---
title: "How to See Which Shopify Orders Came From Chat (Any App)"
slug: "how-to-see-which-shopify-orders-came-from-chat"
url: "https://chatty.net/blog/how-to-see-which-shopify-orders-came-from-chat/"
kind: "post"
author: "Drake Q."
author_url: "https://chatty.net/author/drake/"
published_at: "2026-10-27T09:00:00+07:00"
updated_at: "2026-10-27T09:00:00+07:00"
description: "Find the orders chat helped sell in Shopify admin with order marks, segments, discount codes and UTMs. Then check if chat added sales or only took credit."
categories: ["Marketing"]
tags: ["Shopify", "Artificial intelligence AI", "AI sales agent"]
reading_time: 15
feature_image: "https://cdn.chatty.net/wp-content/uploads/2026/10/how-to-see-which-shopify-orders-came-from-chat.webp"
---

# How to See Which Shopify Orders Came From Chat (Any App)

Find the orders chat helped sell in Shopify admin with order marks, segments, discount codes and UTMs. Then check if chat added sales or only took credit.

To see which Shopify orders came from chat, have your chat tool mark each order it touched, then filter those orders in Shopify admin. After that, check whether chat added sales or only took credit for sales you would have made anyway.

Most stores start from the other end. The chat app shows a revenue figure every month, and nobody on the team can point to the orders behind it.

This guide works with any chat app or [AI sales agent](https://chatty.net/blog/what-is-an-ai-sales-agent-for-shopify/). It covers five ways to find chat orders in Shopify admin, two tests for real lift and a one-line ROI check.

**Key Takeaways**

- **Count orders, not conversations**

  A conversation count describes workload. The sales number is the set of orders chat touched.
- **A revenue figure only counts if you can find its orders in Shopify**

  Ask your app to mark each order, then filter them yourself.
- **Attributed revenue is not extra revenue**

  Shoppers who chat were already likely to buy, so test lift before you celebrate.
- **Divide direct revenue by the full cost**

  That gives you revenue for every $1 you spend, a number finance can check.

## What tools let you see conversation impact on revenue in Shopify?

Three places show how conversations affect revenue and conversion rates in Shopify. The first is your chat app's own dashboard.

The second is a mark the app writes onto each Shopify order or customer, which you can filter. The third is Shopify's own reports, such as the order conversion summary and the Sales by discount report.

None of the three tells you whether chat caused a sale. Only a lift test does that, and you run it yourself. Here is how the layers compare:

| Layer | What it tells you | Can you audit it in Shopify? |
| --- | --- | --- |
| Chat app dashboard | Revenue and orders the app credits to itself | Only if the app also writes to the order |
| Marks on Shopify orders and customers | Which orders and customers had a conversation | Yes, filter Orders and build segments |
| Lift test (two periods, or a holdout) | Whether chat added sales | Yes, with Shopify's own sales numbers |

Most chat apps report a chat-to-sales rate, meaning the share of conversations that led to an order. That is a different number from your store conversion rate. It also differs between vendors, because each one decides which orders count.

Shopify's own data has limits too. The order conversion summary shows a customer's first visit, sessions and UTM parameters, but [Shopify](https://help.shopify.com/en/manual/fulfillment/managing-orders/analytics/conversion-summary) shows a shorter summary when cookies were blocked. Its reports also credit the last click.

Treat the dashboard as the vendor's claim, the order mark as your evidence and the lift test as your proof. The rest of this guide builds each one.

## Why your chat app's revenue never matches Shopify

A chat app's revenue rarely matches Shopify because each app decides for itself which orders count as chat orders. The rule sits in the vendor's documentation, and few merchants read it.

