How to Set Ecommerce Inventory Reorder Points

A reorder point becomes misleading when the underlying inventory count mixes units that can be sold with units that cannot. Shopify separates inventory into states including available, committed, unavailable, and incoming. Available inventory is sellable and not committed to an order. Committed units belong to placed orders. Unavailable units may be reserved, damaged, under quality control, or held as safety stock. Incoming units are moving through a transfer, purchase order, or app workflow and do not become available merely because a supplier says they shipped.

Use available inventory as the starting quantity for a reorder decision, then show incoming supply as a separate dated line. This prevents an unreceived purchase order from masking a near-term stockout. It also prevents safety stock or damaged units from being counted twice. For a multi-location operation, make the calculation at the location level before rolling it up. A network can appear healthy in total while the location serving a high-demand region is below its own threshold and routing orders from farther away.

Lead time should cover the complete interval between approving a replenishment decision and making the received units available for sale. Include internal purchase-order approval, supplier processing, production when applicable, transportation, customs or appointment delays when applicable, receiving, inspection, and putaway. Measuring only the carrier transit leg understates exposure. Use actual timestamps from recent purchase orders when they exist, and keep the median and slower cases visible instead of choosing the fastest delivery as the plan.

Estimate demand across that same interval using the SKU's own order history and known operational changes. A simple starting point is average daily unit demand multiplied by replenishment lead-time days. Keep promotions, product launches, stockout periods, and discontinued variants labeled because they can distort the average. Shopify's demand-forecasting documentation describes using recent order history, comparable historical periods, or a chosen target inventory level. Those are inputs to review, not proof that one forecast method fits every catalog.

Safety stock is the explicit buffer above expected lead-time demand. It protects the customer promise when actual demand is higher than forecast or replenishment arrives later than planned. The buffer should be tied to observed variability and the consequence of a stockout. A stable replenishable item with a reliable domestic supplier may need a smaller buffer than a seasonal item with irregular demand and long inbound lead times. Document the chosen units and the reason so the team can challenge the assumption later.

Do not treat safety stock as available inventory. Shopify identifies safety stock as one reason units may be held in an unavailable state. That distinction matters because a threshold can otherwise consume its own buffer on paper. Set an exception rule for when safety stock may be released, who approves the release, and how it will be replenished. If the team repeatedly consumes the buffer, the reorder point, lead-time assumption, or order quantity needs review rather than another manual override.

The reorder point answers when to act, while the purchase quantity answers how much to buy. Keep those decisions separate. When available inventory plus credible incoming supply falls to the threshold, calculate an order quantity that restores the intended coverage without ignoring supplier minimums, case packs, storage capacity, shelf life, cash constraints, or an upcoming assortment change. A supplier minimum can force inventory above the desired maximum, so record that constraint instead of presenting the resulting quantity as pure demand.

Use a purchase order to preserve the commercial agreement and a linked transfer or receiving workflow to track the physical movement. Shopify documents purchase orders as records of products, quantities, costs, payment terms, and supplier details. It also distinguishes an ordered purchase order from the linked inventory transfer used to track movement. That separation gives the operator a clean audit trail: what was requested, what the supplier confirmed, what entered transit, what was received, and what became available.

A reorder workflow is not complete when the purchase order is sent. Reconcile received quantities with the ordered quantities, record damaged or missing units, and update availability only for units that passed the receiving process. Shopify's receiving guidance distinguishes accepted, rejected, and still-incoming quantities. That evidence prevents the planning system from assuming the entire order is sellable when part of it is delayed, damaged, or rejected.

Review the exception list weekly. Look for SKUs below the reorder point, overdue inbound orders, repeated partial receipts, demand shifts, stockouts during lead time, and excess inventory after receipt. Assign an owner and a next action to each exception. The cadence should also identify stale reorder points: a threshold calculated from an old sales pattern is not safer merely because it is automated. Recalculate after material demand changes, supplier changes, location changes, product launches, promotions, or repeated overrides.

Keep the operating record small enough that a buyer, warehouse lead, and finance owner can read the same decision. For each SKU and location, record available units, committed units, unavailable units, dated incoming units, average daily demand, approved lead-time days, safety-stock units, reorder point, supplier minimum, proposed order quantity, and the evidence date. Add an exception column for a launch, promotion, discontinuation, supplier delay, or unusual return pattern. This table does not replace the commerce or warehouse system. It explains which system values were used and why the resulting action was approved.

Test the control with a bounded group of important SKUs before applying it to the whole catalog. Compare the planned reorder date with the actual approval date, supplier confirmation, shipment, receipt, acceptance, and sellable date. Review where the model was early or late without declaring one period conclusive. A useful canary proves that the data states reconcile, the owner receives an actionable alert, and the purchase-order trail closes after receipt. Once those controls work, expand by supplier or product family while retaining the same definitions.