A product business opens its dashboard and sees an average order value of $60. That sounds like the typical customer spends about $60. But the average may be carrying a different story: many $35 to $45 orders, a smaller group of larger bundles, and one wholesale or corporate purchase that pulls the number upward.
AOV does not describe the shape of your orders. Before changing an offer or production plan, compare the average with the median and the full order value distribution. Learn what most customers actually buy—not just what the dashboard averages.
Start with a consistent average order value
The basic formula is:
AOV = included order revenue ÷ included completed orders
Platform definitions differ. Shopify documents its sales-report definition using gross sales minus discounts divided by orders, with specific treatment for later adjustments. Document your definition before comparing reports. For an internal product-sales review, you might use merchandise revenue after discounts and refunds while excluding sales tax and carrier charges.
The average order value foundation explains the calculation and common uses. This diagnostic begins after the dashboard has already produced the average.
Why the average may not describe a typical order
An average gives every revenue dollar weight. One unusually large order can therefore move it sharply.
Imagine a business has 100 completed direct-to-consumer orders. Ninety orders average $45, nine orders average $100, and one corporate gift order is $1,050.
| Order group | Orders | Average in group | Group revenue |
|---|---|---|---|
| Everyday orders | 90 | $45 | $4,050 |
| Larger orders | 9 | $100 | $900 |
| Corporate order | 1 | $1,050 | $1,050 |
| Total | 100 | — | $6,000 |
The reported AOV is $60:
$6,000 ÷ 100 orders = $60 AOV
Yet 90% of orders are in the group averaging $45. The $1,050 order contributes more than one-sixth of all revenue while representing only 1% of orders. A business that treats $60 as its usual basket may set offers, inventory expectations, or packaging plans for a customer pattern that barely exists.
The corporate order is not bad and the AOV is not wrong. The average simply needs context.
Add the median order value
The median order value is the middle order after all order values are sorted from smallest to largest. With 100 orders, it is the midpoint between orders 50 and 51.
Suppose those two orders are $42 and $43:
Median = ($42 + $43) ÷ 2 = $42.50
Now the review has two useful facts:
- AOV: $60
- Median order value: $42.50
The $17.50 gap signals a right-skewed distribution: a minority of large orders pulls the average above the middle. Inspect the order mix before acting on AOV.
Median is not automatically “better.” It can hide the financial importance of large orders. Use both: the average describes revenue per transaction across the full set, while the median describes the center of the order pattern.
Build order bands you can act on
Average and median still do not show where orders cluster. Place completed orders into simple value bands that match your price points. For example:
| Net product order value | Share of orders | Likely pattern to inspect |
|---|---|---|
| Under $30 | 18% | Trial item, refill, or single low-price product |
| $30–$49 | 47% | Core single-product or two-item order |
| $50–$74 | 20% | Add-on, modest bundle, or quantity increase |
| $75–$124 | 10% | Larger bundle or gift purchase |
| $125 and over | 5% | Bulk, event, corporate, or wholesale-like order |
Review the most common SKUs, units per order, discounts, and customer type within each band. A $90 order containing three profitable refills differs from a $90 gift bundle with extra packaging and absorbed shipping.
Order bands translate customer spending into production and fulfillment questions. If nearly half of orders sit between $30 and $49, those baskets—not the $60 average—should strongly influence everyday box sizes, packing instructions, component availability, and add-on tests.
Run an outlier sensitivity test
Do not quietly delete large orders from the official metric. Instead, calculate a labeled sensitivity view.
In the 100-order example, remove the single $1,050 corporate order only for the diagnostic:
Adjusted revenue = $6,000 − $1,050 = $4,950
Sensitivity AOV = $4,950 ÷ 99 = $50
The official AOV remains $60. The $50 sensitivity result shows how much one transaction affects the average. You could also report direct-to-consumer AOV separately and place corporate or wholesale orders in their own segment.
Different order types have different pricing, terms, packing work, and margins. A blended AOV can be mathematically accurate and operationally weak.
If you need to understand whether a higher AOV also produced better economics, use the order-value and contribution bridge. Pair the distribution view with the contribution margin guide before deciding that larger baskets are automatically healthier.
Use a five-step monthly diagnostic
For the last complete month:
- Export completed orders with a consistent revenue definition.
- Separate materially different channels or order types, especially retail, subscription, marketplace, wholesale, and corporate orders.
- Calculate AOV and median order value for each useful segment.
- Count orders in practical value bands and identify the products, discounts, and fulfillment work behind the largest bands.
- Run one labeled sensitivity calculation when a small number of transactions appears to move the average materially.
Choose the decision that fits the evidence. If orders cluster below a proposed shipping threshold, test whether the gap is realistic. If a bundle creates many $75 orders but weak contribution, recost it. If corporate orders explain the increase, improve that workflow without pretending every shopper changed.
Watch conversion rate as well. A higher threshold or more complicated bundle can increase completed-basket size while discouraging some purchases. The conversion-rate breakdown helps keep that tradeoff visible.
Frequently asked questions
Is median order value better than AOV?
No. AOV measures included revenue per order; median identifies the middle order. Use both when large transactions or a skewed order mix may make the average unrepresentative.
Should wholesale orders be included in AOV?
They can be included in a companywide number, but calculate a separate wholesale and direct-to-consumer view when pricing, quantities, terms, or fulfillment differ materially.
How many orders are enough for this review?
There is no universal cutoff. Use a period with enough completed orders to avoid treating a few transactions as a pattern, and disclose the order count. Smaller businesses may need a longer window.
Should tax and shipping count toward order value?
Follow the documented definition in your platform, then create a consistent management view if needed. For product-economics decisions, separating tax and carrier charges can make merchandise performance clearer.
Can AOV rise while most orders get smaller?
Yes. A few larger transactions can lift the average even if the median falls or lower-value order bands grow. Compare the distribution and segment mix before crediting a pricing or merchandising change.
Practical takeaway
A $60 AOV does not mean most customers spend $60. Calculate the median, build order bands, and test how sensitive the average is to unusually large transactions. Then connect each meaningful band to real products, discounts, packaging, labor, and contribution.
Your next action is simple: export the last 30 days of completed orders and write down four numbers—the AOV, median, share of orders in the largest band, and AOV without the single largest order as a labeled sensitivity test. Those four numbers will tell you whether the dashboard average describes your everyday customer or mainly reflects the edge of the distribution.
