How to set a free shipping threshold that pays for itself

A free shipping threshold should come from the spread of your order values and your contribution margin per order. Multiplying average order value by a fixed factor is not enough.

Most of the cost of a threshold comes from customers who already spend above it and no longer pay for shipping. The return comes from customers who add items to reach it. It also comes from customers who would have dropped out over the shipping fee but now complete the order. The right threshold is the one where the second side outweighs the first.

This guide works through the threshold step by step for one hypothetical store. It shows how many new orders two different thresholds need to pay for themselves, and how to test the threshold you pick.

A threshold splits your orders into three groups. The first group already spends above the threshold. These customers would have placed the same order anyway, but they no longer pay for shipping. You now pay their shipping and get no extra sale in return.

The second group sits just below the threshold and adds items to reach it. The third group is new customers who would have abandoned over the shipping fee but complete the order because of the threshold. The return on a threshold comes only from these two groups.

Academic research shows how sensitive shoppers are to shipping fees. A 2006 study by Lewis, Singh and Fay reports that shipping fees affect both whether people order and how large their baskets are. It finds threshold-based free shipping effective at generating extra sales.

The same study notes that lost shipping revenue can make such promotions unprofitable for the retailer. So extra sales alone are not enough of a reason.

Why the average order value rule misleads

Most guides suggest setting the threshold 20% to 30% above average order value. No published source stands behind that rule. More importantly, average order value does not tell you how many orders already sit above the threshold.

Picture two stores with the same average. In one, most orders cluster around the average. In the other, many small orders sit next to a few large ones.

The same threshold can cover few orders in the first store and many more in the second. So start by splitting the last few months of orders into value bands. You can check average order value (AOV) with the AOV calculator, but the information the decision needs is in the bands themselves.

Fix three definitions before you calculate

The first definition is tax. The threshold shoppers see may include or exclude sales tax, while product cost and carrier rates are usually net of tax. Run the calculation on one definition, ideally net of tax.

Platforms read the amount in different ways too. Shopify says shipping rates are based on the cart value after discounts and before taxes. Check whether discount codes count toward the threshold on your own platform.

The second definition is contribution margin. Here it means the margin left after product cost, payment fees and packaging, excluding shipping. You can find gross margin with the gross margin calculator and subtract the other variable costs from it.

If an order shipped free is returned, you may pay shipping both ways. So in categories with high return rates, calculate contribution margin net of returns.

The third definition is shipping cost itself. For bulky or heavy items, carriers price by size and weight, so cost per order can sit far above the average. If you sell such items, you may need to treat them separately from the threshold.

Hypothetical example: a store takes 2,000 orders a month. Values are net of tax and returns. Customers pay a $6 shipping fee on every order, while the store pays the carrier $9 per order. The store already covers the $3 gap.

In this distribution average order value is $69.05. The "20% to 30% above average" rule would put the threshold at about $83 to $90. The example compares one threshold on each side of that range: $75, close to the average, and $100.

Contribution margin excluding shipping is assumed to be 35%. The example also assumes that 30% of customers in the band just below the threshold add an item worth $25 on average. These assumptions will change with your data, but the structure of the calculation stays the same.

When a customer who adds an item leaves a profit

A customer who adds an item to reach the threshold does two things. They add an item, but they also stop paying the $6 shipping fee. If the margin on the added item does not cover that loss, profit falls even though the order grows.

To find this limit, divide the lost shipping fee by the contribution margin rate. In the example that is $6 divided by 0.35, or about $17.

If a customer adds less than that, the threshold loses money on that order. The customer in the example adds a $25 item. Its margin is $8.75, which leaves $2.75 of extra profit per order after the lost fee.

So when you set the threshold, look at the items that can be added to the cart. If the items recommended to customers just below the threshold sit under this limit, a new product suggestion may matter more than the threshold itself.

Compare two thresholds: which one pays for itself

The table below compares $75 and $100 thresholds for the same store. Customers who add items come from the band just below each threshold.

Both thresholds lose money before any new orders arrive. The real question is how many new orders it takes to close the gap. New orders are assumed to land just above the threshold: $80 for the $75 threshold and $105 for the $100 threshold. The store pays the full shipping cost on these orders, so margin per new order is $19 and $27.75.

That is why the last row matters. The threshold near the average needs a much larger rise in orders to pay off. It takes on the shipping cost of the 700 orders already above it.

How the result changes when the assumptions change

The result in the example rests on three assumptions: the share of customers who add an item, the value of that item and its margin. The table below varies the add-on share and the added item's margin. It recalculates the order increase each threshold needs to pay for itself. The added item stays at $25.

The margin that changes in the table is only the margin on the added item. Contribution margin on new orders stays at 35%. When the added item's margin falls to 20%, a $25 item contributes $5. That is less than the $6 shipping fee the store gives up.

In that case, more customers adding items makes the threshold harder to pay off. Each order that adds an item leaves a loss, not a profit. So pushing the add-on rate up in a cart that recommends low-margin items can shrink profit.

Under the assumptions tested, the $100 threshold needs fewer extra orders to pay for itself. The table does not show on its own which threshold is more profitable.

