Replenishment

Continue selling when out of stock, and for how long

The switch is one click. What Shopify does not tell you is how many days you can take orders for goods you do not have, or what the negative number does to your next order.

Marcus VolstedCo-founder, Restocio Published 10 October 202614 min read

Every source on this page was opened and read, and each one is dated in the Sources section at the foot.

Continue selling when out of stock is safe for as long as you can name the day the goods land and the customer has accepted it. Past that point you are taking money for a delivery you cannot schedule. This is the mirror of stop running out of stock. That page is about the product Shopify refused to sell at zero, so its empty days hide demand. This one is about the product you kept selling at zero, so nothing is hidden. The demand sits in your admin as a negative number, and it belongs inside your next order.

What the switch does in Shopify

Shopify's Help Center page on selling out of stock products, read on 29 September 2026, sets the starting position. An item is out of stock “when inventory is tracked and the inventory level is at zero or below”, and by default “customers can't purchase an item until you adjust the inventory levels to add more stock.” The switch that changes this is called Continue selling when out of stock. You find it in the Inventory section of a product or of a single variant, and it only appears once Inventory tracked is on for that item.

The same page names four reasons you might want it: products on preorder, stock that is coming soon, selling first and ordering supplies later, and not tracking real quantities in Shopify at all. Three of those are a plan. The fourth is the absence of one, and it is worth knowing which of the four you are in before you leave the switch on.

Two caveats on the page matter more than they look. The setting does not apply to orders placed through Shopify POS, where staff can keep selling below zero regardless and get a warning first. And with more than one active location, stock is assigned per location, so an item can read as out of stock online while units sit at a location that does not fulfil online orders. When that happens the fix is stock at the fulfilling location, not the switch.

	Switch off (Shopify's default)	Continue selling when out of stock
What the customer can do at zero	Cannot buy	Can buy
What the stock figure does	Stops at zero	Keeps falling below zero
What a zero day means for your sales rate	A day you could not sell, so it comes out of the divisor	A selling day like any other, so it stays in
What the empty shelf means for the next order	Demand you did not see, so the rate has to be corrected	Demand you did see, so the units are already counted and owed
Who is waiting	Nobody has paid	Every unit below zero is a paying customer

Where the negative number comes from

Shopify counts down through zero. Its inventory reports page, read the same day, gives the plainest example: a product set to 5 in the morning that takes eight orders during the day ends the day at −3, “because you oversold the product by three units.” The page lists that as one of two reasons for a negative ending quantity. The other is a product with tracking off, which starts at 0 and reads −2 after two orders.

So a minus figure in your admin has three possible causes. The switch was on and the product kept selling online, a till sale went through at zero, or the product was never tracked. Only the first is a decision. It is the one this page is about, and it is the one your reorder maths has to be able to recognise.

Why the negative number is not a counting error

A stock figure that has drifted away from the shelf is a different problem, covered in where Shopify's count drifts when goods arrive. Drift is an accident, a short delivery nobody logged or a miscount at receiving, and the right response is to fix the number. Stock at −300 under continue selling is not drift. The shelf really is empty and 300 units really have been sold. Correcting it to zero would be the error.

Read it as two numbers instead of one. The shelf holds nothing, and 300 units are owed. Owed is the useful word, because it says what the figure is: orders taken and paid for, goods not delivered. That is committed demand, and it is the one kind of demand you know to the unit. Under the default policy the same empty stretch would have left you guessing how much you missed. Under this one, the guess is replaced by a count.

Two ways the reorder maths goes wrong, in opposite directions

Most reorder maths, ours included until 31 August 2026, was written for the default policy. It assumed stock stops at zero and that an empty day is a day you could not sell. Feed it a product that kept selling at zero and it makes two mistakes at once, pointing opposite ways.

The first is flooring. A figure of −300 gets rounded up to 0 because negative stock looked like nonsense, and the 300 owed units vanish from the calculation. Nobody orders them. The next delivery lands, the waiting customers take it, and the shelf is empty again before a single new sale.

Then there is counting. A tool that divides sales by in-stock days, which is the right thing to do under the default policy, reads every day at zero or below as a stockout day. A product that was on preorder for the whole window has no in-stock days at all, so its sales get divided by one. On a 60-day window the rate comes out sixty times too high, and every quantity downstream inherits that.

Take a product that sells 10 a day, sits at −300 today, and has been selling at zero for the whole 60-day window, 600 units in all. With 30 days of cover as the target, the two readings look like this.

