What does overselling cost you a month?

I built this after watching my own count drift

I sell online as well as building sku.io, and this summer we had our own oversell problem. Our system and our 3PL's system had stopped agreeing about what was on the shelf, so we kept selling things we thought were in stock. On a call with the 3PL in August I said it out loud:

"I synced it two months ago and now we're way out of sync again. So I need to sync it and then watch it like a hawk to see like, why are we getting out of sync? Because otherwise we're going to keep overselling."

Some of those orders were for stock we physically owned that could not ship to the customer for up to a month, because of a restriction on the 3PL side. Those turned into unhappy customers, and into me re-syncing and then trying to work out why the two systems drifted apart again. None of that has a line in the P&L.

Two other calls, with people outside my own business. An operations lead running two Shopify stores, TikTok Shop, retail and wholesale out of one pool, back in April:

"Traditionally, we've just opened up the taps and then each channel sells as much as it can. And unless somebody's watching it like a hawk. Right. It's very common that we end up overselling by, you know, a thousand units on a channel."

And a consultant who implements inventory software for multichannel brands, on flash sales:

"They disconnect the [IMS] and the integration while the flash sales are going on so they don't oversell. Like it's really crazy manual and then anytime it oversells, we're talking could be hundreds of orders they have to cancel."

If your oversells come in bursts like that, put in the month with the sale, not an average month.

What the calculator does not count

The customer who never orders again, and the listing a marketplace suppresses when your cancellation rate crosses its line. Amazon's line is under 2.5% of seller-fulfilled orders over 7 days. Walmart's is 2% or below over 30 days (Walmart's seller performance standards). I have no way to price either for a store I have not seen, so the result panel shows your cancellation rate next to those two limits and nothing more.

The margin slider is the one judgment call in it. A late order still ships, so only cancelled orders lose their margin. If you think a cancelled customer just buys again next week, slide it down. For first-time buyers I would leave it alone.

About buffers

The usual answer is a buffer. Show 80% of stock on each channel, or hold back five units. It works. It also means a slice of stock you own is never offered for sale, and on a best-seller that shows up as a stockout on a channel while units sit on the shelf.

My view is that a buffer belongs on a channel that genuinely lags, like a marketplace that takes a while to accept an update or a flash sale with orders landing faster than any sync. Everywhere else it is covering for a count that is wrong. Fixing the count means one system owns the stock number and publishes availability to every channel, instead of each channel keeping its own idea of what is on the shelf. That is how sku.io handles channels: it is the master of record, with buffer rules you set per product and per channel for the places that need one. There is more on the prevention side in how multichannel retailers end overselling.

When the number is too small to act on

If you sell on one channel and the calculator lands at a couple of hundred dollars a month, do not buy inventory software to fix it. Turn on your store's own stock tracking, or look at Zoho Inventory or inFlow, which are built for that size. Software like ours earns its keep when the number above is in the thousands, it is spread across several channels, and the cleanup time goes into working out which count to believe.