Ecommerce inventory planning: replenish, hold or promote

Ecommerce inventory planning is deciding, product by product, what to replenish, hold, promote or deprioritise. Forecasting units is only the input; the plan is the decision.

I worked the method through four illustrative products with every formula shown: profit, stock capital and availability read side by side, then one decision rule that turns them into an action.

Which numbers does ecommerce inventory planning use?

Ecommerce inventory planning reads three lenses per product: profit through GMROI, stock capital through excess capital at cost, and availability through revenue at risk. POAS then says whether advertising the product pays.

GMROI is gross margin return on investment: the gross profit each euro held in stock at cost earns in a year. Excess units are sellable units − (weekly demand × weeks to next delivery + safety stock), never below zero.

Units at risk are max(0, weekly demand × weeks to next delivery − sellable units). Revenue at risk is an estimate of exposure, not a booked loss, and stockout cost covers how to estimate it.

Why one lens on its own misleads

Each metric is blind to what the other two measure. GMROI on its own says “buy more” for a product that is about to run out.

Excess inventory on its own says “promote”, even when every ad euro on that product loses money. Revenue at risk on its own says “reorder”, even for a product that only sells through unprofitable ads.

POAS, profit on ad spend, settles the ad question. A POAS of 1.0 is break-even: each euro of ads brings back one euro of profit.

What inputs do you need for each product?

You need eight inputs per product, all for the same date, and each with one precise meaning.

With no order placed yet, weeks to next delivery is the weeks until you order plus the supplier’s lead time. Cover is sellable units ÷ weekly demand, the same idea as days of inventory counted in weeks.

Size the buffer before you judge excess: the safety stock calculator works it out from demand and lead-time variation. Count a purchase order only when its date and quantity are firm.

How do you decide which products to replenish, hold or promote?

Decide on stock position first, then on POAS against break-even: a short product with POAS at or above 1.0 gets replenished, a short one below it gets its economics fixed first.

No product can be in both top rows. Excess starts only above demand to the delivery plus safety stock, so a product with units at risk has none.

GMROI sets the order within a cell, not the cell. The buying decision stays with you.

Which products should you review first?

Start with the products that have the most euros at stake, short or tied up in stock. With hundreds of products, ABC analysis gives the order.

ABC analysis ranks products by one value, such as gross profit, and labels the few that make up 80% of the total as A. The ABC analysis calculator splits your own export into A, B and C.

Worked example: four products through the decision rule

For example, take four products, a trail jacket, a canvas tote, a bottle holder and a merino beanie, and run each lens in turn. All figures are illustrative data.

Profit: GMROI and POAS

GMROI runs from 4.00 for the trail jacket to 0.50 for the bottle holder, and only the jacket and tote have a POAS at or above break-even.

Show the dataHide the data

GMROI per product against the weighted shop GMROI

Product	GMROI
Trail jacket	4.00
Canvas tote	2.00
Merino beanie	1.50
Bottle holder	0.50

The shop’s GMROI is €51,900.00 ÷ €19,600.00 = 2.65. Averaging the four product ratios gives 2.00, because it weights the €1,600.00 beanie the same as the €9,000.00 jacket. Sum first, then divide.

Show the dataHide the data

Share of shop gross profit and of stock at cost per product (%)

Product	Share of gross profit	Share of stock at cost
Trail jacket	69.4%	45.9%
Canvas tote	23.1%	30.6%
Merino beanie	4.6%	8.2%
Bottle holder	2.9%	15.3%

Availability: cover and revenue at risk

Two products run out before their next delivery: the trail jacket with 30 units at risk and the merino beanie with 48.

The jacket needs 10 × 5 = 50 units before its delivery and has 20. The beanie needs 64 and has 16.

Together that is €5,939.22. Valued at gross profit instead, the same units come to €2,699.70 and €575.52, €3,275.22 in total: an alternative view of the same units, never a figure to add.

Stock capital: excess units at cost

The canvas tote and the bottle holder hold €4,878.00 of excess capital at cost between them, and the other two hold none.

The tote’s 225 excess units are 15 weeks of its demand; the bottle holder’s 216 are 72 weeks.

Show the dataHide the data

Revenue at risk against excess capital at cost per product (€)

Product	Revenue at risk	Excess capital at cost
Trail jacket	€4,499.70	€0.00
Canvas tote	€0.00	€3,150.00
Bottle holder	€0.00	€1,728.00
Merino beanie	€1,439.52	€0.00

The verdict per product

Each product lands in a different cell of the matrix.

To apply the method to your own products, download the inventory planning worksheet (XLSX, free, no sign-up). It opens in Excel and Google Sheets and uses the same formulas and decision rule as this example.

Each row gets a suggested review, its reason and what to check first, such as season, returns and how firm the order is. Break-even POAS is a setting, 1.0 by default.

Four products take a few minutes. A shop with 600 products needs a forecast, a delivery date and a cost for each, refreshed every time stock moves, and a spreadsheet is out of date by the time it is filled in.

Product Metrics reports GMROI, excess capital and revenue at risk per product and rolls them up by category, brand and price range. Inventory Insights adds projected demand, lead time, safety buffer, stock cover, reorder dates and slow movers with a reason.

What should you do with each product?

Each case has one action and one move to avoid: the move its best number invites.

Replenish: the trail jacket

Reorder the trail jacket now, because demand before the delivery is 50 units and only 20 can ship. Bring the delivery forward if the supplier can, and set the next order with the reorder point calculator.

Lower its ad priority until the delivery lands. Extra clicks sell no extra units when demand already exceeds stock; they only empty the shelf sooner.

Product Segmentation sets ad priority per product with labels in Google Merchant Center. In Inventory Insights, a stock label you approve pauses a product at zero stock and resumes it when the restock lands.

Fix the economics: the merino beanie

The merino beanie sells out, but with a POAS of 0.9 every euro of ads returns €0.90 of profit. Its gross profit is €11.99 on a €29.99 price, 40%, against 60% for the jacket.

Check the price and the unit cost first, then lower its ad priority so fewer of its sales come from loss-making clicks. Reorder once the economics hold.

Hold and promote: the canvas tote

Skip or shrink the tote’s next order: it has 15 weeks of excess on a 3-week wait. With a POAS of 1.6, its ads pay, so give it more exposure.

Rerun the numbers before a promotion. If it doubles demand to 30 a week, for example, excess falls from 225 to 180 units: 300 − (30 × 3 + 30).

Deprioritise: the bottle holder

The bottle holder holds 72 weeks of excess and earns a GMROI of 0.50, so buy none and lower its ad priority, since its ads lose money.

Stock that has stopped selling altogether needs a different plan, covered in dead stock inventory.

When should you recalculate the plan?

Recalculate each time stock, the demand forecast or a purchase order changes, and before every big promotion. A plan built on last month’s stock goes wrong in six common situations.

For seasonal products, compare GMROI with the same product’s earlier years, not with products sold all year.

High inventory turnover products reach their reorder point fastest, so they need the most frequent check.

Final thoughts

The products that need a decision rarely look wrong on their best number. The jacket’s GMROI hides its shortage; the tote’s healthy POAS hides 15 weeks of surplus.

Start with your ten largest products by gross profit. Work out cover, excess and POAS for each, place them in the matrix, and act on the cell rather than on the single metric that looks best.