Weeks of Supply for Imported Inventory: Build a Comparable SKU Runway

Calculate current runway, then model inbound stock by date

Weeks of supply is the number of weeks that a defined quantity of inventory would cover at a declared weekly demand rate:

Current sellable WOS = units available for sale now / demand units per week

The numerator and denominator must refer to the same SKU, unit, location or channel scope and calculation time. Choose the demand basis for the decision: use comparable trailing demand to describe runway at a recent observed rate, or a named forecast version to test a forward planning scenario. Do not rank the results together when their bases differ.

For an importer, the formula is the easy part. The useful work is deciding which inventory is genuinely available, which demand rate belongs in the denominator, and when inbound units could become available to sell. Record those assumptions beside every result.

Use two views rather than one inflated number:

1. Current sellable WOS describes today's available inventory at the selected location or channel.
2. Projected balance by period places each expected receipt in its dated week and subtracts the relevant demand scenario.

Do not add every open purchase order to today's stock. A shipment expected in six weeks cannot cover a shortage in week four. Its quantity, date and release state belong in a time-phased scenario, not in the current-stock numerator.

Weeks of supply is therefore a monitoring measure. It can show which SKU needs investigation, but it does not establish the right reorder point, order quantity, safety stock or excess-stock action on its own.

Define the numerator as stock available for the decision

Start from the inventory state that matches the question. If the decision is whether ordinary customer orders can be fulfilled now, use units that are available to sell—not every unit physically present or recorded somewhere in the network.

Shopify's current inventory-state documentation illustrates the distinction. In that platform, on-hand stock includes available, committed and unavailable quantities. Incoming inventory is separate and does not become available until receipt. Other systems use different labels, so map their statuses to your operating definitions before calculating.

Create a small data dictionary:

The table is a decision rule, not a command to rename source-system fields. Preserve the original inventory state and transformation so another reviewer can reproduce the numerator.

If a SKU is stocked across a 3PL, marketplace and local warehouse, decide whether the view is location-specific or network-wide. Use inventory snapshots from a consistent cut-off, reconcile transfer IDs and count a physical unit once across the network. A transfer can otherwise appear as stock at the origin and incoming at the destination. Network stock can look comfortable while the channel that owns the next customer demand is empty. A network view and a location view answer different questions.

Record the demand basis beside every result

The denominator determines what the answer means. Shopify's product analytics overview defines its days-of-inventory-remaining metric using ending inventory divided by average units sold per day during the reporting period. Amazon Connect Decisions documents days-of-cover inventory-policy inputs based on either forecast demand or historical demand. These are product examples, not proof that one window or method suits every importer.

For every WOS result, record:

• demand source: shipped orders, fulfilled units, adjusted sales history or a named forecast;
• time basis: units per day or units per week;
• history window or forecast version;
• locations, channels and SKUs included;
• treatment of returns, cancellations, samples and wholesale allocations;
• known stockout periods that suppressed observed sales;
• promotions, launches, delistings or one-off orders that distort comparison; and
• calculation timestamp.

Never divide inventory units by a demand rate expressed in another unit or time basis. Convert the rate explicitly—for example, a per-calendar-day demand rate multiplied by seven before a weeks-of-supply calculation—and retain the original value and calendar convention. Labelling different windows or forecast versions does not make their WOS values comparable; segment the dashboard or recalculate them on one approved basis before ranking SKUs.

Use trailing demand for a descriptive view

A trailing average answers: “How long would today's sellable stock last if the selected recent rate continued?” It is easy to reproduce and useful for a stable item. It can mislead when recent sales were constrained by a stockout, inflated by a promotion, or drawn from the wrong season.

Keep the raw period demand and excluded events visible. If a buyer overrides the rate, preserve both the original calculation and the reason for the override.

Use a forecast for a planning scenario

A forecast-based denominator answers: “How much time would this stock cover under forecast version X?” It can reflect known seasonality, promotions and channel plans, but it inherits the forecast's assumptions and uncertainty.

Use the existing demand-forecasting process for imported SKUs to create that input, then name the version in the WOS record. When forecast error or persistent over- or under-forecasting is material, route it to the forecast accuracy and bias review. Do not quietly change the denominator until the dashboard looks comfortable.

Keep two views instead of one inflated number

Some platforms report cover alongside shipment states. Amazon's current FBA report guide, for example, lists available inventory, shipment statuses, inventory age, excess units and weeks of cover within its inventory reporting. For an importer, that is a reminder to read the cover measure beside the status of the units—not an instruction to treat every inbound shipment as received.

Maintain these fields separately:

• current sellable WOS;
• next expected available date;
• expected available quantity;
• date and source of the latest receipt evidence;
• organisation-defined inbound evidence state;
• projected pre-receipt balance;
• projected post-receipt balance;
• base demand scenario and any named alternative; and
• assumption or evidence exception.

A simple “pipeline WOS” calculated as available plus every open PO divided by weekly demand can hide sequence. It also combines quantities with different evidence quality: one lot may be packed and booked, another may only have a requested ex-factory date, and a third may be held for inspection.

Project the balance in dated periods instead. Only place a receipt in the period when your scenario assumes it becomes available for the relevant demand. Keep that assumption linked to production, freight, receiving and release evidence. Where receipt timing is uncertain, show a base and a relevant late or held-release scenario rather than attaching an unsupported confidence percentage to one date.

Build the minimum weeks-of-supply dashboard

The following schema is OPL analysis and can be implemented in a controlled spreadsheet, inventory system or planning tool.

