Distributed test automation that cut UAT from half a day to 30 minutes

The Problem

A major UK insurance company had a UAT process that took approximately half a day to complete. For a business integrating with price comparison platforms like GoCompare and Compare the Market, where SLAs matter, this testing bottleneck was limiting release velocity.

The Testing Challenge

• Sequential execution: Tests ran one after another
• Long feedback cycles: Half a day before knowing if changes worked
• Platform integration pressure: Price comparison platforms required reliable, fast quoting

The Technology Context

The existing technology stack included QuickTest Professional (QTP) for test automation, running against a codebase spanning procedural PHP, VB6, and early .NET technologies.

Our Approach

We built a distributed test automation system that parallelised test execution.

Distributed Architecture

• Multiple test runners: Work distributed across parallel execution nodes
• Workload distribution: Tests allocated across available runners
• Result aggregation: Centralised reporting from distributed execution

This was early experience with distributed systems architecture - patterns that would later become known as microservices.

Optimised for Longest Test

The system reduced total execution time to approximately the duration of the longest individual test. Parallelisation eliminated the sequential bottleneck.

Integration with Existing Tools

Built to work with the existing QTP infrastructure, requiring no changes to the test suite itself - only to how tests were scheduled and executed.

Outcomes

Execution Time

• UAT reduced from half a day to approximately 30 minutes
• Limited only by the longest individual test
• Parallelisation delivering near-linear speedup

Development Velocity

• Faster feedback cycles for development team
• More releases possible within the same timeframe
• Reduced cost of testing bottleneck

Platform Integration Support

• Faster iteration on price comparison platform integrations
• Ability to verify SLA compliance more frequently
• Quicker response to platform requirement changes

Technical Foundation

This project provided first experience with distributed systems architecture. The patterns learned here - workload distribution, result aggregation, parallel execution - informed architectural decisions in every subsequent engagement.