According to ReBound Returns, returns rose 80% in the first week of December after Black Friday 2025 and roughly one in three shoppers who bought over the Black Friday and Cyber Monday weekend sent something back. Then a second spike followed in early January, with returns up 59% as Christmas purchases came home.
For a finance team, those two numbers add up to a lot of extra work. The busiest trading week of the year is also the one that creates the largest refund liability, and the two arrive just ten days apart.
Not all peak revenue is cash
A Black Friday sale sits in the sales report as revenue the moment it is taken. But with extended return windows running into late January, a slice of that revenue is money the business is holding on the customer's behalf until the window closes.
That slice can be a significant amount too. If one in three peak shoppers returns something, then a meaningful share of the record weekend is provisional. That means the cash forecast that treats it as banked overstates January by roughly the return rate.
What the returns surge does to January
If the weekend after Black Friday was the biggest in the company's history, then it’s entirely possible that ten days later the warehouse is receiving more parcels than it is sending. The corresponding refund run can be larger than a normal month's payroll, and the returned stock is sitting in a quarantine bay waiting to be graded and re-listed. Unfortunately for the accountant, they’ll be asking why the December cash position looks nothing like the December sales report.
In this scenario two things have gone wrong. Cash that was allocated to January supplier payments and staff, on the strength of peak sales, is being refunded. And the stock behind those refunds is tied up twice, once as cash out and once as inventory that cannot be sold until it has been processed.
Why faster returns processing doesn't fix the forecast
The usual responses are operational. That could be processing returns faster so refunds and re-listing happen sooner, shortening the return window, or charging a restocking fee. Each one helps a little, by reducing how much cash is refunded or how long the stock is stuck.
But none of these responses changes when the refund hits the forecast. The team is still reacting to refunds after they arrive, and the forecast is still built on full peak sales. Faster processing can move the correction earlier in December, but it never gets ahead of the sale itself.
When should a refund hit the numbers?
A share of every peak order will come back, and you can estimate that share in advance from your own history. It varies by category and channel: fashion and footwear come back far more than homeware, and a marketplace sale follows the platform's return policy rather than yours.
So the liability exists from the moment the order is taken. Recognising it then, rather than when the parcel arrives, is a forecasting decision rather than an accounting change.
Book the liability at the sale
In practice this means running two December numbers rather than one:
- Apply a return-rate assumption by category and channel to peak revenue as it is booked, and hold that share out of the cash forecast until the return window closes.
- Build the January supplier and payroll calendar on the number that is expected to stick, rather than on gross sales.
- Forecast the recovery of returned stock separately, because a refunded item only becomes saleable cash again once it has been graded and re-listed.
This is revenue-side and forecasting discipline. The tools are the order management system, the returns platform and the cash-flow model. The work belongs with the accountant and whoever owns the forecast, but realistically spend tooling has no part in it.
What changes when you book it early
With a return rate calculated in advance, by category and channel, you can set the peak reserve at the point of sale rather than waiting for the parcels to come back.
This means the refund run in the second week of December lands close to a figure you have already planned for, so the January supplier and payroll calendar can be built on the sales expected to stick rather than on gross peak.
The same assumption covers the returned stock, which is tracked to a re-list date rather than left as an unplanned gap. Peak will still create a large refund liability, but once the forecast carries it, it stops arriving as a shock.
