How to reconcile marketplace payouts without a connector

August 20, 202617 min read

You sell on a marketplace, money arrives in your bank, and it matches nothing. Not your sales figure, not the platform's dashboard, not the number the tax authority will eventually hold about you.

So you go looking for the tool everyone recommends, and find that your channel is not on the list.

This is the situation the reconciliation guides skip. They assume A2X or Link My Books covers your platform, and if it does, most of this article is unnecessary

buy the subscription. What follows is for the sellers those tools leave out, and for anyone who wants to understand what the connector was doing on their behalf.

The work splits in two. One evening to read a settlement report properly, then about twenty minutes per channel each month.

First, check whether you are actually without a connector

Before building anything by hand, spend ten minutes confirming the gap is real. Coverage changes, and the cheap end of this market is not where people look.

Checked on the vendors' own integration pages in August 2026: A2X lists six channels - Amazon, Shopify, eBay, Etsy, Walmart and PayPal - plus a multi-channel option. TikTok Shop is not among them; TikTok orders reach A2X only indirectly, routed through Shopify. Link My Books covers TikTok Shop directly, plus WooCommerce and Square. Neither lists Shopee, Lazada, Mercado Libre, Wish or Alibaba anywhere.

Two traps in that checking.

A "coming soon" page is not coverage. Link My Books has pages for Temu and Whatnot that read Coming Soon with a waiting list. They rank in search like product pages, and a seller skimming results will conclude the integration exists.

The cheap specialists are invisible in search. Seller Ledger, at $10 a month for 250 transactions, covers Whatnot, Poshmark, Mercari, Depop and TCGplayer - channels the well-known tools do not touch, at a price below A2X's entry tier. Check what it actually produces before you buy. It reports in the shape of a Schedule C, and does not post summary journals into your ledger; its Whatnot support runs through a browser extension, with no platform API behind it. That may be exactly what a sole trader needs, or nowhere near what a bookkeeper needs. Either way, look at the small vendors before deciding you are on your own.

If after that your channel genuinely appears nowhere, the rest of this article is your method.

The deposit is the marketplace's revenue figure, not yours

Here is the mistake that makes everything downstream wrong, and it is an easy one to make because the number is right there in the bank feed.

Marketplaces pay net. Commission, transaction fees, advertising, refunds and adjustments come out before the money moves. Book that deposit as revenue and you have recorded a figure that is arithmetically the platform's, not yours.

The accounting standards are unusually blunt about this. IFRS 15 says a principal "recognises revenue in the gross amount of consideration to which it expects to be entitled" (¶B35B), and describes the agent's position as a fee that "might be the net amount of consideration that the entity retains after paying the other party" (¶B36). The three indicators of being a principal - responsibility for fulfilment, inventory risk, discretion over price - describe a seller who buys stock, ships it and absorbs returns. US GAAP arrives at the same place through ASU 2016-08.

If standards feel abstract, the tax forms make the same argument by their shape. On Schedule C, line 1 is "Gross receipts or sales" and line 10 is "Commissions and fees" - nine lines apart, with no line anywhere for a net deposit. The ATO puts it in one sentence, in its ride-sourcing guidance: "GST must be calculated on the full fare, not the net amount you receive after deducting any fees or commission." Their example is a $55 fare where the platform keeps $11 and pays out $44, and the GST is calculated on $55.

Netting also destroys your own numbers. Gross margin computed on a net deposit has the wrong denominator, so it is not a margin at all.

Three official numbers about your sales, and none is the bank

This is the part that surprises even careful sellers: the disagreement between figures is deliberate, and knowing which is which saves an argument later.

In the US, Form 1099-K reports gross. The IRS states the gross amount "isn't adjusted for any: Fees, Credits, Refunds, Shipping, Cash equivalents, Discounts". So the form shows a bigger number than your bank ever saw.

The reporting threshold also moved recently, and much of the internet is out of date on it. The phase-down to $5,000, then $2,500, then $600 was repealed retroactively. For marketplaces and payment apps, the threshold went back to more than $20,000 and more than 200 transactions. Card payments have no threshold at all. And the threshold governs the paperwork, not the tax: income is reportable whether or not a form arrives.

In the EU, DAC7 reports net, quarterly. DAC7 is the rule that makes online platforms hand seller data to tax authorities, and the figure it reports is defined after fees are deducted - the opposite convention to the American form.

In the UK, the seller copy follows the calendar year, while the tax year runs 6 April to 5 April. Two documents about the same shop, cut on different dates, in different directions.

An eBay seller found a fourth axis worth remembering - settlement date against sale date: "no where on the 1099 help page on the eBay site or the 1099-K itself does it say that transactions that settled after the calendar year will not be included."

