Inventory Planning Guide: How to Build an Inventory Plan

An inventory plan is a series of bets.

You are betting on what customers will buy, when they will buy it, which stores or channels will see the demand, and how much inventory risk the business should take to capture those sales.

The math matters. But the job is not simply calculating reorder points or copying last year's inventory into a new spreadsheet.

A useful inventory plan connects expected demand to inventory commitments:

Demand forecast → sales plan → inventory targets → receipts and purchases → allocation and replenishment → actual performance → reforecast

Each step changes the next one. If demand comes in differently than expected, inventory targets need to move. If a supplier pushes a delivery out three weeks, the receipt plan changes. If ecommerce starts selling through a style twice as quickly as stores, allocation should change before one channel stocks out while another sits on excess units.

That is what makes inventory planning an operating process rather than a budgeting exercise.

Start With Demand, Not Last Year's Inventory

Historical sales are useful. Historical inventory is useful too. Neither should dictate the new plan without interpretation.

The starting question is: What do we reasonably expect customers to want?

Last year's numbers can give you a baseline, but retail rarely repeats itself cleanly. Promotions move. Weather changes. Stores open and close. Ecommerce takes a larger share. A fashion silhouette that worked last year may already be losing momentum. A basic that looked mediocre may actually have been constrained by poor availability.

Oracle's retail forecasting documentation specifically calls out recent trends, seasonality, out-of-stocks and promotions as factors that need to be considered in demand forecasting.

Stockouts are one of the easiest ways to misread history.

Suppose an apparel retailer sold 400 units of a style last season. At first glance, 400 looks like the demand signal. But the style was out of medium and large for two weeks during peak selling.

The retailer did not necessarily have demand for 400 units. It had sales of 400 units under constrained availability.

There is a meaningful difference.

The same problem works in reverse with promotions. If a product had an unusually aggressive discount or a major marketing push, blindly carrying that sales rate into the next period can inflate the forecast.

Good planning requires cleaning up those signals rather than treating the POS file as objective truth.

The planner's job is to build the most credible view of unconstrained future demand available, knowing it will still be wrong in places. From there, the business decides how much inventory risk it wants to take against that demand.

That order matters.

Forecast what customers are likely to want first. Then decide how much stock you are willing to own.

Turn the Sales Forecast Into Inventory Targets

A demand forecast tells you what you expect to sell. It does not tell you automatically how much inventory you should carry.

That requires another set of decisions.

Set targets based on the role of the inventory

Planners usually have several ways of looking at inventory coverage.

Weeks of supply or forward weeks of supply tells you how long current and incoming inventory should support expected demand. Inventory turnover tells you how effectively inventory capital is being converted into sales. Sell-through matters heavily for seasonal and fashion inventory because the selling window is finite. Beginning and ending inventory targets help connect one planning period to the next.

None of those metrics should be managed in isolation.

A style sitting at eight WOS might be a serious problem if it has four weeks left in its productive selling window. Eight WOS on a year-round replenishment basic with a reliable vendor is a completely different situation.

This is why blanket inventory rules tend to create bad plans.

A replenishable black T-shirt, a holiday sweater, a fashion dress and a specialty item with intermittent demand should not all carry the same coverage.

For each, the planner is balancing several things: demand confidence, margin, lead time, replenishment frequency, lifecycle, minimum order quantities and the cost of being wrong.

The consequences of a stockout matter too. Running out of a fashion item near the end of its season may be preferable to carrying excess units into markdown. Running out of a proven core item every Friday because the safety stock target is too lean is usually just lost productivity.

Add uncertainty rather than simply adding more stock

Safety stock exists because forecasts and supply chains are imperfect.

But "add 20% to everything" is not much of a safety stock strategy.

A SKU with volatile demand and a 12-week supplier lead time needs different protection than a stable replenishment item available from a domestic supplier in seven days. The buffer should reflect uncertainty in demand and supply.

This is ultimately a service-level versus inventory-cost decision. Oracle's inventory optimization methodology frames the problem similarly, setting replenishment policies at item/location level based on trade-offs between service levels and inventory costs.

Every inventory target is therefore partly a financial choice.

Higher targets can protect availability. They also consume working capital, increase carrying costs and create more aging and markdown exposure when the forecast misses.

There is no universally correct WOS target. There is only a target appropriate to the risk you are taking.

Build the Receipt Plan and Open-to-Buy Budget

Once you know what you expect to sell and how much inventory you want to carry, the next question becomes practical:

What needs to arrive, and when?

At a simplified level:

Planned receipts = planned sales + planned ending inventory + planned reductions − beginning inventory

The formula is straightforward. The inputs deserve more attention.

Beginning inventory is what you expect to own when the period starts. Planned sales consume units or inventory value during the period. Reductions account for markdowns and other inventory adjustments. Planned ending inventory is what you want left to support future sales.

The receipt plan fills the gap.

Timing matters as much as quantity. A receipt arriving after the demand peak is not equivalent to a receipt arriving before it. Anyone who has watched holiday inventory land in January knows that technically having enough stock is not the same as having useful stock.

