E-commerce
June 28, 2026
A ticket taxonomy is not just a reporting tool. If well thought out, it helps support understand the request faster, route the case to the right place, and respond with the right rule.
For the customer, this means fewer repeated questions, less waiting, and a more consistent response across email, chat, and phone.
This guide shows how to build a simple, useful taxonomy that is compatible with a support chatbot.
Summary
Why classify tickets?
Without a clear classification, two agents may name the same problem differently. A "parcel not received" request can become delivery, carrier, claim, refund, or emergency depending on the person processing the ticket.
This dispersion makes the root causes difficult to read and slows down improvements. A well-designed taxonomy transforms conversations into actionable signals.
To classify a ticket well is already to prepare a better response.
Convert over 2,000 customers on average per month with Qstomy.
The world’s 1st Shopify AI dedicated to customer conversion
Empowering 200+ e-commerce merchants
What levels should be planned?
The taxonomy must remain readable. A first level can cover ordering, delivery, returns, refunds, product, payment, account, warranty, security, and pre-purchase advice.
A second level specifies the case: address to modify, delayed package, missing item, refused return, requested invoice, or product compatibility. Beyond that, the complexity may become more costly than useful.
How to think client-side?
Categories must reflect the customer's words as much as internal needs. A customer does not always say “logistics SLA”; they say “my package is not arriving” or “I am leaving tomorrow.”
The chatbot can translate this intent into an internal reason, while maintaining a human and understandable response.
Which data to associate?
Each reason must indicate the necessary information: order number, email, product, proof, photo, tracking, date, channel or status. This avoids asking for too much information on simple cases and not enough on sensitive cases.
The taxonomy must also signal data that should not be requested in the chat, such as passwords or bank codes.
How to use taxonomy to improve?
Volumes by reason reveal recurring issues: unreliable carrier, misunderstood return policy, incomplete product page, confusing payment, or poorly visible warranty.
The taxonomy then becomes a tool for improving the experience, not just a column in a ticketing tool.
It must be reviewed regularly with support, logistics, product, and marketing to avoid obsolete categories.
Which flow to follow?
The flow must start from real requests and remain easy to maintain.
Analyze existing conversations to identify recurring patterns and those that create risk.
Create understandable main categories: order, delivery, return, payment, product, or account.
Add sub-reasons only when they change the response, routing, or reporting.
Associate the necessary data, deadlines, rules, and possible escalations with each reason.
Measure volumes, correct duplicates, and remove categories that are no longer useful.
Which examples should be used?
“Delivery > parcel delivered not received” must trigger a different verification than “Delivery > delayed parcel”. “Return > label not received” does not call for the same response as “Return > used item”.
These distinctions prevent the customer from receiving a generic answer that does not solve their problem.
When to transfer?
Transfer is necessary when the reason involves payment, fraud, personal data, dispute, strong emotion, dangerous product, or an uncovered rule.
The bot must transmit the reason, sub-reason, context, evidence, history, urgency, and expected action.
Which KPIs should be monitored?
Track volume by reason, resolution rate, reopenings, handling time, escalations, satisfaction, "other" categories, and rising reasons.
A too frequent "other" category shows that the taxonomy no longer reflects actual requests.
Which mistakes should be avoided?
Avoid too many categories, incomprehensible internal names, duplicates, patterns without owners, or classifications that do not change any action.
A taxonomy should help to respond, route, and learn.
How can Qstomy help?
Qstomy can connect the chatbot to support reasons, transactional messages, reassurance rules, post-delivery content, product sheets, warranties, compatibilities, accessories, orders, and response bases to guide the customer with reliable information.
The chatbot helps the customer move forward without inventing a lead time, a warranty, a compatibility, a return rule, an order status, or a product recommendation that still needs to be confirmed by a reliable source.
Explore AI support, the AI sales agent, or request a demo.
Key takeaways
Key Takeaways
A useful taxonomy connects reason, sub-reason, data, rule, routing, escalation, and reporting.
What the Customer Needs to Understand
The customer must be understood quickly without having to repeat their problem at every step.
The Right Limit of the Chatbot
The chatbot can classify and collect, but it must transfer sensitive or uncovered reasons.
Enzo
June 28, 2026
