You close January. Consolidated net profit across your US, UK and German accounts comes to €18,400. Three weeks later you open the same report and it says €17,950. Nothing sold, nothing refunded, no data was corrected. The number simply moved.
This confuses sellers more than almost any other reporting behaviour, and it usually gets blamed on the tool. It is not a bug. It is what happens when a single profit figure is assembled from transactions denominated in three currencies, settled on different dates, at exchange rates that keep moving after the sale.
If you sell across marketplaces, understanding this is the difference between trusting your consolidated report and quietly ignoring it.
What multi-currency profit reporting actually involves
A consolidated P&L across marketplaces requires converting every line into one reporting currency. Each conversion needs a rate, and each rate needs a date. There are three defensible choices, and they produce three different answers:
| Method | Rate used | Effect |
|---|---|---|
| Transaction-date rate | Rate on the day of each sale, fee and refund | Most accurate economically; the figure stops moving once the period closes |
| Settlement-date rate | Rate on the day Amazon paid out | Matches your bank; profit shifts as settlements land |
| Period-average rate | One average rate for the month | Smooth and simple; obscures within-month swings |
| Current spot rate | Today’s rate applied to all history | Whole history restates every day — avoid |
The fourth row is the culprit behind most “my past profit changed” complaints. If a report applies today’s rate to a sale from eight months ago, then every report of every historical period is a function of today’s currency market. Last year’s best month can become a worse month because of something that happened in the FX market this morning.
Why the number moves even when the method is sound
Even with transaction-date conversion, a freshly closed month can shift for legitimate reasons:
- Late-arriving charges. A storage fee or long-term storage surcharge assessed after month end but relating to that month gets booked back into it, converted at its own date’s rate.
- Refunds against earlier sales. A February refund on a January order reduces January’s profit if you report on an accrual basis. The refund converts at February’s rate while the original sale converted at January’s — so the two do not net to zero even for identical amounts.
- Reimbursements. Same problem, longer lag.
- Reserve releases. Amazon’s withheld reserve moves between periods and currencies.
A month is not truly final until refunds and fee assessments against it have stopped arriving. That is typically 60 to 90 days, not 30.
The refund asymmetry, with numbers
This one costs real money and almost nobody models it.
Assume a UK sale reported in EUR:
- January: item sells for £40. GBP/EUR = 1.19. Reported revenue: €47.60
- March: item is refunded in full, £40. GBP/EUR = 1.14. Reported refund: €45.60
Reported revenue: €47.60
Reported refund: −€45.60
Net reported: €2.00
You sold nothing, refunded everything, and your consolidated P&L shows €2.00 of revenue. You also paid a return processing fee and lost the unit. The €2.00 is pure FX noise, and across a few thousand refunds it is not noise any more.
Now run it the other way, with GBP strengthening to 1.24 by March:
Reported revenue: €47.60
Reported refund: −€49.60
Net reported: −€2.00
The same commercial event now shows a €2.00 loss. Neither figure is wrong. Both are artefacts of reporting a two-currency business in one currency.
What this means for real decisions
Never make pricing decisions from the consolidated view. Price is set in the marketplace’s own currency, against costs that may be in a third currency. Converting to a reporting currency first adds a variable that has nothing to do with whether the price works. Set prices per marketplace, in local currency.
Watch where your costs sit relative to your revenue. A seller buying in USD and selling in EUR has a margin that moves with EUR/USD whether or not anything about the business changes. If that exposure is meaningful, it belongs in your break-even calculation as a buffer, not as a surprise.
Compare marketplaces on local-currency margin percentage, not converted absolute profit. Margin percentage is currency-neutral. Converted profit is not, and ranking marketplaces by it will sometimes reorder them for no operational reason.
Reconcile to the bank on settlement-date rates, and to performance on transaction-date rates. These are two different questions. Trying to answer both with one number produces a figure that answers neither.
Best practices
- Fix a reporting currency and don’t change it. Every change re-bases your entire history.
- Insist on transaction-date conversion for performance reporting. If your reporting can’t do that, treat historical comparisons with suspicion.
- Treat months as provisional for 90 days. Label them that way internally so nobody builds a decision on a figure that is still settling.
- Keep a local-currency view per marketplace alongside the consolidated one. It is the version you actually manage from.
- Book costs in the currency you paid them in, converting at the payment date, not at today’s rate.
- Separate FX effect from operating performance when a period looks unexpectedly good or bad. The question is always: did margin move, or did the rate move?
Tools like sellerboard handle per-marketplace P&Ls with consolidated reporting on top, which is the structure that lets you answer the local-currency question and the group question without conflating them.
Common mistakes
- Applying today’s rate to history. The single most common cause of “my numbers keep changing.”
- Netting refunds against sales at a single rate. Hides the asymmetry above.
- Comparing this January to last January in converted terms without checking whether the rate moved more than the business did.
- Holding one blended cost per SKU across marketplaces when inbound freight, duty and local fulfillment differ by destination. The same unit has a different landed cost in the UK than in the US.
- Treating the consolidated figure as the operating number. It is a reporting output, not a management input.
- Ignoring FX on supplier payments. If you pay in USD and a purchase order takes 90 days from deposit to delivery, your landed cost was set by the rate on two different dates.
FAQ
Which conversion method should I use? Transaction-date for performance reporting, settlement-date for bank reconciliation. They serve different purposes and you need both.
Why does my consolidated profit differ from the sum of my per-marketplace profits? If each marketplace is reported in its own currency, summing them requires conversion, and the result depends entirely on which rates were applied. The two figures should reconcile once you know the method; if they don’t, the method is inconsistent somewhere.
Should I hedge currency exposure? For most sellers the exposure isn’t large enough to justify the complexity. The practical alternative is building an FX buffer into your target margin and holding cost reserves in the currency you buy in.
Does the marketplace currency converter cost me anything? Amazon’s own currency conversion for disbursements carries a spread. Local bank accounts in each currency avoid it, at the cost of more accounts to manage. It is a fee line worth quantifying before assuming it is negligible.
How long until a month is genuinely final? Practically, once refunds and late fee assessments against that period have stopped — usually 60 to 90 days depending on your return rate and categories.
Conclusion
Multi-currency reporting introduces a variable that has nothing to do with how well you sell: the exchange rate, measured at a date somebody had to choose. Once you know which date your reports use, the moving numbers stop being mysterious and become predictable — and predictable is enough to work with.
The discipline is simple. Manage each marketplace in its own currency. Use the consolidated view to see the shape of the whole business, not to make the decisions. And when a period moves after it closed, check the rate before you check your assumptions about the business.
