ASDA: Real-time Inventory and Shelf Availability Platform

Real-time retail, at scale.

An inventory and shelf-availability platform that turns high-volume retail data into timely operational decisions.

Retail / Grocery
Data engineering
Project

6.5Mtransactions per hour

1,000+stores

130Kitems

The Challenge

What stood in the way.

The central problem was not knowing what inventory should exist. It was processing enough operational data, quickly enough, to see when system stock and shelf reality were drifting apart, and to act before a customer noticed.

Inventory inaccuracy

System records showed stock as available when the shelf was empty, so decisions were made on numbers that did not match reality.

Processing latency

Large data volumes took hours to process. By the time insight arrived, the shelf had changed and the moment to act had passed.

Manual gap checks

Store colleagues spent time walking aisles to find gaps that data could have flagged earlier and more precisely.

Pricing and stock errors

High-volume price and stock updates handled slowly or inconsistently caused errors at the shelf edge and in the system.

Our Solution

How we answered it.

We treated StockWise as an operational decision system rather than a reporting project. One question guided every design choice: how do we shorten the time between a signal in the data and an action at the shelf?

ProblemTrading peaks multiply transaction and update volumes.
Our answerWe used distributed compute and elastic cloud capacity so the platform handles 6.5 million transactions an hour.
ProblemStale data describes a shelf that has already changed.
Our answerWe made data freshness the primary design target, bringing processing down from hours to minutes.
Problem12,000 users need fast, reliable views of the data.
Our answerRedis caching, Azure Front Door edge delivery and containerised services keep dashboards responsive for thousands of concurrent users.
ProblemStore operations depend on the platform every day.
Our answerWe built in production observability through Azure Monitor from the first architecture, not after a pilot.

How It Works

The parts that made the difference.

Centralised data interfaces

Inventory, pricing, sales and operational data brought into one processing environment, replacing fragmented feeds with a single, consistent source for decisions.

Distributed high-volume processing

Apache Spark on Databricks, partitioned by store and time, cuts large-volume processing from hours to minutes.

Transparent rules engine

Automated replenishment tasks and price-change alerts driven by explainable business logic that operations teams can govern and refine.

Role-focused dashboards

Azure-hosted dashboards give store, inventory and commercial teams clear, timely views of stock levels, sales trends and price changes.

Results

What changed for the client.

StockWise changed the speed, consistency and scale of ASDA's inventory operations. Store teams moved from searching aisles to fixing known gaps, with clear priorities on where to start.

Large-volume processing cut from hours to minutes
6.5 million transactions an hour processed on one platform
Fewer pricing and inventory errors through automated, rule-based updates
Improved on-shelf availability, with fewer gaps reaching the customer
Less manual effort in inventory and price management
1,000+ stores, 130,000 items and 12,000 users served from a single estate-wide platform