None of these numbers is wrong. They answer different questions, and only your books answer yours.

Read one settlement report properly, once

The method is unglamorous: export a single settlement report from each platform and write down what every column means. That is most of the work, and it is a one-off. Afterwards the monthly job is mechanical.

Three things make the columns harder than they look.

The column set is not stable. Shopee documents this itself: components that do not apply to your orders will not appear in the statement. A spreadsheet built around last month's export breaks silently when a fee type disappears or returns.

One order is not one row. Whatnot instructs sellers to sum the transaction amounts of all rows sharing an order ID. TikTok's export runs across five or six sheets. A naive one-row-per-order join produces a wrong total without producing an error, which is the worst failure shape there is.

The adjustment is not where the order is. Shopee's documentation is explicit that an adjustment to an order can land in a different statement than the order itself. Anyone matching statements to orders one-to-one will find a permanent, moving discrepancy.

Write the dictionary as a plain table: column name, what it means, whether it is money in or money out, and which account it maps to. Keep it next to the books.

The effort is the same one that decides whether automated document handling works at all. Getting invoice data out of PDFs succeeds or fails on how clean and well understood the incoming fields are, and a settlement report is the same problem wearing a spreadsheet.

The journal entry that survives a change of platform

Once the columns are named, the bookkeeping shape is stable across platforms, which is what makes this method outlast any particular integration.

Record the sales at gross, put the money in a holding account, then take the fees out of that holding account as they are documented. The holding account is the piece people skip. Accounting software usually ships with one called undeposited funds: a waiting room for money that is yours but has not reached the bank yet. A bookkeeper described the whole routine in two sentences: enter the sales as a journal entry with all money going to undeposited funds, transfer from there to the bank as funds arrive, then take the fees against that same account.

Done this way, the fees come out of the waiting room and leave the sales figure untouched, so it stays gross - which is where every tax form expects to find it.

In practice, one monthly entry per channel:

Line	Direction	Source column
Gross sales	credit revenue	gross merchandise value
Sales tax or VAT collected by the platform	credit liability, or omit entirely - see below	tax columns
Marketplace commission	debit expense	commission, referral fee
Payment processing	debit expense	transaction fee
Advertising taken from the payout	debit expense	ads, promotions
Refunds and returns	debit revenue contra	refunds
Reserve held back	debit holding account	reserve, on hold
Net cash received	debit bank	payout total

The entry balances only when every deduction has a home. If it does not balance, the gap is a column you have not identified yet - that is the signal to go back to the dictionary, not to force the total.

Money the platform is holding, and the report that does not exist

Reserves are where reconciliations quietly go wrong, because held-back money is real, yours, and frequently unreported.

TikTok Shop holds funds in a reserve for a period after delivery, and shows the balance in several places in its seller centre - statements, the finances summary, an on-hold view and the payouts screen. What it does not do is put that balance in the settlement report you reconcile from, so a seller working report-first can close a month without ever seeing money that is theirs and pending. Shopee's escrow releases roughly a week to a fortnight after delivery, and can run longer. Whatnot pays out manually, which means the payout schedule is a decision someone has to remember to make.

Two consequences worth planning for.

The money exists in your accounts before it exists in your bank, so a reserve needs a line of its own; folding it into an adjustment at the end hides it. And a payout can fail to arrive for reasons that look like nothing at all - a payout threshold set too high, or a payout method never configured, leaves funds sitting on the platform with no alert anywhere. That is an automation failing silently wearing a different hat: the absence of a deposit produces no error message.

Which exchange rate, and on which date

Sellers expect a rule and find a choice. No authority mandates a single rate; they mandate consistency and evidence.

The IRS says it in as many words: "The Internal Revenue Service has no official exchange rate. Generally, it accepts any posted exchange rate that is used consistently." Amounts get translated using the rate prevailing when you receive, pay or accrue the item.

HMRC is narrower for VAT. Notice 700 permits three methods: the UK market selling rate at the time of supply, HMRC's period rate for customs, or a rate agreed in writing with HMRC. Adopting the customs rate locks you in until HMRC agrees otherwise.

The European Central Bank rate is not among those three, despite appearing in a great deal of advice. The ECB says so itself, describing its reference rates as published for information and discouraging their use for transactions.

Income tax is looser than VAT. HMRC's manuals allow London closing rates, a bank's quoted rate, or the monthly rates published for VAT, and say HMRC queries a rate only when it diverges markedly from reputable sources.

One more trap worth knowing before you copy a link from an old article: the gov.uk exchange-rate collection that most of them point to was withdrawn in 2023. The live rates sit in the Trade Tariff service, published on the penultimate Thursday of each month for the month that follows.