	Read as a default-policy product	Read as a continue-selling product
Selling days in the window	0, every day at or below zero treated as a stockout	60
Sales rate	600 ÷ 1 = 600 a day	600 ÷ 60 = 10 a day
Shelf and owed	Shelf 0, owed units floored away	Shelf 0, 300 owed
Starting quantity for 30 days of cover	600 × 30 = 18,000 units	10 × 30 + 300 = 600 units

Fix only the flooring and you order the 300 you owe. Fix only the counting and you stop over-ordering by a factor of sixty. You need both, and the second one is invisible unless your tool knows which policy the product was on, on each day of the window.

That starting quantity is not the order, either. It is the seed. The full chain sizes the boat on what the shelf will hold when it docks and adds faster legs on top when the boat is too late, which is worked through on the replenishment page. With this product, 10 a day and 300 owed, on that page's example settings (70 days by sea, 20 by rail, 5 by air, with 30 spare days on the boat, 7 on rail and 3 on air), the chain lands at 1,300 by sea plus 570 by rail and 230 by air, 2,100 in all. Your own lead times and spare days move the figure; the shape does not change. Because the shelf is already empty the fast legs run from today, so the first units to land go to the 300 customers already waiting.

How long is it safe to keep selling?

For as long as three dates agree. The day you can deliver, the day you told the customer, and the day the law fills in when you told them nothing.

The first is the only one you control with stock. Goods already on the water with a purchase order behind them give you a landing date, and receiving adds a day or two. Goods ordered but not yet shipped push the date out by the factory's time as well as the crossing. Goods you have not ordered yet leave you with no date at all, only the promise of one: the air lead time from the day you finally order at best, and the sea lead time if the boat is the only route you use. For an importer that is a spread from about a week to more than two months before anything goes wrong. Which part of an order to fly is the same question in a different coat.

Your product page sets the second date, and Shopify's switch says nothing about it. It lets the customer buy. Whether they are told the goods are not here depends on your theme or a preorder app, and the safe assumption is that the page says nothing you did not write yourself. A date you have written is a promise. A blank is an invitation to assume the goods ship tomorrow.

Merchants forget the third date. In the United States, the FTC's Mail, Internet, or Telephone Order Merchandise Rule, read on 29 September 2026, requires a reasonable basis to expect you can ship within the time you advertised or, if you advertised none, within 30 days, and when you cannot, you must obtain the buyer's consent to the delay or refund the unshipped goods. In the EU, Article 18 of the Consumer Rights Directive, 2011/83/EU, read the same day, says that unless you agreed a delivery time with the customer you must deliver “not later than 30 days from the conclusion of the contract”, and that a customer who has called for delivery within a further period and still has nothing may terminate the contract and be reimbursed. The Commission's Your Europe guidance on shipping and delivery says the same in plain words. Other markets have their own defaults, and the detail of both rules is for your own adviser, but the shape is the same. No date written down means 30 days, and 30 days is shorter than one sea crossing.

Which situation you are in

Your situation	What sets the safe window	What to do with the switch
Goods on the water, purchase order placed, arrival date known	The landing date plus receiving, checked against what the product page promises	Leave it on, and put the date on the page
Goods ordered, not yet shipped	Production plus the crossing, with a margin for a late boat	On only if the promised date carries that margin
Not ordered yet, selling first	Your fastest route's lead time from the day you order, at best	Order today or turn it off; every sale is a promise with no date behind it
A preorder for a product not yet made	The date on the product page, and the 30-day default if there is none	On, with the date written where the customer sees it before paying
On by accident since setup	Nothing; the store is overselling without knowing	Off, one variant at a time, and check what is already owed

The last row is more common than it sounds, because the switch is set per product or per variant and there is no reason to look at it again once a delivery lands and the number goes positive. It waits for the next zero, and then the store is quietly taking orders for goods with no boat behind them.

When to turn it off

When the first date stops being a date. A boat that slips two weeks is a new promise to every customer who bought since the old one was made, and there are only two honest moves: tell them and keep selling, or stop selling until the date is real again. What is not honest is leaving the switch on because the number in the admin still looks like a plan.

The cleanest habit is to tie the switch to a purchase order. On when the order is placed and the arrival date is on the product page, off when the arrival date passes without goods, on again when the next order exists. It is more clicks than leaving it on everywhere, and it is the difference between a preorder programme and an overdraft on your customers' patience.

How Restocio reads a negative stock figure

It reads the policy first. For every variant Restocio pulls Shopify's inventory policy alongside the quantity, so it knows whether a product is allowed to sell below zero before it looks at the number. A negative quantity is then split in two: the shelf, floored at zero, and the units owed. The owed units are added to the need at every step, in the starting quantity and in the projection that sizes the boat, so the customers already waiting are inside the order rather than outside it. Because the shelf is empty, days of stock reads zero and the fast legs are sized from today, which is why the first units to land go to them.