Six tools show how wide the range is. They are sorted by how long each one keeps counting after the chat:

| Tool | What starts the count | How long it counts | A stricter "bought what the chat showed" view |
| --- | --- | --- | --- |
| Quickchat | The AI agent added items to the cart, and that cart checked out | Not stated | The rule itself is the strict version |
| Alhena | A session where the AI gave a knowledge-grounded answer from store data | 24 hours | The grounded-answer condition |
| Gorgias | A ticket is created | 3 days | Not described |
| Tidio | A live chat, a ticket, or a sales flow that gave a coupon | 7 days | Lyro AI-assisted sales, after Lyro's product recommendations |
| Chatty AI Sales Agent | A chat with Chatty, or a view of your FAQs | 7 days | Chatty Direct, for orders containing a product Chatty showed |
| ChatBot.com | A conversation | During it, or within 30 days after | A Sold view of chats where the AI agent closed the order |

We read every source on 27 October 2026. First four rows: [Quickchat](https://docs.quickchat.ai/conversations/insights/assisted-revenue/) (page dated 19 August 2026), [Alhena](https://alhena.ai/blog/shopify-ai-chatbot-roi/) (updated 12 May 2026), [Gorgias](https://docs.gorgias.com/en-US/revenue-statistics-81744) and [Tidio](https://help.tidio.com/hc/en-us/articles/5445427745052-The-Analytics-section) (updated 14 October 2026). The last two rows come from the [Chatty help center](https://help.chatty.net/analytics/Chatty-Attribution-Metafields/) and [ChatBot.com](https://www.chatbot.com/integrations/shopify/).

"Not described" means the documentation does not say, not that the tool lacks it.

Picture one shopper who asks about sizing on Monday, adds the shoes to the cart herself and buys on Friday. Quickchat, Alhena and Gorgias would not count that order. Tidio, Chatty and ChatBot.com would.

Neither kind of rule is wrong. A long window counts more orders, including more that would have happened anyway. A strict cart rule counts fewer and misses shoppers who took the advice and clicked Add to cart themselves.

The most useful split inside any rule is assisted versus direct. A shopper asks about sizing, gets a recommendation and buys that pair: the order is assisted and direct. Another shopper asks about shipping times, then buys something the chat never mentioned: the order is assisted only.

When you report to finance, lead with direct revenue. It is the stricter number.

A few housekeeping gaps widen the difference further:

- **Refunds.** Gorgias drops fully refunded orders from its report. Other tools keep the total from checkout, so ask which one yours does.
- **Permissions.** An app that writes to orders can only mark orders placed after you gave it access.
- **Dates.** Set the same date range and time zone on both sides before you compare.
- **Last click.** Shopify's marketing reports credit the last click, so a chat order can show up as "Direct".

## Five ways to find chat orders in Shopify admin

Five methods let you find chat orders in Shopify admin, ranked from the most reliable to the least. Three need no code at all, and none needs a developer:

| Method | Where in Shopify admin | What it catches | What it misses | Effort |
| --- | --- | --- | --- | --- |
| Order marks your chat app writes (metafields or tags) | Orders, filtered by the field or tag | Every order the app's rule counts, one by one | Orders outside the app's rule; orders before the app had access | Low, if the app writes them |
| Cart note turned into a tag with Shopify Flow | Order page, Additional details; then Orders filtered by tag | Orders from a cart the chat app marked | Shoppers who chatted, then checked out in a new cart or on another device | Medium |
| Customer segment of shoppers who chatted | Customers > Segments | Everyone who chatted, ready to compare | Which order the chat influenced | Low, if the app writes a customer field |
| A discount code you only give in chat | Analytics > Reports > Sales by discount | Orders that used the code | Chat buyers who skipped the code; leaked codes | Low, but costs margin |
| UTM tags on links that leave your store | Order page, Conversion summary | Shoppers who come back from an email or social reply | Anything on your own store; blocked cookies | Low, but easy to get wrong |

Use the first method if your app supports it. Add one of the others as a cross-check.

### Order marks your chat app writes onto the order

An order mark written by the chat app is the best evidence you can get, because it sits on the order itself. The mark is usually an order metafield or an order tag.

For a metafield, go to Settings > Metafields and metaobjects > Orders in Shopify admin. Open the app's field, turn on "Filter on the order list and in the Admin API" and save. Then go to Orders, click Add filter and pick that field.

![Example screen of Shopify admin Order metafield definitions, rebuilt from the Chatty help center, with the Chatty Direct and Chatty Assisted true or false fields highlighted](https://cdn.chatty.net/wp-content/uploads/2026/10/shopify-order-metafield-definitions-chat-fields.webp)

Shopify allows that filter on only 5 order metafield definitions per store, according to [Shopify](https://help.shopify.com/en/manual/fulfillment/managing-orders/viewing-orders/filtering-orders-with-metafields). Tags are simpler: type the tag into the order list search.

