How to Analyze a Shopify Order Export in Excel with AI

Step 1: Export the orders

In your Shopify admin, find the orders list and look for an export option. It downloads your orders as a CSV file for the date range you choose. Menus change, so follow Shopify's own help page for the current steps. Choose a range long enough to be useful, such as the last 90 to 180 days, because trends need time to show.

You can open the file in Excel to look at it. You do not need to edit it first.

The traps in an order export

Order exports have a few quirks that break a spreadsheet analysis if you do not know about them:

• One row per line item, not per order. An order with three products takes several rows, and in the standard export the order-level fields, such as the order total, may appear only on the first of them. Counting rows over-counts orders. Count unique order numbers, and read order-level fields such as the total from the first row of each order.
• Taxes and shipping are separate columns. Decide whether your revenue figure includes them before you compare months.
• Refunds may live in a separate status or file. If you skip them, every revenue figure is overstated.
• No product cost. The standard export does not know what a product cost you. To compute margin you need to add cost per SKU, from your own cost sheet.
• Blank discount code does not mean full price. It means no code was used. Automatic discounts and other price changes can still be in the discount amount.

The exact columns differ by store and by export setting, so look at the header row before you start.

Step 2: Ask instead of pivoting

Upload the CSV to the Analyst and ask your questions in plain English. The Analyst reads every row, flags data problems first, and writes the answer with a chart where it helps. If you would rather stay in Excel, ask for the formula and paste it in.

A sample store as a guide

The Analyst's sample file is an order export from Cascade Supply Co., with 24,623 rows across 15 columns and 180 days of orders. It is a cleaned export with one row per order, and it shows what a good answer looks like:

Those six rows are a better first meeting with your own data than any dashboard total.

Six questions to start with

1. "What are gross sales, orders and average order value by month?" This sets the baseline.
2. "Which products have the lowest margin?" It finds the products that look fine on revenue and are not.
3. "Which discount codes cost the most, and how deep do they cut?" Discounts are the first leak to look at.
4. "What share of customers order more than once?" In the sample store it is 23.4%.
5. "Which channel brings the most orders, and which brings the best ones?" Volume and value often disagree.
6. "What are the biggest outliers in this file, and why?" Outliers are either errors or the most interesting rows.

Ask them in order and use each answer to sharpen the next. If an answer looks wrong, ask for the rows behind it, which lets you see what was counted.

When to stay in Excel

Keep a workbook when the numbers feed a live model that other people edit, or when you need to trace each cell. Use the AI for the first read, the unusual-row hunt and the charts. The two work together: the AI tells you which pivot to build, and you can ask it to write the formula to build it.

Try it

Open the sample file to see the Analyst answer these questions without signing up. The Analyst page has the 14-day trial. For the full list of store metrics worth tracking, read 12 e-commerce metrics that matter, and the e-commerce analytics page shows what else it reads from an order export. For the basics of asking questions of a spreadsheet, see how to analyze Excel data with AI.