A sold-out product page is not a dead end unless you build it as one. The customer arriving there has already done the expensive part — they found you, they found the product, and they decided they want it. What happens in the next ten seconds decides whether that work produces a sale later or is handed to whoever ranks second.
Three things stores do with sold-out products, and what each costs
Hide the product
Unpublishing the product removes the disappointment and takes the page with it. The URL 404s, the ranking it had built decays, and when the product returns you start again from a new or re-crawled page.
This is almost always the wrong choice for anything you intend to restock. The page is an asset; the out-of-stock state is temporary. Hiding is defensible only for genuinely discontinued lines, and even then a redirect to the closest replacement beats a 404.
Show it with a sold-out badge and nothing else
The page stays, the ranking stays, and every visitor leaves. This is the common case and it is the one worth fixing, because the page is doing all the work of attracting the customer and none of the work of keeping them.
Show it with something to do
The same page, plus a way for the customer to act on the intent they arrived with. That action is one of three, and which one is right depends on a single question.
The question that decides the instrument
Do you know when it is coming back?
- Yes, with a confirmed date and confirmed stock — preorder. Take the money now, state the ship date plainly, and let the customer stop shopping. You capture the sale rather than a chance at it.
- Probably, but you cannot commit to a date — a restock alert. You are asking the customer to wait, so the honest offer is "we'll tell you the moment it lands", not a date you will miss.
- No, or not at all — a substitute. Put the closest alternatives in front of them now, because there is nothing to wait for.
Most stores pick one of these and apply it to everything. The cost of that is quiet: preorders on uncertain stock produce refund requests and support load, and restock alerts on discontinued products produce a list you will never email.
Making the alert worth subscribing to
The widget itself is the easy part. Four details separate one that collects addresses from one that collects them at two or three times the rate:
Put it where the Add to Cart button was. Not below the fold, not in a tab. The customer's eye is already there and the muscle memory is already pointed at it.
Handle variants properly. A product that is available in three of five sizes shows no sold-out state at all, so the customer who wants the missing size has nothing to click. They are the most motivated person on the page and most stores drop them silently.
Say what happens next. "We'll email you the moment it's back, and you'll have first claim in the order you joined" answers the two questions the customer is about to ask support.
Be honest about dates. If you have a confirmed arrival, show the month. If you do not, say you do not. An invented date does not buy patience; it buys a complaint on the day it passes.
Selling the alternative without insulting the customer
Recommendations on a sold-out page are worth real revenue and they are easy to get wrong. The rule is that the alternative has to solve the same problem — the same use, the same size range, a comparable price. A £200 jacket replaced with a £40 one reads as a store trying to clear stock rather than help.
Show two or three, not twelve. And put them below the notify-me form, not above it: the customer came for this product, and the first thing you offer should be the version of the answer they actually wanted.
The restock is where the recovery is won or lost
Everything above only matters if the notification works. Two failures undo all of it.
Notifying more people than you have stock for. Four hundred emails about thirty units means three hundred and seventy people reach a sold-out page for the second time, by invitation. Those units sold either way; what you lost is the list's willingness to open the next one. Match the send to the stock, or send in waves and stop when it runs out.
Restocking without sending at all. Inventory goes up during a routine receiving process, the product quietly becomes buyable, and the list is never told. The trigger needs to be the inventory change itself, not somebody remembering.
What the sold-out page should still do for search
A page you keep is a page that keeps ranking, but only if it still reads as a real product page. Two things are worth getting right while the product is unavailable.
Leave the description, images and reviews in place. Stripping the page back to a sold-out notice removes the content that earned the ranking in the first place, and the page will drift down the results before the stock returns.
Keep the product's structured data honest. Shopify emits availability in the product schema, and marking something in stock while it is not is the kind of mismatch that costs you the rich result rather than saving the click. An out-of-stock product with accurate markup keeps its listing and recovers immediately when stock returns; one with wrong markup can lose the enhancement for longer than the stockout lasted.
Measuring recovery without flattering yourself
"Recovered revenue" is the number every tool in this category reports and the easiest one to overstate. Three rules keep it usable:
- Count the order, not the cart. Revenue from customers who were notified and then purchased, not the value of everything they browsed afterwards.
- Pick one attribution window and keep it. A long window credits the alert with purchases the customer had already decided on. A short one understates it. Consistency matters more than the number you choose.
- Do not add two systems together. If your email platform and your alert app both report restock revenue, they are reporting the same orders. Pick one as the source of truth.
The figure worth watching alongside it is purchase rate among people notified. Recovered revenue rises with list size; purchase rate only rises when the programme is actually working.
A short order of work
- Stop unpublishing sold-out products you intend to restock. This costs nothing and is the largest single recovery on this list.
- Add the notify-me form where Add to Cart was, and check it appears at variant level rather than only when every variant is gone.
- Write the widget copy once — what happens next, and whether you have a date.
- Add two or three genuine alternatives below it, matched on use and price rather than on whatever the recommendation engine suggests.
- Decide the send rule before the first restock: everyone, or matched to the units available.
- Pick your attribution window and write it down, so the number you report in three months means the same thing as the one you report today.
Where to go next
Stok captures restock and price-drop subscribers at the variant level and fires on the inventory change rather than a manual send. For choosing between waiting and paying now, see alerts or preorders; for the send-sizing problem, more people waiting than stock.