A mark is only as honest as the rule behind it, so read it next to the table above. Ask your vendor one question: which field or tag do you write, and when?

### Cart notes turned into order tags with Shopify Flow

Some chat apps write to the cart rather than the order. Shopify carries those cart attributes onto the order, where they appear under Additional details.

A Shopify Flow workflow can turn that note into an order tag you can filter. [Crescendo](https://docs.crescendo.ai/assistants/shopify-attribution) documents this exact setup for its chat widget (April 2026).

The method catches orders from a cart the chat app marked. Crescendo marks the cart as soon as a conversation starts. It misses shoppers who chatted, then checked out in a new cart or on another device.

### Customer segments of shoppers who chatted

A customer segment groups everyone who chatted, so you can study them as one audience. Go to Customers > Segments, click Create segment and filter on the customer field your app writes, such as "has chatted".

Use the segment for an email flow or an ad audience. You can also compare its repeat purchase rate and AOV with everyone else.

Be careful with that comparison, though. Shoppers who chat are not a random group, and the next section explains why the gap flatters chat.

### A discount code you only give in chat

A discount code that appears only in chat is the cheapest method, and it works with any app. Shopify's Sales by discount report then shows the orders and total sales for that code, according to [Shopify](https://help.shopify.com/en/manual/reports-and-analytics/shopify-reports/report-types/default-reports/sales-report).

The method undercounts, because most chat buyers never ask for a code. It also costs you the discount, and codes leak to coupon sites. An order with combined discounts appears once in each discount's row.

### UTM links, but only for links that leave your store

UTM parameters help only on links your team sends off-site, such as an emailed transcript or an Instagram or WhatsApp reply. Read the result on the order page under Conversion summary.

Never add UTMs to product links inside the chat on your own store. The shopper who arrived from your Meta ad gets relabelled as a chat visit, and your ad report loses the sale. The [GOV.UK data team](https://docs.data-community.publishing.service.gov.uk/processes/utm-usage) warns against campaign tags on internal links for this reason.

UTMs also depend on the storefront tracking pixel. Crescendo notes that ad blockers, browser tracking prevention and declined cookie banners routinely block it. Treat UTMs as a cross-check, not as your main number.

## Attributed revenue vs extra revenue: how to check whether chat added sales

Attributed revenue counts the orders chat touched. Extra revenue counts the orders that would not exist without chat. Every method in the previous section measures the first, and only a test measures the second.

The gap can be large. [Polar Analytics](https://www.polaranalytics.com/post/shopify-ai-chatbot-and-assistant-whats-actually-worth-using-in-2026) describes a bot that was "adding a step rather than saving a sale" (July 2026).

Its verdict: "The deflection was real. The incremental revenue was not."

![A bar of all orders chat touched, split into orders that would have happened anyway and orders added by chat, with three steps: mark the orders, test the lift, divide by cost. Bar widths are for illustration, not data](https://cdn.chatty.net/wp-content/uploads/2026/10/attributed-vs-extra-chat-revenue.webp)

Three ideas make the gap measurable:

### Why shoppers who chat buy more, with or without chat

Shoppers who chat buy more mainly because they were already closer to buying. They are on a product page, asking about a size, a delivery date or a discount code.

So a chart that says "chatters convert five times better than non-chatters" mostly measures intent. The same bias applies to the segment comparison from the previous section.

Keep the chatter comparison for spotting trends inside chat. Do not use it to prove that chat earns its cost.

### Compare two periods with the same traffic

Comparing two periods is the test you can run with no extra tool. Take your store conversion rate and revenue per session from Shopify Analytics, with chat on and with chat off.

Keep the periods alike. Use the same length and the same weekdays, with no sale, launch or holiday inside either one. Check that paid traffic stayed roughly the same.

You can also alternate on and off weeks for a month. The result is noisy and seasonal, so read it as a direction rather than a decimal.

### Run a holdout: hide chat from a share of shoppers

A holdout is the strongest test. A share of shoppers, for example 1 in 10, never sees the chat or its proactive messages. You then compare their conversion and revenue per session with everyone else over the same weeks.

