When to reorder: days of cover vs supplier lead time

Most stockouts are not surprises. They are mistimed reorders. Knowing when to reorder inventory is less about holding more safety stock and more about comparing two numbers on every line, every day: how many days of cover you have left, and how long supply actually takes to arrive.

Days of cover in plain terms

Days of cover tells you how long the stock you can actually sell will last at the pace you are actually selling it. The formula is simple: on-hand quantity divided by average daily sales over a recent window, usually the last four to twelve weeks.

Two details decide whether the number is useful. First, use sellable units. Quantity in quarantine, reserved for a contract, or sitting in a damaged bin is not cover. Second, subtract open customer backorders and add goods already in transit. A line with 10 days of cover and 14 days of lead time is fine if a purchase order lands on day 8. Without that PO, it is a stockout scheduled for next week.

Lead time from history, not from the supplier card

Every ERP has a lead time field on the supplier or product record. It was filled in once, often by the supplier's own promise, and rarely reviewed. Real lead time lives in your purchase orders: the gap between the order date and the goods-receipt date, per supplier and per SKU.

Compute the median and the spread for each supplier-SKU pair over the last year. A supplier whose deliveries land in 12 to 16 days is predictable. One whose deliveries land in 9 to 31 days needs a buffer, no matter what the card says. The spread matters as much as the average, because a reorder timed to the median will be late half the time.

The reorder point formula, worked through

The classic reorder point formula combines expected demand during lead time with a safety buffer that covers demand variability. For a 95% service level the safety factor is 1.645.

Reorder point = average daily sales × lead time + 1.645 × σ(daily sales) × √lead time

Take a mid-velocity line: 12 units a day on average, a standard deviation of 5 units a day, and a 14-day lead time from receipt history.

The quantity to order follows from the cover target, not from the reorder point. If the category targets 60 days of cover, the order is roughly 12 × 60 = 720 units plus the 31-unit safety stock, rounded to the supplier's pack size and minimum order quantity. A line you are exiting gets a shorter target; an A-class line with a volatile supplier gets a longer one.

A simple rule for when to reorder inventory

Reorder when days of cover < lead time + receiving buffer + policy buffer, after accounting for open purchase orders. The receiving buffer covers the days between the truck arriving and the stock being pickable. The policy buffer is where judgment lives: three days for lines with a service-level commitment, zero for lines you are running down.

The rule is deliberately boring. Its value comes from being applied to every SKU on every refresh, which no buyer can do by hand across 2,000 lines.

Common mistakes

• Reordering on a calendar rhythm. "We place orders on Mondays" means a line that crosses its threshold on Tuesday waits six days. Let cover crossing the threshold trigger the draft, and batch the sending if the supplier requires it.
• Ignoring open POs and backorders. Cover math without in-transit stock double-orders; without backorders it under-orders.
• One global safety stock field. A single "minimum quantity" for SKUs with different velocity, margin, and supplier reliability is wrong for most of them.
• Reading the trend too late. Average daily sales lags a demand shift. Pair the reorder rule with the four-week versus twelve-week velocity trend so an accelerating line is caught early.
• Sorting the list alphabetically. Rank by revenue at risk and by the stock risk score, then cap the queue to what the dock can receive.

How Flowra handles this

Flowra recalculates cover, lead time, and the reorder point for every SKU on each nightly refresh, or on demand after a new upload. Lead time comes from your purchase-order history, not the master-data field. Stock is reconstructed from net transaction flow, so a clean opening count is not a prerequisite, and the result is labelled as a reconstructed position rather than a physical count. Lines whose days of supply fall at or below one lead time are flagged critical; those within two lead times are flagged reorder soon. Each flag arrives as a drafted purchase proposal with its evidence, delivered to the buyer in Slack, Teams, or email. Nothing is written back to the ERP unless the connector supports it, your organisation has opted in, and the approving user holds the right role. See which ERP and file sources feed the calculation.