An out-of-stock page gets treated as a failure state. Something went wrong upstream, a customer arrived at the wrong moment, and the job now is damage control. That framing costs brands more than the lost sale does, because the sold-out page is the one place on your site where a customer tells you precisely what they want, at the price you're asking, with no discount attached, and receives nothing in return. That's not a dead end. That's the cleanest piece of first-party demand data your brand owns.
Consider what makes it clean. Wishlist saves are cheap and ambiguous. Add-to-cart is polluted by browsing behavior and price-checking. Survey responses tell you what people say they'd buy. A back-in-stock signup costs the customer something real — they handed over a contact detail to get a product they cannot have right now. There's no other reason to do it. And the signal is unbiased by your own merchandising, because nobody signed up as a result of a homepage feature or an email push. The product wasn't available to promote. Demand arrived on its own.
Start with the most obvious thing that data can tell you and that most brands never extract: variant-level demand. Aggregate product demand tells you to reorder the style. Variant-level demand tells you the size curve, the shade ratio, the scent split, the color mix. Most brands buy roughly the same ratio they bought last time, sell through the middle sizes, and then discount the tail for the rest of the season. Your notify-me list is a free forecast for that exact split, built from people who were ready to pay full price, and it exists before you commit a dollar of inventory capital.
It tells you about price too, if you read it right. How quickly a list builds, and how deep it gets before it stalls, is a rough measure of how much slack sits in your pricing. A variant that quietly accumulates hundreds of waiting customers at full retail while receiving no promotional support is telling you that price is not the constraint on that product. Meanwhile a product that sold out fast and then generated a thin list wasn't a hit — you simply bought too few of something ordinary, and buying deeper into it is a mistake dressed up as a data-driven decision.
The problem in most organizations isn't the data, it's the geography. The notify-me list lives in the marketing platform, owned by the person writing the alert emails. The person deciding next season's buy is in a spreadsheet, working from last year's sell-through and a supplier deadline. Those two people often never speak about this. Fixing it costs nothing: a weekly export, ranked by variant, with signup dates attached, dropped into the buying meeting. It is probably the cheapest forecasting improvement available to a growing brand, and the mechanics of the alerts themselves are worth getting right alongside it — BMO's rundown on back-in-stock alerts covers that side properly.
One thing will limit how much of this signal you actually capture: a large share of people who land on a sold-out page will not give you an email address. They're on a phone, mid-scroll, and typing an address into a field to be told about a product later is more friction than the moment can support. Those people leave, and their demand never gets counted, which means your sample is skewed toward customers who were already willing to join your list.
That's the case for capturing the signal in more than one way. Web push notifications take a single tap, no typing, no inbox, and no commitment that feels like a relationship — which both widens the number of people you can alert and, more valuable long term, widens the demand sample you get to read. Two customers wanting the same variant should both register, whether or not one of them wanted to be on your email list.
Here's the practical first step. Export the last 90 days of notify-me signups, broken out by variant, and set it beside your most recent purchase order. If the two documents disagree, you already know which one was written from evidence.
Top comments (0)