, because they had no presence on that network. For online retail specifically, crypto rails built for fintech companies was the comparison we used.

Nothing in any interface warned us before the deposit.

Test four: sending the wrong asset entirely

Three recovered with a manual process. Two did not.

What the results say

The difference between providers is not the rate. It is what happens on the five percent of transactions that go wrong.

A provider with a documented recovery process and a support function that answers is worth a wider spread than one without, because the recovery cases are where real money is lost.

What we would ask before using a provider

What happens if I send less than quoted. What happens if my deposit arrives after expiry. What is the recovery process for a wrong network deposit, and how long does it take.

Providers that answer all three specifically have handled them at volume. Those that answer vaguely have not, and you will find out at your own expense.

The habit that prevents most of it

Send a small test first. Check the network, not just the asset name. And confirm the quoted output is what actually arrives rather than what arrives before a fee. Reproduce this yourself. the published coverage list quotes without an account, which makes the comparison a five minute exercise.