With the advent of DevOps, testing plays more of a starring role in developing financial services software, but many still view it as an annoying bottleneck. Within that, the biggest bottleneck of all is locating quality test data (the information against which to carry out a test). Searching or waiting for someone to generate that information is a big issue for testers, so it can be very tempting to compromise on those processes, leading to problems down the line. Fortunately, like so many aspects of software development, test data has also evolved. Here are five essential things to know about test data in 2022: 

1 Masking and subsetting data can help

This process works by taking a subset (in other words, just part) of a production database and then masking the data (full names, credit card details and so on), making it unrecognisable for confidentiality and compliance purposes. The benefits of masking and subsetting are that testers are presented with sufficient information, plus of course, it is real and accurate data. The downside to masking and subsetting data is that it can be difficult to preserve referential integrity, leading to possible security holes and incorrect data. 

2 Synthetic testing adds speed, accuracy and compliance

Another technique growing in popularity when testing financial software is creating artificial or synthetic data, which means the data is built from scratch without any production data. Synthetic bypasses compliance issues, and the advantage of synthetic information is that it can be spun up relatively fast. Plus, once it has been created, it can be used again and again across different test scenarios. Synthetic data generation is particularly useful in small teams, Agile and DevOps operations because it does not hold up a sprint or other fast turnaround requirements. In addition, synthetic data generated through cloud-based tools can be scaled up and down as required. 

3 Developers and testers can both make use of test data

Traditionally, testing was left to dedicated professionals. However, that has changed rapidly over the past few years, with more testing happening earlier in the development process and getting more developers involved, even those with no testing background. This has become a necessity because of financial software's sheer scale and complexity. A number of testing tools are cloud-based, which means that they are accessible to a broad range of users with greater ease.

4 Test data can be used for many different areas of application testing

Conventionally, test data is mainly used for functional, performance and regression testing, but it can also be used for mock services, also known as virtual services. A mock service imitates a request/response from an actual service, such as checking a log-in sequence with a backend server. Carrying out that task against the real service could create a delay due to the unavailability of the server and any associated access costs. Obviously, a mock service is just a simulation, but it allows developers and testers to keep moving forward when the actual service is not available. If the mock service can utilise dynamic test data, that simulation can become far more accurate and support a variety of potential requests and responses. 

5 Test data solutions can be multi-layered

Which to choose: on-premises versus cloud, synthetic versus masked data? The answer is that there is no need to choose, and most banks and other financial services organisations will benefit from having a blend of capabilities. Of course, generating millions or even billions of lines of test data is going to require an on-premises test data management solution. However, when greater agility and speed are needed, synthetic data generation through cloud-based tools has an increasingly valuable role in the testing sphere. Traditional and more recent evolutions in test data solutions can happily co-exist and complement one another. 

Financial software development requires teams to provide continuous updates to their web and mobile applications faster. Coupled with increasingly complex code and strict compliance regulations, test data can be a real challenge for many financial organisations. So, new ways to accelerate and improve code validation processes become essential because that is where many bottlenecks lie. The good news for anyone involved in checking software quality is that generating compliant and accurate test data no longer needs to be a roadblock.