How to forecast demand for your Shopify store (a practical guide)

Demand forecasting sounds like data science, but at its core it is one question: how many units will I sell before my next shipment arrives? Get it wrong in one direction and you stock out and lose sales. Get it wrong in the other and you freeze cash in stock that sits. You do not need a model with a hundred variables to do this well, you need a repeatable method and clean data.

Start with clean sales history

Pull at least twelve months of unit sales per SKU if you have it, so you capture a full seasonal cycle. Then fix the single biggest trap: do not forecast off days you were out of stock. A month that looks low because you sold nothing for ten days is not low demand, it is missing data. Estimate what you would have sold across those days and use the adjusted figure.

The simplest method: a moving average

Average your recent sales to forecast the next period. Say a SKU sold 80, 95, and 110 units over the last three months:

3-month moving average = (80 + 95 + 110) ÷ 3 = 95 units

It is crude, but for a stable, year-round product it is often good enough, and far better than guessing.

Weight the recent months more heavily

A plain average treats a sale from three months ago the same as last week. If demand is trending, weight recent periods higher. Using weights of 0.5, 0.3, and 0.2 from newest to oldest:

(110 × 0.5) + (95 × 0.3) + (80 × 0.2) = 55 + 28.5 + 16 = 99.5 ≈ 100 units

The weighted forecast of 100 reflects the upward trend that the simple average of 95 misses.

Add seasonality

Most stores are not flat across the year. Calculate a seasonal index for each period by dividing its average sales by your overall monthly average. If December typically sells twice an average month, its index is 2.0. Forecast a normal baseline, then multiply:

December forecast = 100 baseline × 2.0 index = 200 units

This is how you avoid the classic mistake of running out in your single biggest month because you planned off a flat average.

Adjust for trend and known events

History is the starting point, not the whole story. Layer in what you already know: a planned promotion, a paid-ads push, a new sales channel, a price change, or a product going viral. A forecast you can override with judgment beats a black box you cannot.

Turn the forecast into a reorder

A forecast on its own does not place an order. Combine it with your lead time and a buffer to get a trigger point, the work we cover in reorder points vs. safety stock. Then, when cash is tight and you cannot reorder everything at once, rank the candidates by return on inventory investment so your money funds the SKUs that pay back fastest.

Common forecasting mistakes

• Forecasting off a flat average and ignoring seasonality, then stocking out in peak season.
• Counting stockout days as low demand, which quietly trains your forecast to under-buy.
• One method for every SKU. A steady staple and a spiky seasonal item need different treatment.
• Setting it once. Demand and lead times drift; re-forecast on a schedule.

When to let software do it

A spreadsheet works for a handful of SKUs. Across a few hundred, with seasonality, trend, and stockout adjustments per item, it becomes the task that gets skipped under pressure. A dedicated tool, see our comparison of the best Shopify inventory apps, does it automatically. InventoryIQ forecasts demand per SKU from your real Shopify history, then ranks the resulting reorders by ROI under your cash runway. Start a free trial and see your own numbers.