How a denim brand cut sizing tickets to near-zero in three weeks.
A 480K-MRR denim & basics brand was drowning in Sunday-night sizing DMs. Three weeks after installing Arbyn, the Head of CX hadn't touched a sizing ticket in months.
Three weeks in, the inbox looked like a different company's.
Two people. Five thousand tickets a month. Most of them sizing.
The brand's CX team was two full-time people plus a weekend contractor brought in after a particularly bad December. About 5,200 tickets came through per month, split across email, Instagram DMs, and an on-site chat widget. The makeup wasn't unusual for fashion: roughly 68% of inbound was sizing, fit, or exchange-related. The rest was WISMO, return requests, and the long tail.
The team had built solid macros and a thorough size chart. None of that helped at 11 PM on a Sunday when 40 customers were asking "does the slim straight in 27 run small?" with nobody to answer.
Weekend first-reply times had crept up to 28 hours. Customers were placing orders, getting cold feet, and returning instead of exchanging because no one had answered their fit question in time. The Head of CX described the inbox on a Sunday night as "watching money walk out the door at the speed of email."
The cost wasn't headcount. It was the silence.
The team had quoted three options to the founder: hire a third full-time CX rep ($72k loaded), bring on an offshore overflow team ($28-35 per ticket, low CSAT risk), or try one of the chatbots. They'd tried two chatbots already. Both got the size charts wrong, mangled the brand voice, and bounced every other question to "a human will be in touch."
What changed the math was a different way of looking at the problem. Sizing tickets weren't a customer service problem. They were a sales conversion problem: every unanswered fit question on a Sunday was a customer who didn't check out at all, or who ordered two sizes and returned one out of caution. Both were money the brand had already earned and lost.
They installed Arbyn that Wednesday. Connected Shopify in two clicks, pointed Arbyn at the size chart pages, fed it 18 months of past tickets, and let it watch for a week before going live.
Catalog ingestion, then voice, then the inbox.
Three things had to be right for this to work in fashion: the answer had to be accurate to the SKU, the tone had to match the brand, and the agent had to handle the exchange action — not just answer the question.
Step one: catalog ingestion. Arbyn read the size chart pages, the JSON metafields on every product, the model measurements from the PDP copy, and the per-SKU fit feedback the team had been tagging in Gorgias for years. By the time it went live, Arbyn could tell a customer that the slim straight ran 1.5cm narrower at the ankle than the high-rise straight in the same waist — because it was reading the spec, not making it up.
Step two: voice tuning. Arbyn was trained on 18 months of past replies the CX team had sent. The team flagged 30 representative responses as "this is exactly how we sound" and 10 as "this is too formal" and "this is too casual." After a week, even the Head of CX couldn't reliably tell which replies were hers and which were Arbyn's.
Step three: actions, not just answers. The team turned on Loop Returns and Klaviyo integrations on day one. From day one, Arbyn could quote a fit, start an exchange, reserve inventory, send a prepaid label, and subscribe a customer to a restock alert — all in a single reply, without escalating.
How the inbox changed, week by week.
Stock Arbyn. Plus three native integrations.
Nothing custom. The brand turned on features that ship with Arbyn out of the box, plus three integrations that took a click each.
Sunday night reads "all caught up."
Three weeks in, the inbox math had flipped. 71% of tickets resolved entirely without a human. The 29% that escalated were the ones that should: angry customers, complex returns, the occasional press inquiry, a single legal question. Exactly the work the two-person team had wanted to be doing all along.
The numbers downstream of the inbox moved too. Return-to-exchange rate climbed from 18% to 51% — Arbyn's exchange-first phrasing converted a third of would-be refund customers into kept-money customers. Weekly revenue lift attributable to recovered carts and saved exchanges was roughly $9-12k.
The contractor was let go. The two full-time CX reps stayed and were re-tasked: one to community/loyalty, one to VIP and proactive outreach. Headcount didn't shrink — but the work changed shape.
The brand has been on Arbyn since February. Asked what's changed most, the Head of CX didn't reach for a number: "I sleep on Sunday."
Want a Sunday night that reads all caught up?
Install Arbyn on your Shopify store, connect your size chart and return policy, and watch your first sizing ticket get answered — in your voice — within the hour. $149/mo. Cancel anytime.