Shopify Oversell Policy: 5 Rules to Set Before It Happens

Without rules, each oversold order becomes a fresh decision. One customer gets a code, the next one doesn't. One ticket stays open for a week because nobody knows what to do when the customer goes quiet. This guide covers the rules themselves. For what to write to the customer, our guide to handling Shopify oversold item support, listed at the end of this article, has the reply that closes the ticket.

Why you need the rules before the oversell, not after

Oversold orders tend to come in batches. A promotion, a stock miscount or a slow channel sync can create ten of them in one afternoon. If you decide each case on the day, you make ten slightly different decisions under pressure.

Written rules fix three problems:

• Customers get the same treatment. Two people with the same problem never compare notes and find out that one of them got more.
• Tickets close. Every case has an end point, even when the customer never replies.
• You can delegate. A new hire, a freelancer or an AI agent can only follow rules that exist on paper.

Keep the whole policy on one page. If it needs more than that, it's too complex to apply consistently.

Rule 1: Set a refund cap

The refund cap is the order value up to which an oversold item is refunded without anyone on your team checking it first. Above the cap, a person looks at the order before money moves.

How to choose it:

• Start from your average order value. If most orders sit below a certain amount, a cap just above it covers the routine cases and leaves the unusual ones for review.
• Look at your margins, not just the price. A low-margin product with a high price may deserve a lower cap than a cheap, high-margin one.
• Write it as one number. "Refunds up to this amount are approved automatically" is a rule. "Small refunds are fine" is not.

Example rule: "Oversold items in orders below our cap are refunded to the original payment method. Orders above the cap go to the store owner for approval the same day they're flagged."

Rule 2: Set the default when the customer doesn't reply

This is the rule most stores forget. You email the customer with their options, and they never answer. Without a default, the order sits unfulfilled and the ticket stays open.

Your default needs two parts:

• A number of days. Choose it once and use it for every oversold order.
• An action when time runs out. For most stores, that means refunding the missing item and shipping the rest. A full cancellation only makes sense when the missing item was the whole order.

State the default in your first email, so the customer knows what will happen if they stay silent. A line like this works without anything to fill in: "If we don't hear from you, we'll refund the item that's out of stock and ship the rest of your order."

If you offer a wait for a restock, the same logic applies. Only offer it with a date confirmed by your supplier or warehouse. If the date isn't confirmed yet, say this instead: "We will confirm the restock date by email as soon as our supplier confirms it." Never let anyone guess a date to fill the gap.

Rule 3: List the cases that go to a person

Some oversold orders shouldn't be handled by a fixed rule at all. Write down which ones, so nobody has to judge it on the spot.

Common handoff cases:

• Order above your refund cap. Why a person: the amount justifies a second look.
• Swap to a "close match" product. Why a person: material, fit or color differences need judgment.
• Repeat customer with a previous problem. Why a person: one more template reply can lose them.
• Customer already angry, or asking for a manager. Why a person: tone matters more than speed here.
• Items sold as a set or a gift. Why a person: shipping part of the order may be worse than waiting.
• Legal threat or chargeback mention. Why a person: it needs a decision from the owner.

Our step-by-step guide to human handoff rules in a Shopify AI agent, listed at the end of this article, shows how to turn a list like this into clear triggers.

Rule 4: Decide your discount code rule

Should an oversold customer get a code to say sorry? There's no single right answer, but there is a wrong way to do it: deciding ticket by ticket.

Pick one of these and write it down:

• No code. The refund or swap is the fix. Simple and consistent.
• A code for every oversold order. Same value, every time.
• A code by order value. One amount below your refund cap, a different one above it.
• A code only on the second problem. For customers who already had an issue with a past order.

Also decide what the code applies to (next order only, minimum spend or not) and how long it stays valid. A code sent as goodwill usually costs you less than a refund of the same amount, because it only costs you if the customer buys again. Our guide to out-of-warranty claims and goodwill codes, listed at the end of this article, explains how to set this per product.

Rule 5: Name who owns restock dates

One person on your team confirms restock dates with the supplier and records them in one place. Everyone else, human or AI, only repeats what's recorded there. This one rule stops most "any news on my item?" follow-ups, because customers stop getting dates that later slip.

Where an AI agent fits

Once the rules exist, most oversold tickets become repetitive work: the same reply, the same refund, the same follow-up. That's the work Relay is built for. Relay answers, refunds and resolves your Shopify tickets 24/7, following your rules.

With your oversell policy in its knowledge base, Relay can:

• Understand the request, whether the customer asks where an item is or why part of the order is missing.
• Find the order in your Shopify store.
• Apply your store's rules: your refund cap, your no-reply default, your discount code rule.
• Act within the permissions you give it, such as a refund, a cancellation or a discount code.
• Hand off to your team for every case on your handoff list.

You decide the permissions. If you only want Relay to answer and never refund, that's a setting, not a workaround.

How to prevent the next oversell

Rules handle oversells. These checks reduce how often you need them:

• Review "Continue selling when out of stock". According to Shopify's help center, this option lets customers buy a product at zero stock, and it can only be turned on when "Track quantity" is active. Check which variants have it on, and turn it off wherever you don't mean to take backorders.
• Turn on "Track quantity" for every product you hold in stock. Shopify's inventory tracking guide in its help center shows where to find it.
• Check your channel sync. If a marketplace or wholesale channel sells the same stock, find out how often inventory syncs, and keep a small buffer on your fastest sellers.
• Recount before promotions. Your best sellers are the ones most likely to be miscounted when volume jumps.
• Check bundles. A kit can show as available while one of its parts is out of stock.
• Tag oversold orders in Shopify, so you can see which products cause them and fix the cause.

Before a peak like Black Friday, run all of these checks together. Our article on what happens to your helpdesk bill at Black Friday volume, listed at the end of this article, covers the support side of that week.

FAQ

What should a Shopify oversell policy include? At minimum: a refund cap, a default action when the customer doesn't reply, a list of cases that go to a person, a discount code rule, and the name of whoever confirms restock dates.

How many days should I wait before applying the default? Pick a number that fits your shipping times and use it for every order. What matters is that it's fixed and stated in the first email.

Should every oversold customer get a discount code? Not necessarily. Choose one rule (never, always, by order value, or only on a second problem) and apply it to everyone.

Can an AI agent follow my oversell policy? Yes, if the rules are written down and its permissions are set. Relay applies your store's rules, acts only within the permissions you give it, and hands off the cases on your list to your team.

Write the policy once, and oversold orders stop needing a decision every time.