Safety stock

Demand and supply are never perfectly predictable. Safety stock absorbs that variability: a sudden rush, a supplier who ships late, or two orders landing on the last unit at once. Without a buffer, those normal events turn into overselling and orders you cannot fulfil.

The trade-off is a little tied-up inventory in exchange for not selling units you cannot ship, which is usually the more expensive mistake. Tied-up is the right way to think about it too, because that buffer is cash sitting on a shelf at your cost of goods sold, and what your stock is worth is the number that tells you how much.

Two formulas, and which one you want depends on how much data you have.

The quick version covers most small stores: multiply your average daily sales by your supplier lead time in days. A product selling four a day with a five-day lead time needs about twenty units of buffer, which keeps you selling for the whole time a reorder is in transit.

The version that accounts for things going wrong uses the spread of your demand rather than the average: safety stock equals Z times the standard deviation of daily demand times the square root of lead time. Z is a service factor you choose, and the point of it is that a product whose sales swing between one and fifteen a day needs a much bigger buffer than one that reliably sells four, even when both average four.

A middle option if standard deviations are more maths than you want: take your worst week rather than your average. Multiply your highest daily sales by your longest observed lead time, then subtract average daily sales times average lead time. It lands close to the statistical answer and you can do it from a sales export.

The service factor sets how often you are willing to run out. Common values are 1.28 for a 90% service level, 1.65 for 95%, 2.05 for 98% and 2.33 for 99%.

The cost curve is the thing to understand: moving from 95% to 99% roughly doubles the buffer for a four-point gain. That is worth it on a hero product where a stockout loses a customer permanently, and wasteful on a slow-moving accessory where nobody minds waiting.

Set it per product rather than store-wide. Most catalogues want a high level on the top ten sellers and a low one on the long tail.

These get used interchangeably and they are not the same thing. Safety stock is the quantity you refuse to sell below. The reorder point is the level at which you place a new order, and it sits higher, because it has to cover the sales that happen while you wait for delivery.

The relationship: reorder point equals expected sales during lead time plus safety stock. Selling four a day with a five-day lead time and twenty units of buffer gives a reorder point of forty, because twenty units get sold while the delivery is in transit and twenty are the cushion underneath.

The practical difference is what each one triggers. Hitting the reorder point should send you to your supplier. Hitting the safety-stock level should stop the store selling, because at that point you are selling units that are already promised to the gap.

Partly, and this is worth getting right because most articles say flatly that it does not. Open a product in your admin and the stock at each location is split into groups. Available is the only one your store can sell from. Beside it sits Committed, which is stock that already belongs to an order you have taken, and an Unavailable group that includes Safety stock along with damaged units and anything held for inspection. Move units from Available into Safety stock and your On hand total does not change, but what the store can sell does.

What that gives you is a parking bucket, not a rule. It is a one-time manual move: nothing recalculates it as stock sells, nothing raises it when a product starts moving faster, and nothing puts units back when a delivery lands. You are doing arithmetic by hand in a supported place, which is better than the usual workaround but is still arithmetic by hand.

The usual workaround is worse. Under-counting, telling Shopify you have 80 when you really have 100, breaks the moment your warehouse, your 3PL or a stock count writes the real number back, and it makes every inventory report wrong in the meantime. At least parking units in Safety stock leaves your On hand figure honest.

The reliable version is an app that holds a floor against your true count. Easy Stock Buffer stops the order at checkout once your Available count reaches the floor you set, so the floor still holds when someone pays through Shop Pay, PayPal or Google Pay instead of the normal checkout. Those buttons jump straight from the product page to payment, which is why a check that only lives in your theme misses them. See how to set a safety stock level on Shopify for the step-by-step.

It is not free, and there are cases where the buffer costs more than the stockout. Slow movers are the clearest: holding twenty units of something that sells once a month ties up cash for two years to prevent an event nobody would have noticed.

Made-to-order and print-on-demand products do not need it either, because there is no finite pool to protect. Neither do high-value items where a customer will happily wait two weeks for the right thing, and where the buffer would sit as dead capital.

The products that earn a buffer are the ones that sell steadily, get reordered often, and where a stockout sends the customer somewhere else rather than into a waiting list.

A buffer is also the wrong tool for a problem that looks similar. Capping how many units one customer can put in a basket protects a launch from a single buyer clearing the shelf, which is a different job with a different setting, and order limits against safety stock puts the two side by side.