Ecommerce Sales Analysis: What Your Order Export Really Says
By PlainSight — Insightful Actions · Updated July 2026 · ~7 min read
Store dashboards are built to make you feel good: big revenue number, cheerful upward chart. Your raw order export is more honest. It tells you whether growth came from more customers or bigger baskets, how much of it you refunded, and how much you bought with discounts.
The export
Every major platform exports orders to CSV. Two files are ideal: an order-level export (one row per order, with date, total, discount, and customer) and a line-item export (one row per product sold). The first answers customer questions; the second answers product questions.
What to read
1. Orders versus revenue
What good looks like: both moving in the same direction.
When revenue grows but order count is flat, you are selling more per customer — efficient, and worth understanding so you can repeat it. When orders grow but revenue lags, your basket is shrinking or discounting is deepening. The divergence is the insight.
2. Average order value
What good looks like: stable or rising, and read after refunds and discounts.
AOV computed on gross revenue flatters you. Compute it on what you actually kept. Bundles, thresholds for free shipping, and genuine cross-sells move this number; sitewide discounts move it the wrong way.
3. Product concentration
What good looks like: a strong core without total dependence on one SKU.
Rank products by revenue and by units. If a single product carries an outsized share, your business is exposed to one supplier, one ad set, or one trend cycle. That is not necessarily wrong — but it should be a known risk, not a surprise.
4. Repeat purchase rate
What good looks like: a real and growing share of orders from returning customers.
Group orders by customer. Repeat purchase is the cheapest revenue in ecommerce and the clearest signal of genuine product-market fit. A store where nearly every order is a first order is renting its growth from ad platforms.
5. Refund and return rate
What good looks like: low, stable, and not concentrated in one product.
Refunds hit twice: lost revenue and the cost of shipping and handling it. Track by product — a single item with an outsized return rate usually has a sizing, description, or quality problem you can fix directly.
6. Discount dependency
What good looks like: a modest share of orders carrying a code, without revenue collapsing between promotions.
Compute what share of orders used a discount and what share of revenue that represents. If sales fall off a cliff between promotions, you have trained your customers to wait — and undone your own pricing.
7. Seasonality and the quiet months
What good looks like: a pattern you plan inventory and cash around.
Group revenue by month across as long a history as you have. The point is not the peak, which you already know about, but the trough — that is where cash gets tight and where promotions or new products actually earn their keep.
Three things worth checking
- Growth that is entirely first-time buyers — a treadmill, and it stops the moment ad costs rise.
- One product returning far more than the rest — often a fixable listing problem rather than a bad product.
- Revenue that only appears during promotions — you are buying sales rather than earning them, and the underlying price is no longer believed.
A monthly rhythm
Export the last full month at order level, read orders versus revenue, AOV after discounts, repeat rate, and refunds. Choose one number to move and one change to make. Ecommerce rewards small compounding improvements to conversion, basket, and retention far more than any single campaign.
Let PlainSight read your order export
Upload your store's order export and PlainSight surfaces AOV, product concentration, trends and top performers in seconds, benchmarks against typical ecommerce performance, and writes plain-English next steps — all in your browser.
Try it free on your own numbers →
Frequently asked questions
- Which export should I use?
- An order-level CSV for customer and basket questions, and a line-item CSV for product questions. Most platforms produce both from their orders or reports area.
- Should I use gross or net revenue?
- Net, after discounts and refunds, for anything you plan to act on. Gross is useful only for comparing against your platform's own dashboard.
- How do I measure repeat purchase rate?
- Group orders by customer identifier or email and count how many customers placed more than one order in the period. Longer windows give a truer picture than a single month.
- Is my customer data safe with PlainSight?
- Yes. Your file is processed entirely in your browser and never uploaded. Optional AI features send only anonymized summary totals, never customer names or emails.
This guide is general information for business owners, not financial, tax, or legal advice. Figures described as “typical” are rules of thumb that vary by market and model — always read your own numbers in context.