We have just finished our participation in three almost consecutive events dedicated to software quality. We began in London with the National Testing Software Conference (NTSC) and Testing, Agile, DevOps & Low Code Showcase, and ended in Madrid with ExpoQA23. We attended with the intention of learning from speakers, sponsors, and attendees. We also wanted to test some of the hypotheses with which we work to design products that, like icaria TDM, contribute to improving software quality.
The World Quality Report 2021-22, published by Capgemini and Sogeti, has a positive and constructive tone, exciting about the challenges of SQA. However, some indicators reveal that there are still areas for improvement. One of these indicators is particularly worrying: only 15-20% of regression test cases have been automated. Given that our industry is precisely dedicated to automation, this data is a warning signal.
The three most important inputs of a test case are: software, user actions (whether human or automatic), and data. The software undergoes numerous revisions and unit tests, and user actions are specifically designed for each test case.
In preparing our presentations, we started from the hypothesis that testers frequently use low-quality test data. As we well know, if the input data is of low quality, the results will be equally deficient, making efficient automated execution impossible.
We focused our presentations based on this hypothesis, seeking validation or challenge from QA specialists, both in London and Madrid. Although our conversations did not provide us with statistically significant data, the qualitative information we obtained was extremely valuable, given the high level of expertise of our interlocutors.
We identified that a recurring request was to "generate" test data more efficiently. QA teams frequently use Excel sheets, create data manually, or generate data through automation scripts. However, these data present several problems: they are structurally simple, they are usually not consistent between different applications, they do not have age, they lack functional variety and richness, and they are not created with older versions of the application. In addition, they cannot be reused.
On the other hand, production data, which are usually at the origin of the errors detected by end customers or users, are complex, consistent between applications, old, varied, and have been created with previous versions of the software.
It is no surprise that artificially "generated" data is not suitable for effectively testing software.
The reality is that we have an almost unlimited amount of suitable production data for tests. The challenge lies in learning how to use them efficiently and safely, complying with GDPR: locating them, dissociating them, and delivering them on demand in the appropriate testing environment, as many times as necessary.
Icaria TDM knows how to find the client you need for each test case in production, dissociate sensitive data, and deliver it on demand in the appropriate testing environment. As many times as necessary.
If you have challenges related to test data management and want to overcome them, get in touch with us.