The IT sector accounts for about 4% of global emissions — a number that could potentially double by 2025. This underscores the urgency of a sustainable green IT transformation.
As the digital world grapples with its environmental footprint, the Sustainable QA Practices for Green IT webinar brought together thought leaders from Sogeti, Wipro, and Tricentis to shed light on integrating sustainability into quality assurance and testing. This blog post is a synthesized interpretation of the key discussions from the webinar, aiming to encapsulate the essence and insights of the expert panel.
The drive towards sustainable QA practices: Why the urgency?
Richard Wilkinson from Wipro ignited the discussion, emphasizing that sustainable QA is the right thing to do, aside from the growing external pressures from customers, employees, and regulators that is driving the IT industry towards sustainability.
Andrew Fullen of Sogeti UK pointed out the IT sector’s significant role in environmental pollution, comparing its impact to that of the airline industry. This comparison highlights a pressing need for the IT industry to be proactive in reducing its environmental impact, versus being reactive. He claimed we need to look at it 10 years from now, setting clear goals to build a positive agenda for change.
In the webinar, Fullen also illuminated insights from the 2023 World Quality Report around sustainability in the IT sector. He noted that a significant majority of the industry was lagging in addressing sustainability, though there were some exceptions. However, most stakeholders recognized the critical need to engage in this conversation, driven not only by the demands of customers and investors but also by the necessity to attract and retain talent. Employees are increasingly evaluating the green credentials of organizations, making sustainability a factor that could directly impact the bottom line.
Surprisingly, while some were still adopting a wait-and-see approach, about two-thirds of respondents expressed a strong desire to take immediate action. This urgency was echoed by senior stakeholders who exhibited a genuine passion for addressing environmental concerns, acknowledging the noticeable changes occurring within their lifetimes and the pressing need to respond proactively. Andrew’s comments highlight a growing awareness and a shift towards prioritizing sustainability within the IT sector.
Beyond greenwashing: How serious are we about this?
The panel tackled the sensitive topic of greenwashing, when green PR and green marketing are deceptively used to persuade the public that an organization itself is green. While acknowledging its negative connotations, Fullen suggested that any movement towards green awareness, even if initially superficial, is a step in the right direction. He pointed out results from the 2023 World Quality Report stating that 67% of investors in organizations are really focused on improving sustainability, while 97% of people working with organizations think about sustainability of a project. This perspective suggests that starting with green intentions can lead to more substantial and meaningful environmental actions.
How to integrate sustainability into QA processes: Challenges and strategies
When discussing how to incorporate sustainability into QA processes, Fullen highlighted a general lack of preparedness but also a passionate desire for change among industry leaders. He highlighted some practical measures, such as developing metrics to gauge the environmental impact of QA activities. This could include measuring how long it takes to run something, particularly understanding the green credentials of the regression pack you’re about to run. Reducing the number of tests that are being run could equate to a savings of 10 consumption hours. This may not seem like a lot, but you cannot improve what you don’t measure.
Effective tools and methodologies for sustainable QA: Tricentis insights
Kashif Husain from Tricentis shared his expertise on implementing sustainable practices in QA, emphasizing a multi-faceted approach. A primary focus is on reducing energy usage, where less time spent on test execution and creation directly translates to lower energy consumption. He highlighted the significant problem of electronic waste, noting that a vast amount of e-waste is generated annually and only a small percentage is recycled.
In terms of specific methodologies, Tricentis adopts a model-based approach to ensure reusability, repeatability, and maintainability of tests. This approach, coupled with risk-based testing like smart impact analysis, can significantly reduce the number of necessary tests in a regression pack by about 80-90%, leading to less time and energy spent on test creation. Additionally, Husain pointed out that automation can drastically reduce the time for test execution, mentioning a reduction from weeks to just a couple of hours, which is a substantial decrease in energy consumption.
Further, he discussed the efficiency of cloud-based testing over local machines, citing the competition among major cloud service providers to be the greenest as a factor in their efficiency. He also highlighted Tricentis Device Cloud, which allows for efficient mobile device access in a cloud infrastructure by reducing the need for maintaining a large inventory of physical devices. Husain stressed the importance of performance testing, not only for QA efficiency but also for making applications more energy-efficient. He concluded by highlighting the issue of redundant, old data in systems, advocating for data quality tools like Tricentis Data Integrity to manage and reduce this data, thereby improving system efficiency and reducing emissions.
Metrics and KPIs: Measuring sustainability in QA
Richard Wilkinson shared insights on the evolution of sustainability metrics in the IT industry. He stressed the importance of tracking how much certain testing or development activities cost in terms of energy usage, along with their impact, which can provide clear targets for optimization.
Wilkinson then went on to provide some examples from the banking sector where it is important to track how much energy is used in certain user journeys, from joining a bank to receiving an annual report on all the customer’s accounts. Once the energy expenditure for different user journeys is understood, it’s possible to optimize those processes and reallocate that energy use to other products and services that are important. Optimization efforts can involve data optimization or rewriting code into a more efficient language. Rewriting code of course introduces risk, so even though you may not use less energy, you need to ensure that there’s no business impact. Having test tools in place where you target specifically the areas that have the most impact is certainly useful. Once you’ve detailed all these important KPIs across the software development and testing process, it will be clear what to improve.
The Gen AI dilemma: Balancing AI advancements with sustainability goals
How do we reconcile sustainability goals with the energy-intensive nature of Gen AI technologies and QA? As Simona Domazetoska from Tricentis posed this question, she pointed out that training a large language model like ChatGPT can consume energy equivalent to the yearly electricity consumption of over 1,000 U.S. households. Additionally, the 2023 World Quality Report highlighted that the second biggest roadblock to adopting generative AI in quality was indeed sustainability, with 35% respondents raising concerns.
The panel agreed that while AI’s benefits are undeniable, its environmental costs must be considered. Wilkinson emphasized the need for cost-benefit analyses that include environmental impacts to ensure responsible AI utilization. Wilkinson also stressed the importance of balancing the environmental impact of AI against its benefits. He suggested using a dual analysis approach, considering both financial and environmental costs, to justify AI implementation. The key message was that if AI’s benefits surpass its environmental costs, its use is warranted.
Husain highlighted that there’s an arms race going on around AI, with so many vendors in the marketplace. This can help with speed of execution, analysis, and coding practices, but the worry is that people will not be using Gen AI for those purposes. There is potential for a high carbon footprint of inefficient AI use. Will AI, from an investment perspective but also considering energy consumption, benefit us in the long run? This is something we need to be wary of, while also adopting Gen AI to make QA more efficient.
Fullen highlighted the potential drawbacks of AI, particularly its substantial carbon footprint when things go awry, as exemplified by recent issues with ChatGPT. That situation prompted a reevaluation among organizations that had heavily invested in AI. Andrew emphasized that while AI can be beneficial for tasks like data analysis or generating content, often the lessons learned from AI are more effectively applied as efficient, simpler rules. He cited the camera industry as an example, where AI is used for training but the end product incorporates more energy-efficient, low-cost solutions that are derived from AI learnings. Fullen also referenced space exploration, where the high energy costs and constraints necessitate highly efficient coding and problem-solving, akin to the meticulous planning depicted in the movie Apollo 13. This analogy serves as a reminder of the importance of precision and efficiency in technology, especially in high-stakes environments.
A call for continued dialogue and action
All panelists underscored the critical need for the IT industry to actively embrace sustainable practices. The journey towards a greener IT sector is complex and challenging, but as the insights from our panelists show, it is both necessary and achievable. We invite our readers to continue this crucial conversation and explore sustainable solutions in their QA and testing practices.