How Old Is Your “In Stock”?

    Every in-stock claim has an age. Learn why real time inventory accuracy is really about source and freshness, and the rule JENNi uses for search.

    October 23, 20264 min read

    Every in-stock claim has an age, whether a site shows it or not. The badge says "In stock," and it feels like a fact about right now. It's really a reading someone took a while ago.

    People talk a lot about real time inventory accuracy. We think the more useful question is smaller: how old is this reading, and where did it come from?

    At JENNi we keep the source and freshness of every store offer. And search only returns items that are in stock. This post explains why we built it that way.

    Stock is a reading

    A shelf doesn't report itself. Somebody or something measures it: a register sale, a nightly file, a page that gets checked, a count by hand. Each measurement happens at a moment. After that moment, the real shelf keeps moving and the reading doesn't.

    So "in stock" always means "was in stock when we looked." The gap between then and now is the age of the claim.

    Age matters more for some items than others. A pallet of paper towels can sit for days. The last two of a popular toy can go in an hour. The same label, "in stock," carries very different risk.

    Most systems hide the age. They show a clean badge and a confident button. That works until it doesn't, and the failure lands on the shopper.

    Not all sources are equal

    Where a reading comes from changes how much weight it deserves.

    • The store's own system is close to the transaction, so it tends to follow what's really on hand.
    • A public storefront shows what a store chooses to display. It can lag, and it may only say whether something is available at all.

    Neither is bad. They're different, and a system that treats them the same is throwing away information. That's why every store offer in the SKU Graph carries its source next to its price and stock.

    It also carries when we last saw it. A fresh reading from a strong source is worth more than an old reading from a weak one, and the graph should be able to say so.

    "Available" and "how many" are different facts

    This one trips people up. A public page often tells you only that an item is available. It doesn't tell you a count. "Available" is a yes. "Eleven on the shelf" is a number.

    They answer different questions. "Can I probably get this?" needs the yes. "Can I get four of these?" needs more. Treating a yes as if it were a count is how a shopper ends up with three of the four they ordered.

    So we don't blur them. A store offer says what we know and how we know it. If we only know it's available, that's what we record.

    Our rule for real time inventory accuracy

    Here's the rule we hold to: search only shows what's in stock, for shoppers and agents alike.

    That means an out-of-stock item doesn't get listed as a maybe. It isn't in the results. A shopper browsing shop.jennipro.com doesn't see it, and an agent calling search_catalog through our MCP server doesn't get it back.

    We're careful with the claim, though. A product opened directly by its exact id can still be quoted, so we say "search," and we don't say "anywhere."

    Before any money moves, there's a second check. When a buyer opens the hosted checkout page, it re-verifies current price and availability. Checkout refuses stale prices, and nothing is charged until the shopper approves the final total.

    So the rule has two layers. Search keeps out-of-stock items from showing up. Checkout catches the case where a reading aged between the search and the purchase.

    What we don't claim

    We don't say our stock is real time, and we don't say we cover every store. No honest system can say the first, and no system we know of can say the second.

    What we can say is narrower and more useful. Every offer has a source. Every offer has a last-seen time. Search leaves out what isn't in stock. Checkout re-checks before it charges.

    And every order helps. When an order is delivered, the graph learns that the reading held. When one fails, the graph learns that it didn't. Both feed back in, which is how confidence gets earned instead of assumed.

    For retailers

    If your shelves feed digital demand, the question to ask your own systems is the one in the title. How old is the "in stock" you're showing? Where did it come from? Could you say so out loud?

    Want your shelves visible to AI agents? Get in touch at jennipro.com.

    Ready to connect?

    See how JENNi works for your business

    Whether you're a retailer with inventory, a brand, or building a channel — we'll show you exactly how the network fits.

    See the API