According to research work by the National Institute of Standards and Technology, the US economy loses from $22.5 billion to $59.5 billion annually due to improperly organized testing – despite the fact that 25-40% of budget funds are allocated to methods and tools for QA organization. What does this mean? Only that companies don’t optimize their testing processes well enough. We will tell you about six QA pitfalls that can affect the quality of your software and slow down the release of your product.
The main problems in testing
As per research by John Capers – an American software development specialist – testing helps to detect from a quarter to a half of all bugs. If this process is not established well, the software is going to have 2-7 defects per 1000 lines of code (KLOC). The number may seem small, but this results in software-dependent systems entering the market with hundreds or thousands of missed bugs. And if some of these vulnerabilities are related to security, the consequences become more serious.
Donald Firesmith, an author of The Institute of Software Engineering’s blog, has identified common testing issues through surveys on Linkedln and on the International Council on Systems Engineering (INCOSE). His observations are several years old, but they are still relevant. As shown in a recent survey by Diffblue, 61% of US and UK companies resist the culture of software testing. The rest of the survey participants implemented test-driven development (41%); a small part (8%) write tests before coding. And only 30% of organizations believe they are leaders in testing.
Let’s take a look at six common challenges in the QA process that most companies face:
Planning problems often arise when:
- there is no separate test plan;
- the application has complex functionality and there are no requirements for it;
- there are frequent changes in requirements, delays in the development schedule, etc.
The absence of clearly defined roles and responsibilities leads to the fact that important tasks are not completed, and the critical characteristics of the system are not properly tested. Without planning, the team doesn’t know when testing will end. The process schedule is important because it helps to complete tasks at a stated time and without sacrificing software quality, functionality, and efficiency.
- Incorrect test management
The aforementioned survey by Diffblue found that only 35% of respondents always include testing time in their release schedules. Basically, 81% of developers agreed that the biggest obstacle to developing a QA culture is that the management doesn’t have enough financial resources. Working according to the leftover principle (“if there is time and money”) can’t guarantee the high-quality functioning of programs.
Another common problem is the absence of risk management. Planning risks is critical to the success of testing and its deployment. Inaccurate estimates of time and costs, constantly changing requirements, lack of technical resources, changes in the needs and interests of customers – these are just some factors significantly slowing down development. If you don’t learn to anticipate such moments, the same problems will occur during each project, preventing QA specialists from accelerating the product release.
- Lack of experience
The quality and speed of application release depend on the skills of team members. Experienced developers make fewer mistakes, and competent QA engineers find them faster.
Testers should not only be able to plan work and conduct all types of testing but also be ready to develop their skills and gain knowledge. This is because the world is following the trend towards applying DevOps culture, Artificial Intelligence, test automation, and optimization of QA processes. Hence specialists who were not previously familiar with new technologies and tools should be able to adapt to such conditions.
Amir Ghahrai, the author of the article on the role of a tester in an Agile software development model, notes that today, testers working in an Agile environment must have versatile skills, technical knowledge, collaborative abilities, and flexible thinking. Testers are under tremendous pressure to release applications faster, and companies are pushing them to change their mindset, from skillset to coding, to understand how a business operates and how to interact with customers. So testers have to develop.
The testing market is gradually shifting towards a general-purpose specialist who can automate, develop test tools, and even write production code when needed. That’s why Google, for example, ensures the consistent quality of its products by blurring the lines between testers and developers.
The obstacle of QA competencies’ shortage is also confirmed by the authors of the World Quality Report 2020-2021. They found that 42% of respondents admit their lack of professional knowledge of testing in Agile teams.
- Insufficient integration of testing and development processes
Testing is an important part of the transition from development to IT operations. DevOps culture is the best way to unite developers, testers and other team members into a tight-knit team focused on quick and high-quality results. Successful DevOps requires testing to be fully integrated into software development and delivery. This is the only way to achieve the high results illustrated by DORA’s research – release software 46 times more often, testing it 2.5 times more rapidly, and minimizing software defects.
- Problems with test automation
In recent years, test automation has become an important area in QA. For example, new ML methods are used to recognize objects and determine the scope of automated tests. However, only a few organizations have reached this level of maturity; many others are still judging the lay of the land.
According to the World Quality Report 2019-2020, companies’ problems with the implementation of automation are associated with a lack of knowledge among employees and issues of choosing the best tool. The report for 2020-2021 indicates that the automation process slows down due to either the instability of applications or the frequency of their changes. In general, research participants note that they have only 15% automation of all testing. At the same time, only 3% of respondents automate 21% or more of their test actions. We can conclude that automation brings significant benefits to businesses and economies around the world, but it doesn’t happen overnight.
- Low-quality product requirements
Quite often, product requirements are either absent or ambiguous, incomplete, and contradictory. Their misinterpretation can significantly affect the number of defects in the program.
Improved requirements are a major contributor to the reduction of software defects. This issue must get your special attention – it is necessary to minimize ambiguity, subjectivity, and misinterpretation. As a result, the number of bugs in the product decreases, despite the significant increase in functionality and complexity for new software versions. The efficiency of defect covering enhances due to the improvement of the quality of requirements, their availability in a searchable database, and a continuous assembly process.
How to solve testing problems
Watts Humphrey, the author of A Discipline for Software Engineering, stated that software developers make an average of 100-150 errors for every 1000 lines of code. Sometimes even the smallest mistyping can lead to financial losses and pose a threat to people’s lives. For example, due to a breakdown in the system of the UK National Health Service (NHS) in 2018, more than 10,000 patients were at risk of receiving the wrong medication. Taking incorrect medications could be unsafe for patients’ health and cause a public scandal.
There is no universal path in software testing: each project has its set of goals and functions. In order to ensure that QA processes run without problems, losses of quality, and slowed-down product release, it is necessary to organize them correctly. Teams need appropriate tools, skills, testing strategy, and other equally important components that optimize work and deliver results.
And we will conclude that testing is a way of thinking, a key to digital transformation, and a culture the entire team should be involved in.