How to Forecast Sales for a Small Shopify Store

A small Shopify store can forecast sales well enough to plan with from three simple inputs: what it sold in the same period last year, adjusted for how the business has grown since; its current run-rate; and the events it already knows are coming. Put the result down as a range, not a single figure, because the honest answer is always 'somewhere between'. This guide walks through each method with one running example, explains why any forecast is an estimate, and shows how to use the range for the two decisions that depend on it: how much stock to buy and how much cash to keep.

Which sales figure should you forecast?

Net sales, and the same one every time. Shopify's net sales are gross sales less discounts and returns, before shipping and tax are added. That is the figure closest to what the business earned from selling products, and it is not inflated by tax you pass on.

How does the 'same period last year' method work?

Take last year's sales for the period and multiply by a growth factor. Last year's figure carries the season with it: the holiday peak, the summer lull, whatever shape your year has. The growth factor updates it for the size of the business today.

To find the growth factor, compare a recent stretch with the same stretch a year earlier.

Here is the running example, with made-up round figures. Say you are forecasting November.

• Net sales last November: $40,000.
• Net sales for the last three months: $90,000.
• Net sales for the same three months last year: $75,000.
• Growth factor: 90,000 ÷ 75,000 = 1.2.
• Forecast: 40,000 × 1.2 = $48,000.

The weakness is that it assumes this year's season will have the same shape as last year's. If last November included a one-off, such as a press mention or a clearance you will not repeat, take it out of the base first.

When is a run-rate forecast better?

When you have less than a year of history, or the business has changed so much that last year is no guide. A run-rate forecast takes recent average daily sales and multiplies by the days in the period.

In the example, say the last four weeks averaged $1,400 a day. Over a 30-day November that gives 1,400 × 30 = $42,000.

A run-rate knows nothing about seasons. It answers 'what happens if next month looks like this month', which is right for a steady business and wrong for one heading into its peak or out of it.

How do you adjust for known events?

List what you already know will be different, and add or subtract each one with a written reason. The usual candidates:

• a sale or promotion, or the absence of one you ran last year;
• a product launch;
• a change in ad budget;
• an expected stockout on a product that sells well;
• the calendar itself, such as five weekends in place of four, or a holiday moving.

Size each adjustment from your own past where you can. If last year's equivalent promotion added a certain amount over a normal week, start from that. Where you have no history, as with a new product, keep the adjustment small and widen the range instead.

Why give a range and not one number?

Because one number hides how uncertain it is, and the uncertainty is what you plan around. Three figures are enough: low, expected and high.

In the example, the two methods already give you a spread. You might set:

• Low: $42,000. The run-rate, as if the season barely lifts sales.
• Expected: $48,000. Last year adjusted for growth.
• High: $52,000. The expected case plus a launch that goes well.

A wide gap between low and high is information in itself. It says your data is thin or your month depends on one event, and that commitments should be cautious.

Why is every forecast an estimate?

Because each method assumes the past continues, and for a small store the past is noisy. One wholesale order, one post that travels, one week out of stock on the best seller, and a month moves by more than any method would have predicted. Small numbers swing more, in proportion, than large ones, so the fewer orders a store takes in a month, the wider its misses are likely to be.

A forecast still earns its keep by making assumptions explicit, so that when sales differ you know which assumption failed. Keep a simple log of forecast against actual for each month. After a few months you will see whether you lean optimistic or cautious, and can correct for it.

Software can produce the baseline for you. Marka Metrics includes sales forecasts on its paid plans, and labels them as estimates.

How do you use the forecast for stock and cash?

Use different ends of the range for different decisions.

Stock. Running out during a peak costs sales you cannot recover, so buy core products towards the expected-to-high end, and keep risky or new lines nearer the low end. Convert revenue to units by product, then check each against its reorder point. Our guide to calculating a reorder point takes it from there.

Cash. Fixed commitments, such as rent, salaries and ad contracts, should be affordable on the low case. Remember the timing: stock is usually paid for before the sales it supports arrive, so a strong forecast month can be preceded by a tight one.

Targets. A forecast is what you expect; a target is what you are aiming for. Keep them as separate numbers. The guide to setting monthly sales targets shows how to build one from traffic, conversion and order value.

A short checklist

1. Forecast net sales, by month.
2. Start from the same period last year multiplied by a growth factor.
3. Cross-check with a run-rate from recent weeks.
4. Adjust for known events, each with a written reason.
5. State a low, expected and high figure.
6. Buy stock towards the upper end and commit cash on the low end.
7. Log forecast against actual every month.

Forecasting sits alongside the other planning gaps covered in our guide to Shopify analytics beyond the built-in reports. If you want a baseline forecast to start your range from, have a look at Marka Metrics.