What is this transfer actually costing me?
Measure it as the gap between what the payment was worth at the mid-market rate and what actually reached your account, as a percentage. Call it the in-transit cost. It captures any sending fee deducted from the payment, the exchange rate markup, and anything else taken out along the way. It deliberately does not include charges billed to your account separately, which you add on top to get your total. Everything else is diagnosis.
That sounds obvious, and it is. It is also not how the cost is usually presented to you, which is why most people cannot answer the question about their own payments.
Why can't I just compare the fees?
Because the fee is one component of the cost, and providers are free to choose which component they compete on.
A provider charging a $4.50 fee and a 1.1% markup can advertise a low fee. A provider charging nothing and a 3% markup can advertise no fees at all. Both statements are true. Neither tells you which is cheaper, and the second one is cheaper-sounding while being more expensive.
Calculating the exchange rate markup gets you halfway, and it is the half most commonly explained. It leaves the fees to one side. The two have to be added together before any comparison means anything, and once you add them the advertised numbers stop mattering.
How do I work out the in-transit cost?
Compare what arrived against what would have arrived at the mid-market rate. The gap is the value lost between the amount sent and the amount received, whatever its components were.
You need three figures, all of which you already have or can look up:
- What was sent, in the sending currency
- What arrived, in your currency
- The mid-market rate at the time the conversion happened
Then:
In-transit cost % = 1 − (amount received ÷ (amount sent × mid-market rate))
This captures charges deducted from the payment. Charges billed separately, whether to your client by their bank or to you by yours, sit outside it and have to be added on.
Worked through: your client sends $1,000 through a provider charging a $4.50 fee and a 1.1% markup. This provider takes the fee first and converts what is left, so $995.50 is converted at 1.1% below the mid-market rate of 84.50. At the mid-market rate the full $1,000 would have brought you ₹84,500. You actually received about ₹83,194.
The gap is about ₹1,306, which is ₹1,306 ÷ ₹84,500 = 1.55%, or about $15.45. That is more than the advertised 1.1%, because the fee is in there too.
The order matters, and it is not standardised. A provider that applies its markup to the full amount and adds the fee separately produces a slightly different result from the same headline numbers. You do not need to know which order your provider uses, because the in-transit figure captures the outcome either way. That is the point of measuring it from the two ends rather than from the components.
That number covers every cost taken out of the payment itself: the sending fee, the markup, and anything deducted in between. You did not have to identify any of them individually. What it does not cover is a charge billed to your account rather than removed from the payment, which is the next section.
What if I don't know the mid-market rate from that day?
Look it up rather than guessing. Rate-tracking sites publish historical mid-market rates. Same-day precision is enough to compare providers or spot a large markup; if you are auditing a specific transaction, use the rate at the time the conversion was executed. Some central banks publish daily reference rates for accounting and statistical use. Those are an approximation rather than a substitute for the market rate at the moment of your conversion.
Two things to be careful about. Use the rate from the day the conversion happened, not the day the payment was initiated, since those can differ by several days. And use the mid-market rate rather than the rate your provider displayed, which is the number you are trying to test.
If a charge was billed to your account separately rather than deducted from the payment, it will not show up in this calculation. Add it manually. This is the one gap in the method, and it matters most when your bank charges an incoming payment fee it bills to you rather than deducting.
How do I compare two providers with this?
Convert both to the same unit. Cost per $1,000 works well, because it is a number you can hold in your head and carry between providers.
Take a $2,500 payment through two providers:
| Provider | How it is advertised | You receive | In-transit cost | Cost per $1,000 |
|---|---|---|---|---|
| A | $4.50 fee | ₹208,550 | 1.28% | $12.78 |
| B | No fees at all | ₹204,912 | 3.00% | $30.00 |
Provider A costs $31.95 on this payment and Provider B costs $75.00. The last column normalises them, so you can carry the figure across to any other provider or amount.
Provider B advertises the better deal and delivers more than twice the cost. Nothing either of them said was false.
Cost per $1,000 lets you compare things that are not directly comparable, like a bank wire against a platform withdrawal against a multi-currency account. They have different fee structures, different rails and different timings, but they all reduce to a single number once you measure what arrived.
Does the answer change with the amount?
Yes, and this is the part most people miss.
A flat fee is a large percentage of a small payment and a rounding error on a large one. A markup is the same percentage regardless. So the cheapest provider at $300 a month is often not the cheapest at $4,000 a month, and the difference is easy to miss unless you check.
The point where the answer flips can be calculated. Where F is the flat fee, s₁ is the lower markup and s₂ is the higher one:
Crossover amount = F ÷ (s₂ − s₁)
Worked through with an example, a $5 fee against markups of 0.5% and 3% puts the crossover at $200. Below that, the no-fee option wins. Above it, the markup does more damage than the fee ever did.
Re-run this whenever your monthly volume changes materially. It is the single check with the best return on the two minutes it takes.
What do I do with the number once I have it?
Annualise it, because that is the version that motivates action.
If you receive $3,000 a month consistently, Provider A's structure works out at about 1.25% on that amount, or roughly $450 a year. Provider B stays at 3.00%, or roughly $1,080. That is a difference of about $630 a year for the same work, the same client and the same amount of effort.
Note that A's percentage moved between the two examples, from 1.28% at $2,500 to 1.25% at $3,000. The flat fee is a smaller share of the larger payment. This is the same effect described in the previous section, and it is why the percentage has to be recalculated at the amount you actually receive.
Then decide what is worth changing. In rough order of how much they usually move the number:
- The rail. If your client can pay into an account local to them, that leg can settle through their own domestic payment system rather than travelling through a chain of correspondent banks. This does not make the money free: you still convert it and move it eventually, so the cost shifts rather than disappearing. It often shrinks, because you control the conversion instead of inheriting it.
- The provider. Once you have in-transit cost per $1,000 for two or three options, this is a straightforward comparison rather than a marketing one.
- When you convert. If you receive in one currency and convert separately, the timing of that conversion is the one variable you set yourself.
- The invoice size. Batching monthly instead of per milestone reduces how many times a flat fee applies. It also delays your cash and concentrates the whole month into one exchange rate, so it is a trade rather than a free saving.
- The charge code, if the payment travels as a wire. Whether the option exists depends on the route and on your client's bank, and on some corridors it is restricted by regulation rather than by choice.
How often should I check?
Three comparable payments, meaning the same provider, the same corridor and similar amounts, are enough for an initial baseline, and twelve months of them become an annual audit. After that, once a quarter, or whenever something changes.
The reason to keep the record is not the individual number, which will vary. It is that a provider's pricing drifts, your volume changes and rates move. Without a baseline you have no way to notice any of it. A row in a spreadsheet with the date, the amount sent, the amount received and the calculated percentage is the entire system.
Almost nobody keeps it, which is why almost nobody knows what getting paid is costing them.
This article explains a calculation method. It is not financial advice, and it cannot account for your individual circumstances. The figures used are illustrative.