A minimum stock level is a promise to your future self: below this number, order. Set it well and the low-stock list is a short, honest shopping list. Set it badly and it becomes forty lines of noise that everybody learns to scroll past — which is worse than having no alerts at all, because now you are not watching and you think you are.
Start with lead time, not with feeling
The only number that matters is this: how much of this product do we get through between placing an order and the order arriving? That is your floor. Everything else is padding on top of it.
Work it out roughly. If you use six bottles a day and the supplier takes three days, you burn eighteen bottles waiting. A minimum of eighteen means you hit zero exactly as the delivery lands — which is too tight, so you add a buffer.
The buffer is for the bad week, not the average one
Averages are the enemy here. Your usage is not six a day; it is four on a Tuesday and eleven on a Saturday. Size the buffer for the version of the week that actually hurts, and be honest that some products deserve a bigger buffer than others.
Three things that make a minimum level wrong
- It was set at launch and never revisited. Your menu, your range and your suppliers have all changed since. The number has not.
- It was copied from the old system. Imported minimums carry over assumptions from a business that may no longer exist.
- It ignores the ordering unit. If the supplier ships in cases of twelve, a minimum of five tells you to order when you were always going to order twelve anyway. Round to something you can actually buy.
Not everything needs a minimum
This is the part that keeps the list readable. Set minimums on the products where running out costs you something a customer notices — the menu item, the bestseller, the part that stops the line. Leave them off the long tail. A product you sell twice a month does not need to be on a weekly alert; it needs to be on a quarterly count.
A rough split that holds up: minimums on perhaps a fifth of your catalog, covering most of your movement. If more than half your products have a minimum, the list will be too long to act on.
Let the count tell you if you were right
Set the numbers, then leave them alone for two counting cycles. After that you have real evidence rather than a hunch: which items hit their minimum and were ordered in time, which ran out anyway, and which have been sitting three times above their minimum for a month and could stand to come down.
In trakr, the minimum level lives on the product alongside cost, price and barcode, and the weekly low-stock summary lands by email and push. The summary is the test. If people read it, your minimums are set about right. If they stopped reading it, you have too many.