For its annual Global Quality Report, Capgemini and Sogeti surveyed thousands of technologists and technology leaders to learn about the state of software quality practices. The Global Quality Report 2021-22, sponsored by Micro Focus, is the 13th such effort from Capgemini and Sogeti, giving authors more than a decade of data to show not only the state of software quality, but how the trends – and the culture of quality– have changed.
At 72 pages, the report is very comprehensive, and what this year’s report reveals may surprise you. You’ll want to dive into all the details, but for a first-line summary, here are eight key points from the latest Global Quality Report – everything you need to know, plus comments from the authors and other experts. , who shared their views. results.
1. Testing and quality play a bigger role
The first number in the report concerns the goals of quality assurance: detecting problems early, moving faster, and contributing to business growth or customer satisfaction. The percentages only range from 58% to 62% and haven’t changed much over the past four years.
Figure 1: The objectives of quality assurance and testing. Source: Capgemini / Sogeti
In other words, almost everyone wants everything. As a consultant, I would be tempted to read this as immaturity, but the authors of the report see it differently: that management ultimately sees QA as more than just finding flaws. The slight decline in QA’s role as a “quality keeper” could be due to the fact that testing has been pushed beyond one specific person, on all members of the delivery team. software.
Andrew Fullen, head of innovation and technology for Sogeti in the UK, said this means a broader role and perception of quality.
“The industry feels like it has grown up now, it has matured, it looks at a bigger picture. She rejects things that are not important, enters a more interconnected world with more tools to support it. Just because all of these things are important doesn’t mean that every project will do all of these things in the same report. “
2. Increasing the quality of AI expands what testers can accomplish
For years, industry leaders and consultants have complained about the hype around AI. But the report found that AI offloads certain cognitive tasks from humans to the computer. “These results show the need for organizations to advance their modernization initiatives through continuous quality tools and AI-based test automation,” said Rohit De Souza, Senior Vice President and General Manager of Micro Focus’s ITOM and ADM Product Groups and Head of the Technical and Product Safety Director’s Office.
“There are a few big things when it comes to AI-powered testing,” Micro Focus chief technologist said. Don jackson. “The first, and probably the heaviest, is the use of computer vision, so that the computer sees things the same way as a human. In the past, testers had to use properties exposed by the operating system, and each time they crossed a platform, those properties changed for the same object. You can have up to four different scripts for the same business process, depending on the platform you are running on. “With computer vision, it works on the visual characteristics of objects just like a human does,” he said.
Tools without code, for example, can use natural language processing (NLP), allowing a non-technical human to create and update tests. Computer vision can recognize a shopping cart image, so if the icon changes or moves, the software can still find the button to click.
The combination means that the tests don’t need to use precise localization strings or classic code (Java, C #, Ruby), which makes them less brittle. Log query and analysis tools allow you to perform risk-based testing with usage data. None of these strategies are intended to replace humans. Rather, they are approaches that increase human performance.
3. Tests can and do improve.
The following figure assesses success in achieving the goals described above. Only half of those polled said the tests were optimized for speed and cost. This indicates that there is a lot of room for improvement. Yet the same response last year only reached 37%. This is a significant improvement in just a year, and it could be due to the way organizations have responded to COVID-19.
Figure 2: The success of the QA group in achieving specific goals. Source: Capgemini / Sogeti
4. The COVID-19 response has been integrated around the world
Last year’s report had an entire section on how COVID-19 had affected software quality. For example, financial services and industry have been slow to embrace cloud computing, which has made the transition to remote working more difficult. The pressures of time, resources and availability induced by the pandemic have made companies more open to out-of-the-box approaches, not only to remote working, but also to participatory testing.
As more and more employees had to perform remote testing, there was also a demand for data masking services, to transform and anonymize production data. A year later, the Covid problem had “faded in the woods” of software development, and most companies have incorporated these changes into the way they deliver software, Jackson said. What used to be housing, like remote working, is now part of everyday work. So instead of a separate section for pandemic-related responses, responses to COVID-related questions have been incorporated into other aspects of the report.
Over the past year, many companies in Silicon Valley have switched to tele-support for testers. This is a huge change, allowing skilled workers across the United States to compete for work at Apple, Google, Facebook and other now distant companies.
This will reduce the supply of local testing and quality related jobs on the high end, especially jobs where employees have to come to the office. Businesses that don’t allow remote working or that insist on a return-to-office strategy will face recruitment, retention and morale challenges over the coming year. Many employees are now used to and prefer to work from home, or have moved to areas where the cost of living is lower or the quality of life is higher.
5. The boundaries have blurred between testing, quality, agility, DevOps and digital transformation
It is striking how much of these topics overlapped in previous World Quality Reports and how much they converged. Instead of a “Testing Agile” survey, a “Digital Transformation” survey, a “Testing in DevOps” survey, and so on, this year the report combined them into one.
While the most recent WQR contained a section on digital transformation, it becomes clear that agility, DevOps, and quality are all prerequisites, or perhaps the mechanism by which digital transformation is possible.
6. Test tooling and automation become the norm
Of all the improvement priorities to work on, test automation ranks lowest, with just 46% seeing it as a priority. The temptation is to read this as “not important”, but again, that can mean companies are testing successfully. In these cases, test tooling has shifted from a strategic initiative to a part of everyday life. As Sogeti’s Fullen puts it, “I’d much rather have a small percentage of good quality automation that makes a difference rather than automating everything.”
Shivakurma Balasubramaniyan, vice president of financial services, insurance and data quality for Capgemini – and another co-author of the report – adds that companies are finally starting to understand they need to hire and develop software development skills in test.
This means that when they recognize a gap, they do not have an “automation initiative”. Instead, they call it hiring, job description, restructuring, or training.
Figure 3: The importance of different factors to make testing more effective. Source: Capgemini / Sogeti
7. It is increasingly recognized that testing is in fact a skill set.
Balasubramaniyan’s comment leads to the team’s most popular success factor – having the right skills. It goes beyond automation, AI, DevOps, continuous integration and other aspects, but more importantly for real testing and test design.
Kirthy Chennian, vice president for North America at Capgemini and co-author of the report, noted that the biggest challenge interviewees faced was getting sufficient agile testing skills within software teams. At first, this might seem odd, considering how many years since many respondents reported completing an Agile Change.
Chennian said it’s because the person doing the job has moved from a specialized role to everyone involved in delivering software. After a few years of assuming that testing was easy and anyone could do it, organizations realize that this is a repeatable skill that can be developed, trained, and hired.
8. Use of Docker / Kubernetes for testing and test environments is gaining momentum
When businesses test with tools, they need servers to run with the latest version. Just over half of the companies surveyed have on-demand and fit-for-use test environments, and the de facto standard for this appears to be the Docker image, likely running in a Kubernetes cluster. This provides a relatively straightforward test path for businesses, especially since all of the major cloud providers support Kubernetes.
What is the next step in quality control
More than half of the companies surveyed have on-demand testing environments, meaning almost half do not. The real questions are what matters to your organization and where your organization stands relative to its competition. If you don’t have an on-demand testing environment, the new World Quality Report can be a call to action.
These key takeaways are top notch and do not fit into particular industries such as automotive, energy, financial services, healthcare, government, high tech, or telecommunications. But the more detailed breakdowns in the report devote 20 pages to sectoral data. Download the full Global Quality Report 2021-22 to dive deep.