How should customer service questions feed back into site changes?
Our answer
Tag each ticket by reason and product and review the top themes weekly. Answer each repeated question on the page where it arises, then check that its tickets fall.
- Confidence
- Medium
- Strongest proof
- 1 official rule
- 8 practitioners
- Sources
- 7 sources
- 9 statements
- GG
POBINNC
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentMany tickets trace to information shoppers could not find before buying. Those are the cheapest problems to fix, with a better description, size chart or photo (Pattern Owl, a feedback analytics software maker). Pattern Owl checks that tickets drop after each fix.
Fixing the top categories at the source cuts volume. Gorgias (ecommerce helpdesk software maker) sorts the last 30 days of tickets and fixes the top two or three. It puts a tracking link, delivery window and return policy link in the order confirmation.
A question answered more than three times belongs on the site. Gladly (customer service software maker) turns those into FAQ content and puts the answers on product pages.
Our judgment
Put each answer where the question starts. Pre-purchase answers go on the product page, and delivery and return answers go in the cart, checkout, order confirmation and tracking page. That combines Pattern Owl's product page fixes, Gladly's advice to answer on product pages rather than help pages and to host tracking on the store's domain, and Gorgias's order confirmation contents. Review the top themes weekly and recheck the list of reasons monthly and after launches, combining Pattern Owl's weekly 15-minute review with Gorgias's monthly category review. Give the same answers to an on-site chat assistant and to the facts AI agents receive, applying Nielsen Norman Group's advice to scope chatbots from past support logs and Shopify's knowledge base for AI agents.
Our judgment, built from Pattern Owl, Gladly, Gorgias, Nielsen Norman Group and Shopify: our reading of these sources, not a finding from any one of them.
What the sources say
9 statements from 7 sourcesShopify's Knowledge base now sits in the admin under the Agentic sales channel, replacing the standalone app after January 15, 2027. Shopify drafts store facts for AI agents, and nothing goes out until you publish it.
“Nothing Shopify drafted reaches AI agents until you publish it.”
Baymard's usability research says unclear order contents and status on the order list page delay users and can lead to unwanted extra orders and more strain on customer service.
“which can result in unwanted additional orders and place more strain on customer service.”
ClickPost, which sells post-purchase software, says support goals need a metric, target and deadline, such as cutting churn 1 to 2 points, and flags customers with repeat contacts for follow-up.
“A goal only becomes useful when it has three things: a metric, a target, and a deadline.”
Nielsen Norman Group's qualitative research says site chatbots are often abandoned. They should hand off to a human when asked, stay in scope, link products directly, and disclose that they are AI.
“Never gatekeep when the user explicitly asks to be connected to a human.”
Gladly, which sells an AI service agent, says most where-is-my-order tickets are preventable with proactive shipping updates, clear delivery dates and tracking hosted on your own domain.
“Most WISMO is preventable with proactive shipping updates, clear delivery dates, and self-service tracking.”
Gorgias, which sells a helpdesk, says to sort 30 days of tickets, write help articles for the top categories, and send tracking and return links after purchase. Electronics brands see nearly one ticket per two orders.
“Your top two or three categories are your highest-leverage targets.”
Pattern Owl, which sells feedback analytics, says group support tickets into patterns, not agent tags; it estimates 20 to 30% stem from missing pre-purchase details, fixed by better descriptions, size charts or photos.
“These tickets aren't product defects. They're communication defects.”
Gladly, which sells customer service software, says FAQs that cut support volume answer purchase blockers, belong on product pages, and should add any question answered more than three times.
“Every time your team answers a product usage question more than three times, that becomes FAQ content.”
Gorgias, which sells a helpdesk, says support can spot churn risk from negative ticket sentiment, tag VIPs for priority, and win back unhappy customers with discounts, perks or free returns.
“Customer service isn't just about reacting to problems.”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer