Work out when your marketplace money actually lands

Sales are good, the dashboard is green, and there is no money to buy stock with. This is the ordinary condition of selling through channels you do not own, and it has a number attached to it. Most sellers have never worked the number out.

The number is not the payout schedule on the platform's help page. That is a promise about the normal case. The number you can plan against is what your own last three months actually did, per channel, and it is sitting in files you can export this afternoon.

The published schedule is not the lag

Every platform states a settlement time. Shopify Payments settles in three business days in the US, UK and Canada, five in Japan, seven in Mexico. Amazon runs a fourteen-day settlement cycle. eBay makes funds available within two business days and pays out weekly on Tuesdays. TikTok Shop pays somewhere between one and thirty-one days after an order is delivered.

Those are real numbers from the vendors' own pages. They are also the best case, and five separate things push your actual experience away from them.

Reserves. Shopify describes holding "a 10% reserve for 120 days", or a fixed amount such as $1000 held for the same period, with terms "communicated to you using email". Amazon holds an account-level reserve on top of its delivery-date rule. TikTok holds part of every delivered order for thirty days from delivery. Stripe may open one on your behalf. A reserve does not change the schedule; it removes a slice of the money from it.

Money coming back after it landed. Refunds and chargebacks are netted against future payouts. TikTok states it directly: "Any negative settlement amount will be deducted from your next payout." Last Tuesday's deposit is not unconditionally yours.

Fees are netted, not invoiced. Shopify's payout export carries Amount, Fee and Net as separate columns, which means your storefront's gross sales figure and your bank deposit are two different numbers on two different dates.

The clock often starts at delivery, not at sale. Amazon's delivery-date reserve and TikTok's settlement window both begin when the parcel arrives. Your carrier's performance sets your cash flow. eBay says the same in reverse: adding tracking "can help you receive your funds faster".

Weekends. Shopify: "Weekends and holidays don't count toward your settlement time." eBay pays Monday to Friday. The same sale made on Thursday and on Friday can land days apart.

None of this is hidden. It is just distributed across a dozen help pages per channel, and none of those pages can tell you what it adds up to in your account.

Both dates you need are already in your export

This is the part worth knowing, because it turns the whole question into subtraction rather than modelling.

Shopify. Finance, then Payouts, then export transactions to CSV. The file carries Transaction Date, Payout Date and Available On as columns. Payout date minus transaction date is the lag, per order, already in the file. There is also a payout reconciliation report under Finance and Documents which breaks out reserves and holds, though note it exports to PDF only and runs three days behind.

Amazon. Reports, then Payments, then Date Range Transaction reports. One request covers up to 365 days, and history goes back years, so pulling three months is a single export. In an announcement from March 2026, in effect since 30 April, Amazon added two things worth knowing about to these reports for the US, Canada, Mexico and Brazil: a flag for whether a transaction is deferred or released, and a column called Transaction Release Date. The announcement notes that "if a transaction remains deferred, the release date field will be blank" - so a blank cell is money still stuck, countable at a glance.

eBay. Seller Hub, Payments tab, Reports, then Financial documents, then Create report next to Transaction reports. Most come out as CSV, carrying transaction date, order number and payout date together.

TikTok Shop. Finances, then Statements or Payouts, then Export. Statements are generated daily against your settlement period. There is also a Reserve Details Sheet giving collection dates, expected release dates and reserved amounts, though it only covers transactions after 25 September 2025.

Stripe. Every balance transaction carries an available_on timestamp. That is the date the money became spendable, which is the date you want.

Four exports, one afternoon. No new software and no application form.

Take the median, not the average

The instinct is to average the days. The average is the one summary that will mislead you here.

The distribution is lopsided. Most orders settle on the normal schedule, and a handful - reserved, disputed, untracked, delivered late - sit far out in the tail. An average drags those outliers across every row and produces a figure that describes no order you have ever shipped. Worse, averaging a three-day channel against a thirty-one-day channel yields a number that is true of neither.

Two figures per channel do the job. The median tells you what a normal week looks like. The 90th percentile tells you what to hold cash against. The gap between them is the size of the surprise you are exposed to.

One Amazon seller put the boundary problem in a sentence: "If a payment you were expecting on the 14th day is actually the 15th day, you will not receive it until day 28." A schedule cannot show you that. A distribution built from your own deposits can.

What sellers actually run into

The reason to measure is that the schedule breaks without notice, on healthy accounts.

