Product inventory management: prevent impossible orders

Product inventory management means knowing what exists, what can be sold now, and what has already been promised to a customer. A count that is correct at noon can be wrong minutes later if the online store, physical shop and returns process update at different times. This guide gives you a practical way to prevent impossible orders before buying another system.

Separate inventory states before adding numbers

A unit physically on hand is not always available to sell. An order may have committed it, a return may be waiting for inspection, and a delivered item may not yet be ready for fulfilment. Shopify distinguishes available, committed, unavailable and on-hand quantities in its inventory state definitions.

Define the unit for each SKU: individual item, pack, size or other variant. A “shirt” total of twelve tells you little if customers buy by size. Keep variant identifiers and locations consistent across your store, point-of-sale system and supplier records.

Write a plain-language availability rule. In many setups, sellable quantity is physical quantity minus committed and temporarily unavailable units, but use the definitions of your actual platform. Avoid inventing a second calculation that conflicts with the store's own state model.

Record every movement

Online and in-store sales, supplier receipts, returns, damaged items, transfers and manual adjustments all change stock. Assign an owner and recording time to each movement. Keep a basic reference: order number, delivery note, return reason or inspection result.

Returns need a separate check. An item coming back is not automatically ready to resell. Inspect it first. If every return goes straight into available inventory, the store may promise a damaged or incomplete product.

Transfers between a shop and warehouse can create a similar gap. An item may have left one location without arriving at the other. Track its in-transit state and receipt. A positive total across locations does not prove you can ship from the right place.

Test simultaneous demand

Choose a variant with one unit remaining. In a test environment or with a dedicated test item, simulate two customers adding it to their carts almost simultaneously. Observe when stock is reserved, whether both orders can complete, and what message the second buyer sees. Repeat across physical and online channels if they share inventory.

Measure the real sync delay. “Automatically synchronized” does not necessarily mean instant. If channels update after several minutes, a small safety buffer may be justified on fast-moving items. Size the buffer using sales velocity and the cost of overselling rather than applying one arbitrary number to the entire catalog.

Check the product page for an out-of-stock variant. Does the purchase button stop, show a truthful restock estimate, or permit a backorder under a stated policy? A page that says “available” while checkout refuses the order damages trust.

Use cycle counts to find recurring causes

An annual count finds repeated errors late. Group SKUs by value, sales frequency and error risk. Count important variants more often. Count physical units without first looking at the system's expected quantity, then compare the results.

Investigate differences rather than simply overwriting the number. Causes can include an unrecorded receipt, wrong variant, unchecked return, damaged item, transfer delay or theft. If the same error recurs, change the process that creates it.

Fictional example: a store finds four fewer units than expected. Two returns were assigned to the wrong size and two in-store sales were entered the next day. The fix is to inspect returns and record sales promptly, not merely to type a new total into the admin.

Connect stock to commercial decisions

Low availability may call for slowing a promotion, showing a reliable delivery date or placing a replenishment order. High availability is not automatically healthy if demand and margin are falling. Review sales velocity, supplier lead time, storage cost and returns alongside the stock count.

Set a reorder point for important products: expected demand during supplier lead time plus a buffer matched to uncertainty. Revisit it when seasonality or supplier performance changes. One fixed threshold will not work for every SKU.

If you sell bundles, make sure all components are available and that a bundle order reduces their individual stock correctly. Otherwise, the same item may be sold twice under different product listings.

Weekly control sheet

Start with ten SKUs responsible for many orders or incidents. Trace each from receipt through sale, then expand the method. Reliable inventory is a sequence of accountable movements, not just a number on an admin screen.

What to do after an impossible order

First confirm the actual item state and whether another location can fulfil it. Contact the buyer promptly with a reliable date, a genuinely comparable alternative or the cancellation route under your store's policy. Do not let an automated delivery promise continue if it is wrong. Internally, connect the incident to the order, variant and last inventory movement. Document the root cause and test the fix with a controlled order before treating the issue as closed.

Avoid a misleading stock dashboard

A dashboard showing “five units” may combine items across locations or states. Put the relevant location and availability state next to each number. For fast-moving variants, show the time of the last update too. If the store sells through a marketplace as well as its own checkout, check whether that channel reserves stock immediately or only after an order import. The displayed count is useful only if everyone knows when it was true and which units it includes.