How can I set reorder points without perfect demand data?

Begin with the decision the number will trigger

A reorder point is useful only if someone knows what to do when inventory reaches it. Define whether the threshold creates a review task, a purchase recommendation, or an approved replenishment action. Identify the person responsible and the supplier or transfer route available. A precise threshold with no operational owner does not prevent a stockout.

Choose the inventory position you will compare with the threshold. Physical stock alone may ignore open commitments and confirmed incoming supply. Be consistent about what your calculation includes. Document whether you are working at a single location, across a shared pool, or for a specific channel allocation. Those choices affect the meaning of the number and should be settled before adding more sophisticated forecasting methods.

Estimate ordinary daily demand from a usable period

Use a recent period that is long enough to contain normal variation and short enough to reflect the current business. Review units by sellable variant, not only revenue or product-family totals. Check whether the item was available throughout the period. Ten units sold during two in-stock days does not imply demand of one unit per day across a ten-day window.

Mark promotions, large one-off orders, price changes, and new channel launches. You may choose to analyze those separately from ordinary demand rather than letting one unusual event dominate the baseline. Shopify's inventory analytics includes days-remaining measures derived from average sales, but any average depends on its observation window. Keep your own period and exclusions visible so the team understands what the estimate represents.

Measure replenishment lead time from the right starting point

Lead time should cover the interval from the decision to replenish until the goods can actually be sold and fulfilled. Supplier production, transport, receiving, inspection, labeling, and put-away can all matter. A delivery appointment is not necessarily the moment stock becomes available.

Use recent completed receipts to estimate the normal range rather than relying only on a supplier's quoted transit time. Separate predictable processing time from uncertain delays. If different suppliers or shipping methods behave differently, keep distinct assumptions. A product with a short nominal lead time but frequent receiving delays may need a different policy from one with a longer but very consistent replenishment cycle.

Use a small worked example to make the rule understandable

Imagine a fictional variant that sells about eight units per day and takes five days to replenish into sellable stock. Expected demand during lead time is forty units. If the team chooses an additional sixteen-unit safety allowance based on its uncertainty and service priorities, the initial reorder point is fifty-six units.

Those figures are an example, not a recommended buffer for every business. The important feature is that each part can be explained and changed. If observed demand rises to ten units per day, or receiving begins taking seven days, the threshold should be reviewed. Keep the calculation beside the operational rule so a colleague can understand why the threshold exists instead of inheriting an unexplained number in a spreadsheet.

Now suppose the same variant has twenty units available, thirty units confirmed incoming, and no additional commitments outside the calculation. The planning position may already reflect fifty units, depending on the method you have defined. That does not automatically mean a new fifty-six-unit purchase is appropriate. The threshold signals a review; the order quantity still depends on the intended coverage and existing supply. This example shows why the inventory-position definition and purchasing decision must remain visible beside the formula. Otherwise staff may react to a low shelf count by placing a second order for stock that is already due to arrive.

Choose safety stock to address specific uncertainty

Safety stock is a deliberate allowance for uncertainty, not an arbitrary percentage that makes the formula look complete. Consider demand variation, supplier reliability, counting accuracy, and the consequence of running out. A readily substitutable accessory may justify a different approach from a variant that is essential to a larger order.

Avoid using extra stock to hide a process problem indefinitely. If records are inaccurate because returns are not processed, improve the records as well as considering a temporary buffer. If lead time is uncertain because a supplier never confirms dispatch, improve that information flow. A useful safety allowance is tied to a reason, reviewed against outcomes, and adjusted when the underlying uncertainty changes.

Separate when to reorder from how much to buy

The reorder point tells you when to act. The reorder quantity answers how much to replenish. Those decisions are related but not identical. Order quantity may depend on supplier minimums, case packs, storage limits, available cash, shelf life, and the period of demand you want the next receipt to cover.

Do not automatically buy the same quantity as the reorder point. That can create repeated small orders or excessive stock depending on the product. Consider the current inventory position and confirmed incoming orders before creating another replenishment request. Keep open purchase orders visible so a threshold alert does not produce duplicate buying. When incoming quantities change, update the planning view rather than assuming the original purchase order will arrive exactly as requested.

Handle new and seasonal products with explicit judgment

A new product may not have enough history for a stable average. Use a comparable item, a small initial order, supplier flexibility, and a more frequent review while you learn. Record that the estimate is provisional. The absence of history should lead to a controlled learning plan, not a falsely precise forecast.

Seasonal products need a different view of time. A trailing average from quiet weeks may understate an approaching peak, while a peak-period average can overstate demand after the event. Use a defined seasonal scenario and an end-of-season plan. Check the timing of replenishment carefully: stock arriving after the main selling window may not solve the problem the purchase was intended to address.

Review the rule against the outcomes it produced

At an agreed interval, compare the threshold with actual stockouts, emergency purchases, excess stock, and supplier performance. Ask whether the cause was a demand surprise, lead-time change, inaccurate record, or a missed action. Changing the formula will not fix an alert that nobody owns or a purchase request that waits unreviewed for a week.

Adjust one meaningful assumption at a time where possible, and preserve the reason for the change. Over time, the rule should become better informed and easier to operate. You do not need perfect demand data to start making better replenishment decisions. You need a transparent calculation, an accountable response, and a habit of checking whether the rule is producing the outcome you intended.