E-commerce & Retail · Fashion & Apparel
Every return starts with a size that guessed wrong.
We build the register that reads every order's fit reason the moment it comes back, offers an exchange ahead of a refund, and gets a restocked piece sellable again the same day.
Returns · 4 today
Wed 8 JulFit agent · today
Order #8823 · flagged size M→L before the label printed · exchange offered ahead of refund
The names above are placeholders. The size-fit call the register makes on each row is the real thing.
The register, today
Exchanges offered first
62%
of eligible returns, before a refund is ever on the table
Resolved same visit
81%
flagged, offered, and confirmed without a follow-up ticket
Fit-driven returns
34% lower
on styles with size guidance live at the product page
Restocked to sale
4.2 hrs avg
from the return scan to a sellable listing again
Where fit breaks down
Six places a shopper stops trusting the size
None of these is a returns problem, which is why a returns policy doesn't fix them. Each is a moment before checkout where a shopper can't answer a question about their own body from what's on the page, so they order the wrong size deliberately or don't order at all.
Charts
A size chart drawn for no one in particular
One S/M/L grid can't account for how a bias-cut dress and a structured blazer fit the same body differently.
Photos
A single photo carrying the whole decision
One angle, one lighting setup, one body type leaves a shopper guessing how a fabric drapes or a shade actually reads.
Search
A catalog too deep to browse
Thousands of near-duplicate basics bury the handful of pieces a shopper is actually looking for.
Checkout
Hesitation that never converts
A shopper unsure of size or fit adds two, orders none, or abandons the cart on the final screen.
Returns
A refund offered before a size fix
Without a fit reason on file, the fastest response defaults to money back, and the right size never goes back out.
Spikes
A drop day the store wasn't sized for
A restock or influencer moment sends traffic vertical, and checkout or support falls behind exactly when it matters most.
What runs the register
The layer behind every order and every return
Fit and size guidance
A short quiz and past-order history predict the right size before checkout, long before a return could arrive.
Discovery and styling
An assistant narrows a deep catalog to a handful of pieces suited to what a shopper already likes.
Return and exchange handling
Every return opens with a captured reason, and an exchange is offered ahead of any refund.
Peak-drop scaling
Search, checkout, and support hold steady when a restock or influencer moment sends traffic vertical.
Runs inside the same register your team already watches, with no separate tool to check.
Every row above reads live from the same register. Nothing here is a separate system to reconcile.
fit file · order #8823
4 stages · resolved in 3 hrs
Wide-Leg Trouser · Size M
4 past orders on file · sized up twice before
Return opened for sizing; the agent predicts M runs small on this style
Exchange to size L offered before a refund option is shown
Customer accepts; replacement queued from the nearest warehouse
Replacement shipped same day; return label attached to the original box
Flagship · the fit file
One return, followed from flag to shipped replacement
The order matters as much as the steps. A captured reason is what makes the exchange offer specific; the exchange has to come before the refund or the sale is already gone; and the restock instruction has to leave with the return so it's waiting when the box arrives.
A return opens and the fit file already knows the shopper's size history, what the style typically runs, and which resolution keeps the sale, all before a human ever opens the ticket.
- 01Every return opens with a captured reason alongside its status
- 02An exchange is offered first, with refund as the fallback
- 03Fit predictions improve from every past order on file
- 04The warehouse gets a restock instruction the moment a return ships
What happens when a real order hits the register
Point it at a live SKU and its size chart, and the register runs the same call it ran on order #8823 (reason captured, exchange offered, replacement queued) against your own catalog.
The size grid
Every size, read against how it actually returns
The register keeps a live read on which sizes run true and which ones quietly drive returns, style by style, so the chart stops treating every size the same.
XS
Runs true
6% returns
S
Runs true
8% returns
M
Runs small
15% returns
L
Runs small
17% returns
XL
Runs true
9% returns
2XL
Limited stock
12% returns
Exchange or refund
The register always tries the exchange first
Exchange
- 1Fit agent predicts the right size from the reason on file
- 2Customer confirms the swap in one tap
- 3Replacement ships same day from the nearest warehouse
- 4Original return label prints and attaches automatically
Keeps the sale and gets the customer into the size that actually fits.
Refund
- 1Reason still logged for the size-chart record
- 2Refund issued once the return scans in
- 3Item queued for inspection before it goes back on sale
- 4Customer notified at each step, so nobody's left guessing
Used when no size or style swap realistically fits the case.
The return slip
What comes back, and where it goes next
#8823
Wide-Leg Trouser · M to L
Reason: Ran small
Exchanged
Primary stock, new size
#8830
Silk Wrap Blouse · S
Reason: Color mismatch
Refunded
Primary stock
#8845
Knit Cardigan · L to M
Reason: Ran large
Exchanged
Primary stock, new size
#8852
Pleated Midi Skirt · M
Reason: Changed mind
Refunded
Outlet, imperfect tag
Related services
What else runs behind the register
Every register we build connects to the wider service work above where it fits.
Questions
What fashion teams ask first
Does it actually reduce returns, or just process them faster?
Both. Fit guidance at the product page heads off size-driven returns before they happen, and the ones that still come in get resolved as an exchange more often than a refund.
How does the register decide exchange over refund?
It reads the captured return reason against the item's size history and stock, and offers an exchange first whenever a better size or a swap is realistically available.
Will it hold up on a drop or restock day?
Search, checkout, and support are load-tested against launch-day traffic, so a spike doesn't turn into a queue of abandoned carts and missed messages.
Does it work across our sales channels?
Orders, returns, and stock stay in sync across web, app, social, and marketplace channels, so a size shows the same everywhere a shopper looks.
Does this replace our commerce platform?
No. It connects to the commerce platform, PIM, and inventory systems you already run, with no migration off them.
Elsewhere in e-commerce
More of E-commerce & Retail
Let's talk
Running a large platform, shaping a first MVP, or getting a product ready for a funding round? Tell us where you are. We'll shape the process around it, and stay with you after launch.
