Reorder Points and Safety Stock: The Math Behind No Stockouts

The two ways to get reordering wrong

Every inventory decision fails in one of two directions. Order too late or too little and your bestseller goes out of stock, the listing loses rank, and the sale walks to a competitor. Order too early or too much and your working capital sits frozen on a shelf, ageing toward markdown. Most Indian D2C brands run on a founder’s gut and a spreadsheet, which means they oscillate between both failures, often on the same SKU across a quarter.

The fix is not more instinct. It is a small piece of arithmetic that turns your own sales history and supplier behaviour into a defensible reorder trigger. It is not advanced, and it pays for itself the first time it prevents a bestseller stockout during a sale.

The reorder point, defined

The reorder point is the on-hand quantity at which you fire a purchase order. Get it right and stock arrives just as the shelf runs low. The formula has two parts.

The first part is lead-time demand, the units you expect to sell while you wait for replenishment. If a SKU sells 40 units a day and your supplier takes 15 days, you will burn 600 units before new stock lands. If you reorder at 600, you arrive at exactly zero, with no room for a bad week or a slow supplier. That is why you add the second part.

The second part is safety stock, the buffer that absorbs the days demand runs hot or the supplier runs late. So the reorder point is lead-time demand plus safety stock. Everything hard is in sizing that buffer.

Sizing safety stock properly

Holding a flat number of extra days for every SKU is the common shortcut, and it is wrong. It overstocks steady sellers and under-protects volatile ones. The sound method sizes safety stock from three inputs: how much daily demand varies, how much supplier lead time varies, and the service level you choose.

Service level is a business decision you convert to a Z-factor. A 95 percent service level uses a Z of about 1.65, and 98 percent uses about 2.05. The higher the service level, the more buffer, and the more cash tied up. Safety stock rises with the standard deviation of demand and, crucially, with the standard deviation of lead time multiplied by average daily demand. In plain terms, an unpredictable supplier costs you more buffer than choppy demand does.

• Demand variability: the spread of daily unit sales around the average.
• Lead-time variability: the spread of actual delivery times around the quoted one.
• Service level: your chosen probability of not stocking out in a cycle.

Why lead time is the Indian bottleneck

For brands sourcing domestically or importing, the supplier is usually the wild card. A vendor who promises 20 days but actually delivers anywhere between 15 and 40 forces a large buffer, because the formula weights lead-time variability heavily. Two levers shrink it. Tighten the vendor, through firm timelines, penalties, or a second source, so the range narrows. Or accept the range and pay for it in safety stock. Measure each supplier’s actual versus promised lead time over the last ten orders and you will see which vendors are quietly costing you working capital through their unreliability.

Imports add customs and port variability on top. Build that into the lead-time number using real clearance history, not the freight forwarder’s optimistic quote.

Tier your service levels

Do not hold every SKU to the same standard. Classify by margin and role. A high-margin hero product that anchors your ads and rank deserves a 98 or 99 percent service level, because a stockout there is expensive twice over, in lost sales and lost ranking. A low-margin, slow-moving long-tail SKU can sit at 85 to 90 percent, because the cost of holding a deep buffer outweighs the rare missed sale. This tiering is where the math protects cash rather than just preventing stockouts. You spend your buffer budget where it earns the most.

Keep the inputs fresh

A reorder point is only as good as the numbers feeding it, and those numbers drift. Recompute at least quarterly, and always before a festive peak. During the Great Indian Festival or Big Billion Days window, daily demand can run three to five times baseline, so a reorder point set on off-season sales will strand you within days. Build a peak scenario with elevated demand and, often, longer supplier lead times as everyone reorders at once.

Two guardrails keep the system honest. First, your on-hand data must be accurate, because a reorder point triggers off a stock number you trust. Cycle counting earns its keep here. Second, watch for demand trends, not just averages. A SKU on a steady climb needs a rising reorder point, and a static formula will always leave it short.

You do not need expensive software to start. A clean sheet with per-SKU average daily demand, demand standard deviation, average and variable lead time, and a chosen service level will compute reorder points that beat instinct comfortably. As SKU count grows past a few hundred, move it into your inventory or ERP tool so it updates on live data. The brands that win the availability game are not the ones with the best hunches. They are the ones who turned reordering into arithmetic, tiered it by what each SKU is worth, and refreshed the inputs before the season changed.