Then there is open-to-buy.

OTB provides the financial control around what the business can still commit. Shopify's OTB guidance similarly uses planned sales, markdowns, beginning inventory and planned ending inventory to calculate available buying capacity.

But OTB is not the inventory plan.

It tells you how much additional inventory you can financially commit while staying within plan. It does not tell you which SKU deserves the money.

That distinction gets lost surprisingly often.

Imagine a fashion retailer with OTB available late in the season. A buyer could use it to reorder a collection that performed reasonably well. But the supplier has a six-week lead time, several slower colors are already carrying high WOS, and the productive full-price selling window is getting short.

Having the OTB does not make the reorder good.

The better decision may be to preserve some buying capacity and chase the products where demand is emerging more clearly.

This is why committing every available dollar before the season begins can be dangerous. You gain certainty on supply but lose flexibility.

In categories where demand is difficult to predict, some OTB is valuable precisely because you have not spent it yet.

Plan Below the Category Level or the Plan Will Hide the Real Problems

Aggregated inventory numbers can look healthy while the actual assortment is broken.

This happens constantly in apparel and footwear.

A footwear style might show six WOS overall. That sounds fine. Then you look at the size curve and discover the retailer is nearly out of the highest-volume sizes while carrying plenty of fringe sizes.

The spreadsheet says six weeks.

The customer sees a stockout.

Inventory planning therefore needs to move through the merchandise hierarchy:

Company → department/category → class → style/product → SKU/size → location/channel

Financial plans may begin at department or category level. Execution cannot stay there.

Size breaks are a good example of why.

If a T-shirt has 1,000 units available but most of those units are XS and XXL while M and L are depleted, style-level inventory overstates the health of the position. Customers do not buy "the style." They buy a particular color and size.

Once core sizes disappear, sell-through can slow even though plenty of units technically remain.

The same issue appears across locations.

Suppose ten stores collectively own the correct quantity of a SKU. Three high-volume stores are stocking out every week while seven lower-volume stores have excess coverage.

At company level, you may not have an inventory problem at all.

You have an allocation problem.

The appropriate response might be a store-to-store transfer, a change to replenishment logic or a different allocation on the next receipt. Buying more units without fixing the imbalance can simply add inventory to the network.

Oracle's inventory optimization framework operates at item/location level and includes purchase orders, transfers and store rebalancing intended to improve sell-through and reduce markdown exposure. Its merchandise hierarchy also explicitly extends through style, color and size.

Omnichannel makes this harder. Inventory may be reserved for ecommerce, shared across channels, fulfilled from stores or sitting in a distribution center while individual locations are short.

The point is simple: total inventory is not enough.

Inventory productivity depends on whether the right SKU is sitting where demand can actually consume it.

Manage the Inventory Plan as a Living Operating Process

The original inventory plan will be wrong.

Not necessarily badly wrong, but wrong somewhere.

That is normal. Forecasts are estimates. Purchase orders get delayed. Products surprise you. Promotions outperform or disappoint. A style that looked like a winner in week two can flatten by week five.

So the operating cycle has to continue:

Plan → buy → receive → sell → compare actual vs. plan → reforecast → adjust

The useful question in a weekly planning meeting is not simply, "How are sales?"

It is: What has changed enough that we should change an inventory decision?

A planner might review actual versus planned sales, forecast error, WOS or FWOS, sell-through, stockouts, inventory turns, aged inventory, on-order quantities, late POs, size availability and markdown exposure.

You do not need another dashboard full of red and green cells. You need to know what deserves action.

If sell-through is running materially ahead of plan and there is enough selling window left, the team might chase inventory.

If total inventory is sufficient but stores are diverging, reallocation may be the answer.

If WOS is climbing while sell-through weakens, there may still be time to cancel or reduce future receipts rather than waiting for the inventory to become a markdown problem.

If a style looks healthy overall but core sizes are disappearing, allocation and replenishment need attention before the top-line style metric catches up.

This is also where spreadsheet-heavy planning starts to strain. When planners are manually exporting sales, joining inventory files, checking POs and rebuilding size-level views, too much of the week goes into identifying what changed rather than deciding what to do about it.

The better model is exception-based planning.

Oracle's forecasting approach explicitly supports exception reporting because retailers can be dealing with volumes of forecasts that make manual review of every position impractical.

That is also the thinking behind Flagship. Predictive monitoring can keep watching demand, inventory, size-level availability and future risk while planners focus their attention on the positions that actually require judgment. The point is not to remove the merchant from the decision. It is to stop asking skilled planners to hunt through spreadsheets just to find the decision.

Because inventory planning is still judgment.

You are deciding where to put capital before you know exactly what customers will do.

The goal is not minimum inventory. It is not maximum availability either. Both can be expensive targets when pursued blindly.

A strong inventory plan puts capital behind the demand signals where the business has a good reason to take risk, protects flexibility where uncertainty is higher, and changes course when actual performance tells you the original bet was wrong.

That is the real discipline of inventory planning.