Alhena calls a holdout "the strongest way to prove incremental lift rather than just correlation." Polar recommends the same thing at session level.

Ask your vendor whether it can hold out a share of shoppers. If it cannot, an A/B testing app can hide the widget for part of your traffic.

Smaller stores need patience here. Run the test until each group has enough orders that the gap stops swinging from week to week.

Report direct revenue as your claim, and the lift test as your check on that claim.

## Is chat paying for itself? A one-line AI ROI check

Chat pays for itself when the revenue you can find in Shopify, adjusted by your lift test, beats the full cost of the tool. That is the whole AI ROI calculation for a sales tool, and it needs two numbers:

- **Revenue for every $1 you spend.** Divide one month's direct revenue by the month's chat bill, including any extra conversations.
- **Break-even orders.** Divide the monthly bill by your average order value.

Here is an example, not a benchmark. A $68.99 monthly plan and an $80 AOV mean one order a month covers the plan.

Then apply the lift test. If your holdout says only part of the attributed revenue is extra, scale the revenue down before you divide.

Count the cost honestly too. Add your team's time on chats they take over, plus the margin you give away through chat-only codes.

Be wary of calculators that fill in a vendor-wide rate. [Zipchat's ROI calculator](https://www.zipchat.ai/tools/roi-calculator) assumes a "16.4% chat conversion rate and 21% AOV lift" from its own platform data (read 27 October 2026). Replace any vendor default with your store's numbers.

If you are still choosing a tool, Chatty's [ROI estimator](https://chatty.net/#home-roi-heading) runs the same arithmetic with your own volume and AOV.

Support savings belong in a separate line. Our guide to [chatbot analytics](https://chatty.net/blog/chatbot-analytics/) covers the support metrics, while this guide stays on sales.

## Six questions to ask any chat app about its revenue number

Six questions tell you whether a chat app's revenue number will survive your finance lead. Paste them into a vendor email or a trial checklist:

1. **What makes an order count, and for how many days?** A good answer names the trigger and the window.
2. **Do you separate orders where the shopper bought what the chat showed?** A good answer is a separate, stricter number.
3. **Where in Shopify admin can I filter those orders myself?** A good answer names a metafield or tag.
4. **Do you take out refunds and cancellations?** A good answer explains when, and how it differs from Shopify.
5. **Can you split what the AI sold from what my team sold?** A good answer is a switch, not a promise.
6. **Can I hold out a share of shoppers to measure lift?** A good answer is yes, or a clear workaround.

An app that cannot answer question 3 is asking you to take its dashboard on trust.

## Chatty AI Sales Agent: the orders it assisted, marked in your Shopify admin

Chatty is the AI sales agent for Shopify stores. Every shopper gets their own sales rep: Chatty greets them, recommends the right product, brings them back to their cart and looks after them once they buy. It answers from your own Shopify data, and marks every sale it helped make on the Shopify order.

AI chatbot and live chat are two standout features of Chatty. The AI chatbot answers from your own Shopify data, and with live chat your team takes over from the AI and closes the sale.

Chatty marks every order it assisted in Shopify, so you can check the return yourself. These are the fields Chatty, the AI sales agent for Shopify, writes onto each order:

| Field on the Shopify order | What it means |
| --- | --- |
| Chatty Assisted | The customer chatted with Chatty, or viewed your FAQs, within 7 days before ordering |
| Chatty Direct | The order contains a product Chatty showed or mentioned in that chat |
| Chatty Conversation | The ID of the chat before the order |
| Chatty Origin | The last Chatty surface that created or added to the cart |

Each customer record also gets Chatty Chat Status, Chatty Resolution Type and Chatty Responded to Proactive.

Those fields cover the first and third methods above. Pin and filter the order fields under Settings > Metafields and metaobjects > Orders.

Then build a segment on Chatty Chat Status set to engaged. Chatty needs permission to update your orders, and orders placed before that stay blank.

Inside Chatty, the Overview tab shows assisted revenue, chat-to-sales rate and share of store sales on every plan, including Free. Paid plans add the Sales tab, which splits revenue, orders and AOV between the AI and your team.

