Hero
Commerce Systems
Inventory, POS & PLM
Apparel Retail

RFID-Unified Inventory and POS for an Apparel Retailer


Let's Connect

Overview

What we built

A 55-store apparel retailer's online shop routinely oversold items that did not exist and hid stock sitting in shops ten minutes from the customer, because the POS, online store and warehouse each kept their own version of the truth. We built a real-time inventory backbone on RFID, and record accuracy rose from roughly 74% to 98%.

In plain terms: this 55-store apparel retailer could not trust its own stock numbers. The online shop would confidently sell an item that was not actually there, while a customer ten minutes from a shop with that exact item in stock had no way to know it. Every delivery pushed store counts further out of step with the ERP within days, a full cycle count took an entire closing day per store to complete, and the POS, e-commerce platform and warehouse each kept their own separate version of what stock actually existed.

We built a real-time inventory backbone on RFID: item-level tags read at goods-in, on the shop floor and at the till stream events into a single central stock ledger that the POS, online store and warehouse all read from together. Ship-from-store routing then exposes that shop inventory to online orders, and handheld readers cut a store's full cycle count from a closing day to under two hours. Inventory record accuracy rose from roughly 74% to 98% across all 55 stores, online order cancellations from overselling dropped from 6.8% to 0.9%, and within six months of ship-from-store launch, 27% of online orders were being fulfilled from store stock.

The Problem

Stock records nobody trusted

This apparel retailer ran 55 stores plus an online shop, and the two channels routinely disagreed about what stock actually existed. The website would confidently oversell items that existed nowhere in the business, while stock genuinely sitting on a shop floor ten minutes from a customer stayed invisible to the online order that could have used it.

The gap started at the store level. Store counts drifted away from the ERP within days of every delivery, so the system of record was already wrong before a single customer walked in. Fixing it manually meant a full cycle count, and that took an entire closing day per store, time the business could not spend selling.

Underneath all of it sat three disconnected systems. The POS, the e-commerce platform and the warehouse each kept their own version of the truth, with nothing forcing them to agree, so an accurate answer to a simple question, how much of an item do we actually have, depended on which system anyone happened to ask.

Overselling online

The online shop routinely sold items that existed nowhere in the business, creating orders that could never actually be fulfilled.

Hidden nearby stock

Stock sitting in a shop ten minutes from the customer stayed completely invisible to the online order that could have used it.

Counts drifting fast

Store counts drifted away from the ERP within days of every delivery, so the system of record was wrong almost as soon as stock arrived.

Three separate truths

The POS, e-commerce platform and warehouse each kept their own version of stock levels, with nothing forcing the three to agree.

What it was costing them

Every oversold online order became a cancellation, a refund and an apology, while stock sitting a few minutes from that same customer went unsold. A full closing day lost to cycle counts at each of 55 stores was a day of trading given up just to find out what was actually on the shelves, and three disagreeing systems meant nobody could act on a stock number with real confidence.

The Solution

Real-time RFID inventory backbone

We built a real-time inventory backbone on RFID rather than trying to patch the gap between three disconnected systems. Item-level tags are read at goods-in, on the shop floor and at the till, and every one of those reads streams straight into a single central stock ledger rather than sitting in a system-specific silo.

The POS, online store and warehouse all read from that same ledger, so for the first time all three agree on what actually exists at any given moment. Ship-from-store routing then puts that shared truth to work, exposing shop stock to online orders instead of leaving it invisible to a nearby customer.

Handheld RFID readers replaced the manual cycle count too, cutting a store's full stock check from a closing day down to under two hours, so accuracy stopped depending on how carefully a full day of manual counting had gone, and stores could recheck stock far more often than once in a blue moon.

Key decisions

01

RFID tags at every touchpoint

Item-level tags are read at goods-in, on the shop floor and at the till, capturing stock movement as it actually happens rather than after the fact.

02

One central stock ledger

Every RFID read streams into a single central ledger, replacing three systems that each held their own version of the truth.

03

Routing orders to shop stock

Ship-from-store routing lets an online order draw on stock sitting in a nearby shop instead of ignoring it because the systems could not see it.

04

Handheld readers for cycle counts

Handheld RFID readers replaced a closing-day manual count with a check that takes under two hours, so counts can happen far more often.

Measurable Impact

What changed after launch

Trust in the numbers came back fast: inventory record accuracy rose from roughly 74% to 98% across all 55 stores, and online order cancellations caused by overselling dropped from 6.8% to 0.9%. Customers stopped ordering items that did not exist, and the business stopped apologising for it.

Ship-from-store turned previously invisible shop stock into a fulfilment channel of its own: within six months of launch, 27% of online orders were being fulfilled from store stock. And the full-store cycle count that used to consume a closing day now takes under two hours with handheld RFID readers.

Inventory accuracy

Roughly 74% inventory record accuracy

98% inventory record accuracy across 55 stores

Order cancellations

6.8% of online orders cancelled from overselling

Cancellations from overselling down to 0.9%

Store fulfilment

Shop stock invisible to online orders

27% of online orders fulfilled from stores

Cycle counts

A full closing day per store

Under two hours with handheld RFID readers

Headline results

Inventory record accuracy rose from roughly 74% to 98% across all 55 stores

Online order cancellations caused by overselling dropped from 6.8% to 0.9%

27% of online orders fulfilled from store stock within 6 months of ship-from-store launch

Full-store cycle counts cut from a closing day to under 2 hours with handheld RFID readers

Tech & Tools Used

What powered the build

Every tool below earned its place in this engagement. Here is the part each one played.

Impinj RAIN RFID readers

Captures item-level reads at goods-in, on the shop floor and at the till across all 55 stores, the raw signal the whole inventory backbone runs on.

Apache Kafka logo

Apache Kafka

Streams every RFID read as an event from goods-in, shop floor and till straight into the central stock ledger without delay.

Node.js logo

Node.js

Runs the services that consume the RFID event stream and keep the central stock ledger consistent for the POS, online store and warehouse.

PostgreSQL logo

PostgreSQL

Stores the central stock ledger itself, the single source the POS, online store and warehouse all read from instead of their own separate records.

Redis logo

Redis

Caches current stock levels so ship-from-store routing can check nearby shop inventory instantly the moment an online order actually comes in.

GraphQL logo

GraphQL

Exposes the unified stock ledger to the POS, online store and ship-from-store routing through one consistent API instead of three separate ones.

Next.js logo

Next.js

Powers the online storefront and the internal dashboards store teams use to see accurate, real-time stock across all 55 stores.

Docker logo

Docker

Packages the inventory backbone's services consistently, so the same stock ledger logic runs identically across every one of the 55 stores.

AWS ECS

Hosts the inventory and ledger services, scaling to handle RFID reads streaming continuously from goods-in, shop floors and tills across 55 stores.

Ready to Build your Apparel Retail Business with Inventory, POS & PLM

Ask Byte

Ask Byte

Typically replies instantly

just Now

Hi! I'm OrganByte's assistant. How can I help you today?

AI-generated content may be incorrect


OrganByte

Building innovative software solutions that transform businesses and drive digital success.

© 2026 YourCompany. All rights reserved.