Australia gives the cleanest rule for this exact case: translate at the earlier of when the income is derived and when it is received. A payout released in March for a February sale is translated at February's rate. Average rates are allowed over periods up to twelve months if the approximation is reasonable, and the ATO asks you to keep the rate and its source with your records.

Whichever you pick, write down the choice and the source. The rule you will be judged against is consistency, and consistency is only demonstrable in writing.

The VAT line in the report may not be yours to declare

If you sell across borders, one column can put tax in your return that legally belongs to the platform.

Under EU rules, a platform that facilitates certain sales is treated as though it bought the goods from you and sold them on. That applies to imported consignments worth up to €150, and to goods already inside the EU sold by a seller based outside it. What happens to your own leg of the sale differs between the two: for goods already in the EU it is exempt with the right of deduction, while for the imported consignment it falls outside the scope of EU VAT altogether. Either way it is not a sale you charge VAT on.

The UK works the same way below £135. The marketplace becomes liable for the VAT, and the overseas seller is treated as making a zero-rated supply to it. The £135 applies to the whole consignment, not to individual items. Anyone on the Flat Rate Scheme - the simplified VAT scheme for smaller businesses - should note those sales drop out of the flat-rate calculation.

So when the platform is treated as the supplier, the VAT figure in your payout report is the platform's tax to hand over, not yours. Copying it into your own return declares tax that is not yours and will never reconcile to anything.

Six ways this goes wrong

Every one of these appears repeatedly in practitioner discussions, and each has a tell you can check for today.

Booking the net deposit as revenue. Understates turnover, and makes margin meaningless because the denominator is wrong. In the US it also puts your return below a gross figure the IRS already holds; in the EU and UK, where the platform reports net, the mismatch is subtler and shows up as books that cannot be tied to anything.

Inventing a customer called "Shopify" or "Amazon". A monthly invoice raised against a fictional customer, never cleared, leaves a growing receivable that represents nothing. A bookkeeper who inherited exactly that found "a large accounts receivable balance sitting against the Shopify customer."

Plugging the difference. A seller whose 1099-K and payout report differed by about $6,000 was advised to force the match with a clearing account. His own instinct was the right one: an unjustifiable plug sitting there every month is a problem postponed, not solved. A difference you cannot explain is information.

Treating the platform's tax report as the books. UK guidance says it directly: those reports do not replace your business records or tax calculations.

Assuming the connector covered the channel. A2X's own documentation warns that TikTok orders arriving via Shopify record the gross sale, shipping and sales tax, but carry "no record of the other fees charged by TikTok". The predictable result - sales counted twice and fees missing entirely - shows up in bookkeeping forums regularly.

Copying the marketplace's VAT into your return, as above.

When doing it by hand is cheaper than a subscription

Manual is not automatically the frugal choice. This one deserves arithmetic, and most people decide it on instinct.

Connector pricing, as displayed in August 2026: A2X starts at $29 a month per channel at its smallest tier; Link My Books prices on a calculator, and at the lowest order band with one channel it quotes £15, US$20 or A$29 depending on your currency; the niche specialists start around $10. Against that, price your own time honestly: the dictionary is a one-off evening, the monthly entry is twenty minutes per channel once it exists, and the month a platform changes its report format costs an hour.

A commenter who does it manually gave the honest version: for a few payouts a month, splitting it by hand is not much of an issue - it depends on your time. The pressure point is not the number of orders, it is the number of channels and the number of separate settlement cycles you have to hold in one month-end.

The same person, on the alternative, is a reminder that buying a tool is not the end of the story: he reported an entire year of payouts his connector could not match correctly.

If you want the general version of this calculation, what automation costs a small business works through the same trade-off, and when not to automate covers the cases where the honest answer is to leave it alone.

What to do this week

Four steps, in order, and none of them requires buying anything:

Check the integration lists of the small vendors, not just the two famous ones - and treat any "coming soon" page as a no.
Export one settlement report per channel and write the column dictionary. Name every column, its direction and its account.
Rebuild last month with a gross entry, a holding account and a line for the reserve. If it does not balance, find the missing column instead of forcing the total.
Write down which exchange rate you use and where it comes from, then use that one every month.

An entry that balances without a plug is not proof that every figure is right - a wrong column mapping balances just as neatly as a right one. It is proof that nothing is unaccounted for, which is the part you cannot get any other way. The method also keeps working when the platform changes its report, and gives you something to check a connector against on the day one finally appears.

Rules on tax, VAT and currency differ by country and change without warning; the sources below are the primary ones, and your accountant knows your situation. Treat this as a method for getting the numbers straight, not as tax advice.

If you would rather have someone map this alongside the rest of your admin, that is what the process audit is for: $299, three business days, and a written account of what is worth automating and what is not.