Why Order Confirmation Emails Show Up as Helpdesk Tickets

The tickets nobody is waiting on

A robot ticket is a ticket your own systems create. A customer didn't write it, nobody will answer it, and nobody is waiting for it. One way it happens is an order notification that lands in the same mailbox your helpdesk reads.

These tickets can be easy to miss. They file themselves and no customer complains about them, so a busy queue can look normal. The damage shows up in the numbers. Ticket volume, tickets per order and any staffing plan built on them all include messages no person needed to read.

What we found

We audited a client's helpdesk on August 27, 2026, over a trailing 30 days. The finding: 48% of one client's monthly tickets (1,046 of 2,185) were their own order-confirmation robot writing to their own helpdesk.

The arithmetic is simple. 1,046 divided by 2,185 is 47.9%, which rounds to 48%. We found it and sized it. The fix was handed to the client, so we are not claiming the robot has been switched off. This is one client in one month. Treat 48% as a reason to count your own tickets. It says nothing about what your store will find.

How an order email ends up in the helpdesk

An email helpdesk turns mail sent to its inbox address into tickets. So the question is which of your systems writes to that address. Shopify sends two emails when a customer places an order: an "Order confirmation" to the customer (Shopify help) and a "New order" notification to staff. Either can reach a helpdesk. Five patterns can do it.

• The staff notification: Shopify sends a "New order" staff notification when a customer places an order, and its help center says you can add any email address as a recipient. Add your helpdesk address and each new order can open a ticket.
• The account owner's address: Shopify can send a staff notification to a staff member's own email address. Check which address the store's account owner uses and whether it is a recipient of "New order". If it is your support address, each order notification reaches the helpdesk.
• A copy rule: an app, an ERP or an email tool can copy customer emails to a shared mailbox. If the helpdesk reads that mailbox, each copy becomes a ticket.
• Other robots: shipping updates, review requests, payment-failed notices and form auto-replies can take the same path. Each one that reaches the helpdesk inbox becomes a ticket.
• Self-addressed mail: if the store sends a notification from your support address to your support address, the helpdesk can log a ticket from a "customer" who is really you.

What a robot ticket costs

The cost depends on your helpdesk. Gorgias's billing documentation says a ticket is billable when at least one message in it is sent from the helpdesk, and that includes an automatic reply from a Rule. Inbound tickets that get no reply from the helpdesk aren't billable. So in Gorgias, a robot ticket is billable if a rule replies to it and not billable if nobody does. Read how your own plan counts before you assume either way.

Fees aside, every robot ticket counts in your volume. Whether it also counts in response time and resolution time depends on whether your helpdesk closes it, replies to it or leaves it open. Find out before you quote either number to anyone. If a large share of your count is robots, a deflection target or a staffing plan built on it starts from the wrong base.

Check this in your own store: count them from one export

You need a spreadsheet and a 30-day ticket export. Check whether your helpdesk exports tickets to CSV. If yours can't, ask its support for a report with the same columns.

• Export the last 30 days of tickets. Keep four columns: created date, requester email, subject, and whether any reply went out.
• Open the export in Google Sheets. Insert a pivot table with requester email as rows and a count of tickets as values, sorted with the largest count first.
• Read the top 10 requesters. If a robot opens one ticket per order, that address appears about as often as you take orders, so it ranks at or near the top. Any address on your own domain up there is a suspect.
• Open the subject lines for the top requester. The same sentence repeated with only an order number changing is a robot.
• Get the share from your own domain with one formula. If requester email is in column B and your domain is yourstore.com, use =COUNTIF(B2:B, "*@yourstore.com*") / COUNTA(B2:B).
• Read 20 of those tickets by hand. Some may be staff forwarding a real customer problem. Take those out of the count.
• Compare the robot count with your orders for the same 30 days. If the two are close, one notification is probably opening one ticket per order. Write the robot share down with today's date.

Fix it at the source, not with a rule

Find the sender first. Open one robot ticket, note the sender address, the subject and the time, and match the time to an order. Then check where that message could have come from: your store's staff notifications, your apps and any copy rule.

In Shopify, go to Settings > Notifications > Staff notifications and read the Recipients section. Use Edit recipient to see whether "New order" is switched on for each one. If your helpdesk address is receiving it, replace that recipient with a mailbox or a channel the helpdesk doesn't read. Don't remove a recipient until you know who reads it, because someone may rely on that notification. Check the account owner's address too.

A helpdesk rule that closes these tickets is a fine stopgap, and it leaves three problems. The ticket is still created. If the rule sends a reply as it closes, a helpdesk like Gorgias may bill it, as above. And customers can reply to confirmation emails. A reply can keep the subject line with "Re:" added, so a rule that closes by subject can close a real customer's message.

If the sender lives in a system you don't control, such as a vendor's, write the request so its owner can act in one sitting: one sample ticket, the field to change and the date you'll re-count.

What the real queue looks like afterwards

Re-count seven days after the change. The robot share should fall to zero. If it doesn't, there's a second route, so go back to the sender list.

Your ticket count falls by the robot share. That drop corrects the measurement, and service hasn't changed. Mark the change date on your charts so a month-over-month comparison doesn't read as a win.

Once the queue holds only people, two numbers become usable. Tickets per 100 orders is real tickets divided by orders in the same period, times 100. First response time describes how fast people were answered, as long as the robots are out of it. Size any staffing or automation decision on those two, not on the raw count.

We diagnose an ecommerce operation from its own Shopify, helpdesk and ERP data, and a ticket count like this one is the kind of thing that data shows. You can also take our free 10-minute assessment of how systemized your team is. For what to automate once the data is clean, see our fulfillment playbook.

Frequently asked questions

How do I tell a real ticket from a notification?

Ask whether anyone is waiting for an answer. A real ticket has a person on the other end, a question or problem in the body, and a reply from your team. A notification comes from your own domain or a system name, repeats its subject with only an order number changing, and gets no reply. When you can't tell, open ten and read them.

Why are my orders emailing my helpdesk?

Something addressed or copied to your helpdesk inbox fires when an order is placed. In Shopify, start with the New order staff notification, which accepts any email address as a recipient, and the account owner's email address. An app that copies customer emails to a shared mailbox is a third place to look. Open one ticket, read the sender and time, then check those three in order.

Will closing these hurt reporting?

It will change your numbers, and the change is a correction. Ticket volume falls by the robot share. Response and resolution times may move, depending on how your helpdesk treated the robot tickets. Note the date of the change on your charts and compare the periods on either side of it separately. Keep the old tickets so your history stays intact.