Matching Settlement Currency: The Change That Saved the Most
We measured the same conversions with and without a currency mismatch. The difference dwarfed every other cost we had been optimising.
We had been comparing providers on spread and withdrawal fee. Then we tested the effect of the receiving account’s currency and found it dominated both. To reproduce any of this, a crypto to fiat exchange is the easiest starting point because nothing has to be requested.
The test
The same conversion, at the same provider, within the same window, settling twice: once to an account in the provider’s settlement currency and once to an account in a different currency. Repeated at three providers.
The results
Provider A: 0.19 percent matched, 1.31 percent mismatched.
Provider B: 0.34 percent matched, 1.12 percent mismatched.
Provider C: 0.28 percent matched, 1.94 percent mismatched.
Between four and seven times the cost, for the same sale at the same provider, decided entirely by which account received it.
Where the difference sat
Almost entirely the currency conversion margin, between 0.8 and 1.6 percent depending on provider.
The withdrawal fee also rose, because the mismatched case used an international rail rather than a domestic one, adding a flat amount and one to three business days. The institutional version we tested through a regulated European crypto platform, with its additional reporting.
Why this is not obvious
The conversion happens after the sale, so it does not appear in the rate quoted for the crypto.
None of the three providers displayed the applied currency rate before execution. All three showed only the final amount, from which the margin can be calculated afterwards but not compared beforehand.
The fix
Open an account in the currency the provider settles in.
For a European business converting through a European provider, that means a euro account. The one-off effort was repaid on the first meaningful conversion in every case.
What we had been doing instead
Comparing spreads across providers, where the differences were fifteen basis points, while paying over a hundred basis points on a conversion we had never measured.
That is the wrong optimisation and we had been making it for months.
The general lesson
Before comparing providers on any dimension, check whether a structural choice dominates the comparison.
Here it was the receiving account. Once matched, the provider differences became the small thing they appear to be, and the comparison we had been doing became worth doing. Coverage decided more of our results than any feature did. the list of countries covered publishes its country list.
Products covered here are bought at full retail price and used for real transactions before anything is written. There are no affiliate links on this site, no sponsored placements and no review units. If that ever changes, it will be disclosed at the top of the article.