Asda
Transforming Asda’s Customer Analytics with Azure Databricks at Enterprise Scale
Following its separation from Walmart, Asda embarked on a major transformation to establish a modern customer analytics platform on Microsoft Azure.
The retailer required a trusted, enterprise-wide view of its customers to unify customer data across business domains, improve reporting and enable data-driven decision-making at scale. Working alongside Asda’s Decision Science & Data teams, Nimble Approach delivered an end-to-end Azure Databricks solution, introducing a Single Member View, governed self-service reporting and a repeatable Declarative Automation Bundles (DABs) deployment pattern that established modern enterprise engineering practices.
The platform significantly improved processing performance, increased storage efficiency, accelerated delivery of new data products and equipped internal teams with the capability to independently maintain and extend the solution.
Asda’s Objective
Enabling Enterprise Customer Analytics
Following its transition to an independent technology estate, Asda required a modern customer analytics platform capable of supporting millions of customers across stores and online. The retailer needed a solution that would unify customer data from multiple business domains, replace legacy customer matching capabilities and provide a trusted, enterprise-wide customer view to support analytics and operational decision-making.
Alongside building this new capability, Asda sought to establish consistent engineering standards, improve reporting efficiency and provide business users with governed access to reliable data. The retailer also needed to modernise its software engineering practices, replacing manually managed Databricks notebooks with a repeatable, source-controlled deployment approach that would support enterprise-scale delivery. The platform needed to support enterprise-scale processing, integrate first- and third-party data sources and enable teams across commercial, marketing and operational functions to make faster, data-informed decisions while developing long-term internal capability.
Opportunities Identified for Asda
- To disrupt the market, but a need to stay ahead on innovation and agility
- To improve speed to market and product resilience through technical and process improvements
- To upskill the relatively junior data product engineering squads, especially in Agile working practices and CI/CD development patterns
- To improve data and tooling reliability, to combat issues stemming from teething problems with a new tech stack
- To standardise best practice engineering patterns
Nimble’s Solution
Delivering a Modern Azure Data Platform
Working as an embedded partner, Nimble Approach helped Asda design, build and adopt an enterprise-scale customer analytics platform on Microsoft Azure. Azure Databricks formed the core of the solution, with Azure Data Factory orchestrating data ingestion and PySpark pipelines transforming customer, transaction and third-party data into trusted analytical assets.
At the centre of the platform was a Single Member View, providing a governed, person-level customer record through a unique master identifier. This enabled consistent customer analytics, customer lifetime value analysis and governed self-service reporting across multiple business domains.
A key element of the transformation was the design and rollout of a repeatable Declarative Automation Bundles (DABs) deployment pattern. This established a gold-standard engineering framework, replacing manually managed notebooks with source-controlled, code-based data pipelines incorporating automated testing, quality gates and reusable templates. The approach dramatically accelerated delivery of new data pipelines while embedding enterprise-grade engineering standards across Asda’s data teams.
Agile coaching, paired delivery and comprehensive documentation enabled knowledge transfer throughout the programme, allowing Asda to confidently own, operate and continue evolving the platform after delivery.
Accelerating Insight Through Trusted Data
The platform established a trusted, enterprise-wide foundation for customer analytics, creating a single governed source of customer data that supports reporting, advanced analytics and business decision-making across multiple domains. The Single Member View introduced one master identifier per customer, creating a consistent view of millions of customer records across previously disconnected data sources.
Performance improvements transformed the speed at which analytics could be delivered. Basket table processing reduced from two days to 1.5 hours, while customer database processing fell from 12 hours to just 28 minutes, enabling daily person-level reporting that was previously impractical. The introduction of standardised Declarative Automation Bundles fundamentally modernised engineering practices, enabling reusable deployment patterns, integrated source control, automated testing and significantly faster delivery of new data pipelines. Work that initially required weeks could subsequently be delivered in hours or days, allowing Asda to respond rapidly to new business requirements while maintaining consistent engineering standards.
The platform also delivered significant operational efficiencies. Storage optimisation reduced Azure costs by 65%, while the redesigned Customer Lifetime Value model expanded historical analysis from four to 40 years whilst reducing storage requirements by over 80%. Business users gained governed self-service reporting capabilities, replacing manual Excel-based processes with auditable analytics, and embedded coaching ensured Asda’s teams now independently own, maintain and continue to evolve the platform.
97% Faster Basket Tables
Runtime down from 2 days to 1.5 hours.
24x Faster Database
From 12 hours to 28 minutes runtime for customer database processing.
65% Improvement in storage efficiency
Reducing ongoing Azure costs
97% Faster Basket Tables
Runtime down from 2 days to 1.5 hours.
”Your hands-on focus on engineering best practice alongside the product development itself is exactly the support I needed
Jonty HaberfieldHead of Data Product and Engineering at Asda
The Technology Stack
In collaboration with DS&D, Nimble carefully evaluated the available technologies in the current tech stack, architecture, and in-house capabilities to select the best technology for sustainable growth.
- Azure – as the foundational cloud environment
- Azure Databricks – for data transformation, machine learning, and governance
- Azure Data Factory – for efficient and low-code data operations and powerful data transformation
- Pyspark – to handle large-scale big data processing
- Python – for automation and rigorous testing
- SQL – for robust database management and data analytics
- GitHub – for cloud-based code repository management
- Declarative Automation Bundles (DABs)
- Jira
Nimble are also continuing to input into the evolution of the DS&D tech stack, including through their ML Ops Operating Model work.
Discover what's possible for you
Speak to one of our projects team to see how Nimble could help your organisation.