The rate is protected on the other side. A day recorded at zero only counts as a stockout day if the product was on the default policy that day. Under continue selling the day stays in the divisor, and a day recorded below zero never counts as a stockout at all, because the product sold. A product launched on preorder with no sales history yet still gets its owed units ordered, on the slowest route you have switched on, rather than a blank because the rate is zero.

On the Replenishment card this shows as an amber chip in the header, “Preorder: 300 already sold, added to this order”, a “Preordered” line beside “In stock”, and a sentence in the order's explanation saying those units are already sold and that the first units to land go to them. We built it after getting it wrong in both of the directions above, and the fix is pinned by a test so it cannot quietly come back.

Three limits, plainly. Restocio has no preorder or notify-me feature of its own. It plans the order behind the promise, and the promise itself, the button on the product page and the date beside it, comes from your theme or a preorder app. Restocio also only started writing the switch down, per product, on 2 September 2026, and Shopify reports it as it stands today, so days before a product's first recording are judged by today's setting. That is right for a switch that has not moved and converges over the 60-day window for one that has. And the stock-on-a-date report reads a negative figure as zero, because there was nothing on the shelf; the owed units live on the Replenishment card, not in that report.

Once the owed units are served the product is an ordinary one again, and the ordinary questions return. How much cover to hold is in how many days of stock should you hold, and why the boat that will serve those customers is missing from your Shopify totals in the meantime is in incoming inventory, and the stock you cannot see.

Common questions

Does continue selling when out of stock make my Shopify inventory negative?

Yes. Shopify keeps counting down through zero, and its own inventory reports page gives the example of a product set to 5 that takes eight orders in a day and ends at minus 3. The negative figure is the number of units sold beyond the shelf. A product with tracking switched off goes negative too, and a till sale can go through at zero regardless of the switch, so check which of the three you are looking at.

Should I keep selling when out of stock if the stock is on the way?

Yes, if you can name the day it lands and that day is on the product page before the customer pays. The switch is not the risk. The missing date is. A purchase order with a confirmed ship date gives you a landing date to promise. Goods you have not ordered yet give you nothing to promise, however sure you are that you will order them.

How long can I take orders for a product I do not have?

Until the date you promised. If you promised none, the default in both the United States and the EU is 30 days: after that, in the US you must offer the buyer a delay or a refund, and in the EU the customer can set a further deadline and then cancel for a refund. Thirty days is shorter than one sea crossing, so an importer selling without a date on the page is past the limit before the boat is halfway. Write the date down and the window is whatever you wrote.

Why does Shopify still show sold out with continue selling turned on?

Usually locations. Shopify assigns stock per location, and when the location that fulfils online orders is at zero or below while the stock sits at one that does not, the item shows as out of stock online whatever the switch says. Check the variant as well, since the switch can be set per product or per variant and a product-level look can miss one size that still has it off.

Does negative inventory mean my stock count is wrong?

Not under continue selling. The shelf really is empty and the units below zero really are sold, so the figure is right and correcting it to zero would lose the customers waiting. A count that is wrong looks different: units the admin says you have that the shelf does not, which comes from what was typed when goods arrived. That is drift, and it has its own fix.

Does Restocio handle preorders?

It plans the order behind them. Restocio reads each variant's inventory policy from Shopify, splits a negative quantity into the empty shelf and the units owed, adds the owed units to the recommended order, and keeps a zero-stock day in the sales rate when the product was allowed to sell. It does not sell the preorder itself. There is no preorder button, no date on your product page and no back-in-stock email from Restocio; those come from your theme or a preorder app.

Sources

Every source below was opened and read on the date shown against it. Rules, rates and schedules change, so check the current page before you act on a number that matters. If anything here is out of date or wrong, tell us and we will correct it.

Shopify Help Center, Selling out of stock products (read 29 September 2026)
Shopify Help Center, Inventory reports (negative ending quantity) (read 29 September 2026)
Federal Trade Commission, Mail, Internet, or Telephone Order Merchandise Rule (16 CFR Part 435) (read 29 September 2026)
EUR-Lex, Directive 2011/83/EU on consumer rights, Article 18 (delivery) (read 29 September 2026)
Your Europe (European Union), Shipping and delivery (read 29 September 2026)

Marcus Volsted

Co-founder, Restocio

Marcus co-founded Restocio and works on it daily with a Swedish importer who plans their purchasing in it every working day. Restocio is built in Sweden by two founders, one Swedish and one Danish, and Marcus is the Danish one. Why we are building it.