Show the formula inputs, not only a coloured result. A reviewer should be able to answer: “Which units and which demand created 6.4 weeks?” without searching another workbook.

Avoid a universal green/amber/red scale. The useful boundary depends on the SKU's replenishment path, review interval, demand variability, order constraints, substitution options and consequences of shortage or excess. Define organisation-owned review bands and document why they exist.

Turn unusual runway into named exceptions

An exception should identify the assumption or decision requiring attention. Useful categories include:

• pre-receipt shortfall: projected available balance reaches zero before the next assumed receipt;
• inbound evidence gap: the scenario depends on a quantity or availability date that is unconfirmed, stale or changed;
• demand-basis change: a promotion, stockout, wholesale order, launch or forecast revision makes the current denominator unrepresentative;
• state mismatch: stock counted in the numerator is reserved, quarantined or held in another location;
• zero or near-zero demand: the formula is undefined or produces a very large, low-value result;
• long-runway review: cover exceeds an organisation-owned investigation band and requires a separate cost, range or disposition decision; or
• data-quality break: negative stock, duplicate SKU mappings, unit-of-measure conflict or missing timestamp prevents a reliable calculation.

Give each exception an owner, next action, due date and evidence link. Do not convert “low WOS” automatically into “place a PO”. Use the reorder-point and safety-stock guide for that decision. If the issue is long runway, use the inventory carrying-cost method to investigate cost rather than assuming every slow SKU is unprofitable.

Use a review cadence that matches the decision horizon

Review often enough to act before the next relevant gate, but do not invent one cadence for every SKU. A fast-moving item with volatile demand and a near shipment decision may need a different cycle from a stable spare part with long cover.

At each review:

1. refresh available units from a timestamped source;
2. retain the prior WOS and input values;
3. update the declared demand basis only through its controlled forecast or history review;
4. move inbound quantities only when new evidence changes the assumed available date or state;
5. recalculate projected balances under the named scenarios;
6. open, update or close exceptions against evidence; and
7. route replenishment, allocation, quality or excess-stock decisions to their owning process.

Use ABC inventory analysis or another documented segmentation as context for attention, not as a substitute for current runway. A high-revenue classification does not tell you whether its inventory data or inbound assumption is correct.

Worked example: an inbound shipment changes the runway, not today's stock

This example is hypothetical and does not represent an OPL customer, supplier or recommended target.

An importer reviews SKU KT-240 at the start of Monday, before week 1 demand is deducted:

• 1,200 units are available to sell;
• 80 additional units are held for quality review and excluded;
• base demand forecast DF-0914-R03 is 150 units per week;
• 900 units are assumed to become available at the start of week 6 in the base scenario; and
• the arrival is a planning assumption linked to the latest shipment evidence, not a guarantee.

Current sellable WOS is:

1,200 available units / 150 units per week = 8.0 weeks

Adding the inbound quantity would produce 14.0 weeks, but that number hides when the units are expected. The time-phased base scenario is clearer:

Under the base assumptions, the projected balance remains positive before the receipt. Now test an approved promotion scenario of 250 units per week. The calculated net balance is 1,200 - (5 x 250) = -50 units at the end of week 5, meaning the scenario contains 50 units of unmet demand before the assumed week-6 receipt. It is not a claim that the physical inventory record will become negative. The right output is a pre-receipt shortfall exception, not an automatic order instruction. The owner must review promotion timing, allocations, replenishment options and the inbound evidence through their respective controls.

Also run a late-receipt or held-release case when that uncertainty matters. If the receipt assumption moves to week 7, update the scenario version and preserve the former date and evidence state. Do not overwrite the assumption without showing why the exception changed.

Handle zero demand, stockouts, launches and seasonality explicitly

Some SKUs should not receive an ordinary numeric WOS:

• Zero demand: division is undefined. Display no valid demand basis, not infinity.
• Very low demand: one sale can swing the result sharply. Show the raw units and consider a longer, documented view.
• Recent stockout: observed sales may understate demand because the product could not be purchased. Mark the affected period.
• New launch: history is absent. Use a named launch scenario and show that it is an assumption.
• Seasonal product: trailing demand from the wrong season can overstate or understate runway. Use an appropriate forecast scenario.
• Product transition: old and replacement SKUs may need a linked phase-out and phase-in view rather than independent cover targets.
• Returns or quarantine: units physically present may not support ordinary demand. Keep their state visible.

Amazon Connect Decisions documentation explicitly distinguishes forecast-based and history-based days-of-cover inputs. That supports exposing the chosen basis; it does not make either option automatically reliable for these edge cases.

Launch the view with one SKU family

Start small enough to reconcile every input:

1. select one SKU family and define the location or channel scope;
2. map source-system states to available, committed, unavailable and incoming quantities;
3. choose and document the demand basis for each SKU;
4. calculate current sellable WOS with a timestamp;
5. place each material inbound receipt into a dated scenario;
6. test base and relevant alternative demand assumptions;
7. create exceptions for timing, evidence, state or denominator problems; and
8. route each exception to its decision owner without letting the dashboard approve the action.

The result should make comparison honest. Every SKU runway has a reproducible numerator, a named denominator and a dated supply scenario. That is enough to focus attention without pretending one number can replace forecasting, purchasing or inventory judgement.

Sources

• Shopify Help: Product analytics overview
• Shopify Help: Understanding inventory states
• Amazon Seller Central: FBA reports
• Amazon Connect Decisions: Data entities for Supply Planning