Skip to main

Download our latest report - Protecting & growing your payments business -

here Opens in a new window
Back to Building Thredd

Test sizing strategies for fast, stable feedback.

Optimising automated quality gates for scalable payment solutions.

Adam Mitchell

January 20, 2025

In fast-paced fintech environments, speed and reliability in software development are non-negotiable. As an issuer processor, Thredd is committed to delivering high-quality, scalable payment solutions that adapt to evolving customer needs. One of the ways we achieve this is by optimising automated quality gates through an innovative Test Sizing strategy—designed to balance speed, stability, and efficiency across our software development lifecycle (SDLC).

The Risks of Inefficient Quality Gates 

Automated quality gates play a critical role in maintaining software reliability, ensuring that every release meets the highest standards before reaching production. However, without a structured testing strategy, these gates can become bottlenecks, leading to: 

  1. 1
    Extended feedback loops

    Slow-running test suites delay engineering workflows, reducing agility.

  2. 2
    Delayed bug detection

    Issues found late in development require costly fixes and additional resources.

  3. 3
    Erosion of trust

    Flaky or inconsistent tests undermine confidence in automation, prompting unnecessary manual interventions.

  4. 4
    Resource inefficiency

    Ineffective tests consume unnecessary computational power and engineering effort.

  5. 5
    Ineffective tests consume unnecessary computational power and engineering effort.

    Without efficient early-stage testing, critical issues can escalate to production.

    To overcome these risks, we have adopted a Test Sizing strategy, ensuring fast and reliable feedback while maintaining test stability and efficiency. 

What is Test Sizing? 

Originally developed at Google, Test Sizing categorises tests based on execution time and dependency complexity, helping teams optimise performance without sacrificing accuracy. 

Tests are classified into three categories based on their constraints and resource requirements: 

  1. 1
    Small Tests

    Fast, isolated, and stable. No external dependencies (e.g., network, database, or filesystem access). Ideal for unit testing. 

  2. 2
    Medium Tests

    Slightly more flexibility, allowing local database interactions but still restricted from external network access. Used for integration testing. 

  3. 3
    Large Tests

    Full-system, end-to-end testing, incorporating all dependencies, including network access and multiple services. Typically used for system validation. 

Test Sizing Matrix 

By prioritising small tests wherever possible, we ensure Quick Confidence in our releases, catching issues early and reducing reliance on time-consuming, resource-intensive large tests.

Feature 

Small 

Medium 

Large 

Network Access 

No 

Localhost only 

Yes 

Database 

No 

Yes 

Yes 

File System Access 

No 

Yes 

Yes 

Use External Systems 

No 

Discouraged 

Yes 

Multiple Threads 

No 

Yes 

Yes 

Sleep Statements 

No 

Yes 

Yes 

System Properties 

No 

Yes 

Yes 

Time Limit (seconds) 

60 

300 

900+ 

 

Applying Test Sizing at Thredd: ‘Think Small First’ 

At Thredd, we have fully embraced a Think Small First philosophy, optimising our SDLC with structured test efficiency principles: 

  1. 1
    Prioritise Small Tests for Early Confidence
    • Ensure foundational functionality is validated with fast, reliable unit tests. 
    • Drive greater testability in designs by reducing dependencies. 
    • Convert traditionally large, end-to-end tests into isolated, stable alternatives. 
  2. 2
    Align Test Sizing with a Robust Framework
    • Combine Mike Cohn’s Test Pyramid with Test Sizing to structure test execution layers. 
    • Ensure quick feedback loops at each pipeline stage, reducing unnecessary test execution. 
    • Promote a structured approach to automated quality gate optimisation. 
  3. 3
    Foster a Scalable, High-Performance Testing Culture
    • Enable engineering teams with tools and training to implement efficient testing practices. 
    • Drive adoption of ‘In-a-Box’ testing methodologies, encapsulating functionality within controlled environments. 
    • Reduce resource-intensive full-system tests while maintaining high validation standards. 

The Impact: Faster, More Reliable Payment Platform Development 

By implementing Test Sizing, we have transformed how Thredd manages quality in a high-speed development environment: 

  • Increased reliability – Reducing flaky tests and unnecessary manual interventions. 
  • Optimised resource allocation – Ensuring computational and human efforts are focused on value-driven activities. 
  • Faster release cycles – Enabling frequent, stable deployments without compromising quality. 
  • Greater developer confidence – Providing clear test categorisation for better predictability and efficiency. 

 

As Thredd continues to scale, our structured testing strategies will ensure we maintain agility and reliability, delivering the best possible experience for our customers and partners. 

Discover more about Thredd’s approach to scalable payment technology. 

More technical insights

News, events and insights.

Sign up to receive Industry news, events and insights delivered straight to your inbox.

Please enter your first name
Please enter your last name
Please enter a valid email address
Please select a country
Please accept the terms and conditions to subscribe