Automatic Test Data Generation for App Development | STAND 8 | Stand8
Sometimes building enterprise software can feel like steering the Titanic, and in the healthcare domain, there are industry-specific challenges that make it even more challenging.
A major healthcare provider had two critical platforms on quarterly release schedules. Unfortunately, updates had slowed to every 6 months due to manual testing. End-to-end testing was automated in their mobile application, but all other testing was done manually.
A closer look at the numbers shows why their releases had stagnated.
60% of the team’s time was spent setting up testing. Why so long? Test data had to be requested from external teams and consisted of multiple upstream business processes. Once teams got the data and ran the manual tests the best they could achieve was 40% test coverage. This return for the effort was unacceptable.
Instead of relying on the external teams, STAND 8's team used metadata-based test data generation. Using actual data, our team built a model that generated test data that met all necessary requirements. This erased dependencies that were making testing take so long.
How was this data delivered?
The synthetic test data generation was embedded in a java-based service that provided data for a specific test when it receives an API call. This data is essentially on-demand. There was practically no wait time. Synthetic test data was created in an as-needed, real-time model eliminating the need to coordinate with multiple teams for manually collected test data.
The result of this solution was huge time savings and lean operational costs. Where before it had taken 3 weeks to gather data from separate teams and run tests, it now only took 4 minutes. Previously, large teams of manual testers were spending as much as 500 hours per release. Now all of that expense was saved or allocated elsewhere.