How many days of warning do you actually need before stock runs out?

Start with your supplier's quoted lead time, then add the days it takes you personally to place the order, then add the days between a pallet arriving and the stock being sellable. For most Shopify merchants that comes out at the supplier number plus five to ten days. A supplier who quotes 21 days needs an alert that fires at roughly 28 to 30 days of cover, not 21.

The reason is simple. The lead time your supplier quotes starts when the purchase order lands in their inbox. It does not start when stock gets low, and it does not start when you notice. Everything between those two moments is yours, and it is almost always the part merchants leave out of the maths.

The four spans in a warning window

Only one of these belongs to your supplier.

• Noticing. How long between a SKU crossing its threshold and a human seeing it. If you check inventory on Monday mornings, this is up to seven days.
• Deciding and placing. Checking the last order, asking whether the supplier has stock, getting sign-off on the spend, sending the PO. One day on a good week, five on a busy one.
• Supplier lead time. Production plus transit plus customs. The number you were quoted.
• Receiving. Delivery to counted, checked and live on the storefront. Rarely zero.

Classic inventory planning already names the first two. Lead time is conventionally split into preprocessing, the time required to release a purchase order from the moment you learn of the requirement, then processing, then postprocessing, the time to make a purchased item available in inventory after you receive it. Most reorder-point advice quietly drops the first and the last and keeps only the middle.

Work it out on one SKU

Take a candle variant selling four units a day. The supplier quotes 21 days from PO to your door. You approve purchase orders when you sit down with the numbers, which in a bad week is six days after the alert. Receiving and counting takes two days.

Your real warning window is 29 days, not 21. At four units a day, that is 116 units of cover, not 84. If your alert fires at 84, you are placing the order eight days late every single time, and you will be short for roughly a week before the delivery lands.

Now run it on a SKU with a 60-day overseas lead time. The same six days of personal latency barely matters as a percentage, but it still costs you 24 units. On short lead times, your own delay dominates. On long ones, demand variance dominates, because you are forecasting eight weeks out from a sales average.

Firing early costs something too

This is a real trade-off, not a free setting. An alert that fires 40 days out on a 21-day supplier means you carry a couple of extra weeks of stock on every reorder. That is cash sitting on a shelf, warehouse space, and on a slow-moving SKU it is the start of a markdown.

The costs are not symmetric, though. Carrying a few extra days of cover on a steady seller is a known, bounded expense. Running out of a hero SKU takes the margin, the ad spend you were pointing at that product page, and the customer who buys the alternative elsewhere. Merchants who catch the reorder point early can avoid that, and the padding is usually the cheaper of the two mistakes.

Where we would land: pad the window to cover your realistic ordering delay, not your best-case one. Measure the delay honestly for a month before you decide.

Where to put your own delay

You have two places to put those extra days, and only one of them is worth the effort.

The clean textbook answer is to keep lead time pure and add a separate safety-days figure. The practical answer, if your tool takes a single lead time number per SKU, is to enter the full span from "the alert reaches me" to "the stock is sellable". For a 21-day supplier with a weekly ordering rhythm, that is 28 to 30. Write the reasoning down somewhere, because in six months you will wonder why that field says 29.

Restoc's reorder point engine uses a merchant-input lead time per SKU, so this is exactly the field to put the padded number in. It also uses a configurable sales velocity window of 7, 14 or 30 days. That choice changes the answer as much as lead time does. A 7-day window reacts fast to a promo or a TikTok spike and overshoots on a quiet week. A 30-day window smooths both and will underfeed a SKU that is genuinely accelerating. For steady staples, use 30. For anything you are actively promoting, use 7 and watch it.

What the alert looks like when it works

The point of getting the window right is that the alert arrives while you still have options. You can email the supplier, ask about a partial shipment, or pay for faster freight, all before the shelf is empty.

When Restoc launches, it is built to do three things with that number. The dashboard shows current stock, days of inventory remaining and reorder status for every active SKU, so you can see which SKUs are inside their window right now. Email and Slack alerts fire when a SKU crosses its reorder point, delivered through an idempotent background job so one crossing produces one message rather than a daily nag. A weekly reorder report lists everything below its reorder point with a suggested order quantity, viewable in the app and exportable as CSV for the PO.

The 60-day free trial exists for the same reason this article does. It guarantees at least one full weekly reorder cycle completes before the first billing charge, so you see the numbers against your own catalogue before you decide whether the window was right.