The $75 threshold reaches more customers and may create a larger rise in orders. Only a test can tell which threshold actually delivers the increase it needs. If most orders cluster just below the threshold and customers add high-margin items, a threshold near the average can still pay off.

The three groups above are clear as concepts, but they are hard to separate in real data. To find orders that added an item, look at cart events. Some orders were below the threshold before the last item was added and above it at checkout. Flag these as threshold-crossing orders. That needs the order of add-to-cart events and the cart value at each step.

The cart value here should be the amount the threshold is based on. A cart that drops back below the threshold after a discount code should not count as crossing it. Also split threshold-crossing orders into new and returning customers. An extra order driven by free shipping can also come from an existing customer buying more often.

This definition shows a behaviour, not a motive. The customer may have planned to buy the last item anyway. Some of these orders would have happened anyway. The only reliable way to separate them is a control group that does not see the threshold.

When the item added to reach the threshold is returned

A threshold can push customers to add an item they do not want. If that item is returned later, the store loses its margin and may pay shipping twice. The $2.75 of extra profit in the example can turn negative even if only a small share of added items come back.

To measure this risk, track the return rate of threshold-crossing orders separately. If it is well above the rate for other orders, there are three options, and their order matters.

The first option is to recommend high-margin, low-return items to customers just below the threshold. That raises the margin on added items without changing the threshold. The second option is to exclude high-return categories from counting toward the threshold. That rule is harder to explain and to apply in the cart. If the cart does not show it, it creates a surprise at checkout.

The third option is to change the threshold, but only by recalculating it with the same table. A lower threshold means fewer customers need to add items, but more existing orders qualify for free shipping. So you may reduce the return problem while raising shipping cost.

The fee below the threshold does not have to stay as it is.

A 2012 study by Koukova, Srivastava and Steul-Fischer compares threshold-based free shipping with a flat fee. Shoppers rate the threshold less favourably when the cart is below it. They rate it more favourably when the cart is above it. So a threshold can feel like a penalty to the customer who cannot reach it.

That is a good reason to consider a tiered structure, such as a lower fee below the threshold and free shipping above it. Shopify's shipping rates page gives an example of this kind of structure. It charges one fee for all orders, a lower fee above one amount and free shipping above a higher amount. The same study also notes that explaining the fee as a delivery cost improves how customers see it.

Changing the threshold moves conversion rate, order value and shipping cost per order at the same time. So the success metric for a test should be contribution per user, not conversion rate or revenue. Conversion rate can rise while contribution per user falls. The threshold may be winning customers while taking profit from somewhere else.

Assign users to groups, not visits. The same person should see the same threshold on every visit, or the two groups mix. The success metric should use the same unit.

Divide the total contribution from a group's users by the total number of users assigned to that group. Users who do not buy stay in the denominator, and repeat visits by the same person do not count as new users.

Returns reach the result later. So give both groups' orders the same return observation window and read the result after it closes. The control group guide explains how to set up the comparison group.

The site's sample size and A/B test significance calculators are built for rates. They can check the conversion rate side of the test. They do not test a monetary metric like contribution per user on their own. Monetary metrics vary much more from user to user. So they usually need more traffic and a method that tests a difference in means, such as a t-test or bootstrap.

If you cannot split traffic, you can change the threshold for a set period and compare it with the period before. Sales and seasonality can distort that comparison. Compare weeks that share the same promotional calendar. Track conversion rate separately with the conversion rate calculator, and track the drop-off from cart to payment with the funnel analysis calculator.

Customers act on a threshold only if they see it. Baymard Institute's cart abandonment research excludes "just browsing" respondents and looks at the remaining reasons. Among these, 40% cited extra costs such as shipping, tax and fees. That figure comes from a survey of 1,083 US adults who shop online.

A separate 12% said they could not see or calculate the total order cost up front. Seeing an extra cost for the first time at checkout is a strong reason to leave.

That is why showing how far the cart is from the threshold matters. In Google Merchant Center you can set a free shipping threshold in account-level shipping settings or with a product-level attribute. Google says not to set a minimum order value if you still ship below the threshold for a fee. A minimum order value means you do not sell below that amount at all.

A threshold is not a value you set once and leave. As carrier rates, product prices and order values change, the same threshold starts to cover a different number of orders. When prices move quickly, review the threshold every few months against a fresh order distribution.

The steps below let you calculate the threshold the same way each time.

1. Split the last three months of orders into value bands, net of tax and returns.
2. Set shipping cost per order, the fee customers pay and contribution margin excluding shipping.
3. Divide the lost shipping fee by the contribution margin rate to find how much an added item must be worth.
4. For each candidate threshold, calculate the shipping cost of orders already above it.
5. Find the number of new orders the threshold needs to pay for itself and decide whether that is realistic.
6. Test the chosen threshold on contribution per user and record the result with the order distribution.

This calculation bases the threshold on your own orders, not on a guess. Average order value can be a starting point, but the value bands and contribution margin tell you whether a threshold pays. The performance marketing metrics guide reads ad metrics with the same logic.