API test generation at ThinkDonate
ThinkDonate had less than 7 months to bring its first-of-kind fundraising platform to market. They needed to develop rapidly, while delivering the impeccable user experience required by charities and their donors.
Rigorous testing required an automated and optimised approach to API test creation, capable of generating data and scripts for a range of technologies, scenarios, and integrations.
Quality Modeller's automated test generation rapidly built a REST Assured framework, maintaining targeted API tests in two-week sprints. In a test-driven approach, test results led the documentation of API validation rules.
The benefits at a glance
- Rapid and targeted API test generation, testing rigorously within two-week sprints.
- Optimised test coverage, avoiding risky under-testing and wasteful over-testing.
- The discovery of API bugs while they remained quick and affordable to fix.
- The avoidance of costly rework in development, working from test-driven API specifications.
- A 25x reduction in test volume relative to an “exhaustive” test suite.
- Close collaboration between product owners, development, and testing.
A business and campaign critical platform launch
Testing at ThinkDonate faces a wide array of test scenarios to choose from sprint-over-sprint, created by a host of integrated technologies and logic.
Too much to test in sprint
Unable to test everything in a two-week iteration, testing must prioritise and formulate the right blend of unit, integrated and system tests. ThinkDonate’s agile methodology further called for iterative API design, occurring in parallel to development and testing. Therefore, QA must prioritise tests, without introducing negative risk.
Manual testing was not an option
Manually creating the test scenarios and data inputs would simply be too time-consuming and would not hit the range of positive and negative scenarios needed for rigorous testing. ThinkDonate were furthermore keen to “go live” as soon as possible, requiring a rapid and rigorous testing strategy to shorten time to market.
Rapid and automated API test generation
ThinkDonate required API test automation that could match this pace of development, while supporting the design and documentation of the platform.
Iterative system modelling
With no existing documentation, visual modelling initially worked backwards from the system to create a visual map of the integrated APIs. This built logically precise pictures of how data could flow through the integrated APIs, producing clear documentation of the complex system.
Optimised API test generation
Risk-based approach to testing
Coverage profiles were used to target particular logic, focussing API testing based on time and risk. This risk-based approach at ThinkDonate “rightsizes” testing during continuous delivery. Rigorous testing is not then a blocker to speed or agility, targeting testing where it is likely to have the greatest impact in finding bugs.
A successful launch – and many more to come
Read the full story to discover how ThinkDonate successfully launched their business-critical fundraising platform with Quality Modeller.
Automated API test generation
- Rigorous API testing within 2-week sprints
- 25x reduction in test volume vs exhaustive testing
- Optimal coverage, without under- or over-testing
Quality Modeller enabled rigorous in-sprint testing, while facilitating cutting-edge development practices like shift left API testing, fail-fast experimentation, and test-driven API design. It worked seamlessly alongside our teams and processes.
Johnny Pitt
Founder of ThinkDonate