Methodology

How we work out what a transfer actually costs, what we can measure, and what we can't. Every number on this page is computed from the same data as the comparison tables, so the two cannot disagree.

1. The all-in cost, not the fee

A transfer costs you two things: the fee the provider names, and the gap between the real exchange rate and the one you are offered. The second is invisible on a receipt, and it is usually the larger of the two.

All-in cost = fee + (reference rate − offered rate) × amount

This is why a “zero fee” transfer can be worse value than one charging a visible fee, and why we rank providers by how much your recipient actually receives rather than by either number alone.

2. What the margin actually is

Measured across 306 live quotes in 79 corridors. Margin is the provider's rate against our reference rate; the fee is shown as a percentage of the amount sent, at a reference amount of 1,000 units.

ProviderQuotesMedian marginMedian feeMargin's share of cost
TalkRemit106.72% (above ref.)4.10%0%
Wise690.00%1.31%0%
TapTap Send370.09%0.00%100%
Sendwave290.42%0.00%100%
WorldRemit480.69%0.10%87%
Ria380.86%not published
Dahabshiil160.98%1.80%35%
Remitly591.65%0.10%94%

For every provider that publishes a fee, the margin accounts for 0%100% of what you pay. The fee is the smaller part of the bill in every case we can measure. Pooled across all providers, the median margin is 0.25%.

Two providers publish no fee in the data we can access, so their fee column reads “not published” rather than zero — we will not record an unknown fee as free.

3. Where our reference rate comes from

The reference (or “mid-market”) rate is the wholesale rate banks trade at. No retail customer gets it; it is the yardstick everyone marks up from. We prefer official sources that have no stake in the rankings:

  • European Central Bank daily euro reference rates — our primary source.
  • IMF representative rates, as reported by each issuing central bank — used where the ECB publishes nothing, notably the dollar-pegged Gulf currencies.
  • Wise's published mid-market rate — the fallback where no official source covers the pair.

Today an official source backs 15 of 79 corridors we could measure; the other 64 use the Wise-derived rate.

A limitation we would rather state than hide

Wise is also one of the providers we rank. Where the reference comes from Wise, the spread column is measuring competitors against a listed competitor's own rate, and Wise will appear to sit exactly at the reference by definition. Those corridors say so in the table footer. Central banks simply do not publish rates for most remittance destination currencies, so this is a gap we disclose rather than one we can close.

4. Where the cost formula breaks

On 31 of 306 quotes, providers offer a better rate than the reference. Applied literally, the formula above would report a negative cost — as if the provider paid you to send money. It does not, so we do not publish it that way.

The clearest case is ETB: 24 quotes sit a median of 9.75% above the reference rate.

CurrencyQuotesMedian above reference
ETB249.75%
COP21.03%
MAD13.70%
GHS12.02%
PHP11.18%
NGN10.81%
PKR10.55%

The reason is currency controls. Where a central bank maintains an administered official rate alongside a parallel market, providers source the currency nearer the street rate, so their quotes beat any official reference. Ethiopia is the clearest example. In those corridors the reference rate is not a rate you could actually get, and treating it as one would overstate how good every offer is.

This is the main reason we rank by how much your recipient receives rather than by a computed “you save X%” figure. Recipient value stays meaningful whether or not the reference rate is achievable.

5. Why margins differ by corridor

  • Liquidity — thinly traded currencies are more expensive for a provider to source, so spreads widen.
  • Compliance cost — licensing and anti-money-laundering obligations in both the sending and receiving country.
  • Currency controls — administered rates and parallel markets, as described above.
  • Payout network — cash pickup carries per-transaction costs that bank and mobile-wallet payouts do not, and those costs are usually built into the rate.
  • Competition — high-volume corridors are fought over and carry thinner margins than low-volume ones.

6. How we rank, and when we decline to

Providers are ordered by the amount the recipient receives after both the fee and the rate. Where the top two are within 0.5% of each other we call it “best available” rather than “best value”: that margin is smaller than the movement we routinely measure between a quote and checkout, so declaring a winner would imply precision we do not have.

7. Freshness and its limits

Rates are collected every four hours and corridors older than 36 hours are hidden rather than shown stale. From our own history, rates drift about 0.04% per hour, with a mean absolute movement of 0.19% between four-hourly snapshots — small relative to the differences between providers, but not zero. Official reference rates are published once per business day, so they do not move intraday or at weekends.

Fees are captured at a reference amount of 1,000 units. Providers whose fees vary by amount may charge differently for your transfer, and the calculator says so when you enter a different figure. Always confirm the final quote on the provider's own checkout page.

8. What we do not measure

  • Banks. We track licensed money transfer providers only. Bank transfer costs are typically far higher, but we do not quote a figure we have not measured ourselves.
  • Hedging and rate locks. Whether a provider fixes your rate at payment or at settlement is not visible in the data we collect, so we make no claim about it.
  • Cash-agent networks we cannot price. Where a provider exposes no quote for a corridor, it is absent from that table rather than estimated.

Corrections

If a rate or fee here disagrees with what a provider quoted you, we want to know — that is a data problem worth fixing. Email support@xrate2send.com.