![Example screen of the Chatty Analytics Overview tab, rebuilt from the Chatty help center, with total conversations, resolution rate, assisted revenue, chat-to-sales rate and share of store sales, assisted revenue highlighted](https://cdn.chatty.net/wp-content/uploads/2026/10/chatty-analytics-overview-assisted-revenue.webp)

From Basic up, Open in Shopify saves a "Revenue from Chatty" report in Shopify Analytics. Shopify then works out the total from your real orders. Chatty's [sales analytics page](https://chatty.net/analytics/) walks through each view.

![Example screen of the Revenue from Chatty report settings in Shopify Analytics, rebuilt from the Chatty help center, with the Chatty Assisted filter set to Yes](https://cdn.chatty.net/wp-content/uploads/2026/10/shopify-analytics-revenue-from-chatty-report.webp)

Chatty Assisted uses a 7-day window, so some of those shoppers would have bought anyway. Lead with Chatty Direct when you report revenue.

Published results show the order-level view in practice. These are attributed figures from Chatty Analytics, not lift tests:

- [Stonehenge Health](https://chatty.net/case-study/stonehenge-health/), a supplement brand, recorded 599 orders and $75K in assisted revenue over 7 months, at a $124.85 AOV. By our arithmetic, 599 orders at $124.85 come to about $74,800, so the revenue figure is made of orders you can count.
- [BARABAS](https://chatty.net/case-study/baraba/), a menswear brand, recorded $300K in AI-assisted revenue in 12 months.
- [Montana West](https://chatty.net/case-study/montana-west/), a western fashion store, made $41,115 in revenue from chat during peak season. Its chat revenue grew 171% while conversations grew 83%.

## Count the chat revenue you can find in Shopify

Chat revenue is worth reporting only when you can open Shopify and find the orders behind it. Mark the orders, read them as attributed, test lift with two periods or a holdout, then divide direct revenue by the full cost.

This week, send the six questions to the chat app you use now. Its answer to question 3 tells you how much of its dashboard you can trust.

Bring your best salesperson to every shopper. Try [Chatty AI Sales Agent](https://chatty.net/ai-sales-agent/), then find the orders it assisted in your own Shopify admin.

## Frequently asked questions

### How do I know if a Shopify order came from a chat?

Look for a mark your chat app writes on the order, such as a metafield or tag, and filter Orders by it. Without one, use a discount code you only give in chat, or UTMs on links your team sends off-site.

### What is the difference between assisted and direct chat revenue?

Assisted revenue counts orders from shoppers who chatted before buying. Direct revenue counts orders that contain a product the chat showed or recommended. Direct is the stricter number, so lead with it when you report to finance.

### Does chat attribution count sales that would have happened anyway?

Yes, some. Shoppers who open chat are already closer to buying, so attributed revenue includes orders they would have placed anyway. Lead with direct revenue, and run a two-period comparison or a holdout to see the real lift.

### How long should a chat attribution window be?

No single window is right. Documented rules run from a cart-add rule with no window to 30 days.

Chatty uses 7 days plus a stricter Chatty Direct flag. Compare tools only after you know each one's rule.

### Can Shopify track chat sales without a chat app's help?

Only partly. Shopify does not know a chat happened unless something writes it to the cart, order or customer. You can still use a chat-only discount code, UTMs on off-site links and the order's conversion summary.

### How do I calculate the ROI of an AI chatbot or AI sales agent?

Divide one month's direct revenue by the month's bill to get revenue for every $1 you spend. Divide the bill by your AOV to get break-even orders.

Chatty is one example. Your plan includes AI conversations, each extra one costs $0.40, and you set the monthly limit. See [Chatty pricing](https://chatty.net/pricing/).

### Why doesn't my chat app's revenue match Shopify's?

The app counts orders by its own rule and window, while Shopify reports credit the last click. Refunds, the date the app gained access to your orders and mismatched date ranges widen the gap. Set the same dates before you compare.

### Is Chatty a chatbot?

Chatty is an AI sales agent rather than a scripted chatbot. It writes each answer from your store's data instead of following a fixed script, and it can take actions such as adding products to the cart and looking up an order. Your team also gets a live chat inbox for the conversations that need a person.