On the Amazon seller forums, a seller reported "well over 75% of my net sales being held" from late March, saying they had had no chargebacks, account health problems or refunds since 2015. In the same thread another wrote that "this morning my entire account balance went in to Account Level Reserve" and a third summarised the release timing as a "complete guessing game". A fourth was blunt about what it costs: "stripping sellers of their working capital. We have payroll to meet."

Amazon's own position, stated in its help pages, is that funds sitting in the account-level reserve are a normal part of selling on the platform. That is not evasion. It is the accurate statement, and it is exactly why the nominal schedule is the wrong planning input.

Shopify is not different in kind. A reserve can appear on an established store that has never had one and does not meet the published criteria, with most of an order's value held against a payout date months away, and a payout can run weeks past its date while support tickets accumulate.

Reserves also scale with the business as it grows: a new seller sees most of a few hundred dollars held back, and an established one doing serious volume sees a reserve in the tens of thousands. The percentage is what stays, not the amount.

What a published fee still does not tell you

Search any of this and the results are dominated by lenders. They are not lying, and they put a number on the lag in their marketing where the platforms bury it in help pages: why should that money sit in the marketplace's bank account for two weeks, the pitch goes, when you need it now.

The bias is structural rather than dishonest. They sell financing, so their analysis of the problem ends where a loan begins. Nobody in that set suggests measuring the lag first, because measurement is not something you can buy.

Credit where it is due on pricing, because the sector is more open than its reputation. Several of these firms do publish a headline figure. Onramp states a fee "typically ranging between 2 and 8%". Wayflyer's help centre works an example at an 8% fee on $100,000. CrediLinq advertises a service fee "as low as 1.5%/month". Shopify Capital spells out that "if you borrow $100,000 USD with a fixed fee of 13% of your loan amount, then the cost of borrowing is $13,000 USD".

The catch is what those numbers are. A flat fee is a cost without a term, and a cost without a term cannot be annualised. Thirteen per cent repaid over four months and thirteen per cent repaid over twelve are wildly different prices wearing the same label. The only number that makes two offers comparable is the annualised one: the total cost of the financing, interest, fees and every other required charge, expressed as a single percentage over a year. Ask for it. A lender that will not give it is telling you something about the price.

So the published figure tells you what you pay. It does not tell you what the money costs, and the two only become comparable once you know how fast you will repay - which is a function of your own sales and your own lag.

To be fair about it: borrowing against a payout gap can be a sensible purchase. A twenty-day gap on stock that turns quickly is close to the textbook case for short-term capital. The sequence is what matters. Measure the lag, size the gap, then price the money against a number you already know. A seller who cannot state their own median days-to-cash has no way to tell a cheap offer from a ruinous one, and is the ideal customer for a fee quoted without a rate.

What the finished report looks like

One row per channel, four columns:

1. Median days from sale to money in the bank.
2. 90th percentile days, which is your planning number.
3. Share of value currently sitting in reserve or on hold.
4. What you can actually commit to purchase orders this week.

That last figure is the deliverable, and it is the one no platform shows you, because no platform can see the others.

Build it once by hand from the exports. If it turns out to matter every week - and for most multi-channel sellers it does - it becomes a scheduled report rather than an afternoon: the exports pulled automatically, the two dates differenced, the distribution recalculated, one figure delivered every Monday morning. That is a small piece of work, and it is the kind worth automating precisely because the answer changes constantly while the method never does.

Two cautions from building this sort of thing. Reserves have to be tracked as a separate line rather than folded into the lag, because they are not slow money, they are withheld money. And the report is only as good as the exports behind it - if a channel's connector quietly stops pulling, the number keeps looking plausible while going stale, which is the failure mode behind automations that break without telling you.

If you would rather have someone build this against your own channels and numbers, that is what the process audit is for: $299, three business days, and a written account of what to automate first.

Sources

Payout timings come from the platforms' own documentation: Shopify Payments payout schedule, Shopify on reserves, Amazon's payments and settlement cycle, eBay on getting paid, TikTok Shop's payment terms, Stripe's payout schedule and settlement times, and PayPal on holds for new sellers.

Export paths for the calculation: viewing and exporting Shopify payout details, the Shopify payout reconciliation report, Amazon's updated payment reports with a transaction release date, and reconciling eBay sales transactions.

Seller accounts quoted here are public posts on the platforms' own forums: account level reserve holding 75% of net sales, and deferred transactions running past the stated window.

The published fees: Onramp on its own 2-8% fee, Wayflyer's worked example of an 8% fee, CrediLinq, and Shopify Capital's 13% worked example.