The DORA research group has published its 2024 report, its tenth year of publication. Based on a global survey of over 39,000 professionals and supplemented by some in-depth interviews, the annual Accelerate State of DevOps report gives a broad and detailed look at the factors influencing team productivity, job satisfaction, and organisational success.
The report summarises the practices and performance of high-performing technology teams, giving insights into how AI, leadership, and user-focused development are shaping the contemporary software industry.
AI Adoption Brings Benefits and Challenges
One of the key findings is the growing impact of artificial intelligence on software development. The report notes that AI is now significantly used in most organisations, and many of them frequently shift their approaches to using AI. Early adopters are seeing some promising results in areas such as developer flow, productivity, and job satisfaction, leading to better overall organisational performance. Another finding was that as developers' trust of AI tools increases, so too does their willingness to use them in their workflows.
"AI introduces risk, but not because of garbage code [...] it’s because batch size seems to increase when AI is used in the coding process. And bigger changesets are riskier, something that DORA’s research has long supported"
- Laura Tacho (DX)
However, the report urges developers to be careful, with some negative trends around AI to report. A significant 39% of respondents said they had low or no trust in AI-generated code and hinted that AI tools may actually be undermining software delivery performance. The data showed a drop in throughput (1.5%) and stability (7.2%) for environments where AI had been adopted.
The team make some cautious and measured practical recommendations for integrating AI - suggesting that organisations look at AI as a tool to reduce administrative burdens rather than as a replacement for human expertise. They also highlight how developers need time and space to assess AI tools properly rather than unquestioningly adopting them if performance isn't to be negatively affected.
User-Centric Approaches Boost Performance
Another key finding is the importance of a user-centric approach to software development. Organisations prioritising the end-user experience seem to produce higher-quality products. Software engineers working like this also seem more productive, satisfied, and less likely to burn out.
Keeping firmly focused on users' needs and expectations has a broad emphasis in the report. Organisations improve across the full range of metrics by keeping the customer in the middle of making decisions.
Transformational Leadership and Stable Priorities Matter
The study found that teams working under considerate and adaptable leadership with a transforming and progressive mindset are doing better, both in terms of delivery and job satisfaction. Channeling this style into a clear and consistent strategic direction is also deemed important. The report examined the "move fast and constantly pivot" approach, which has had much uptake in contemporary agile software engineering and found it often detrimental to developer well-being and overall team performance. According to the report, frequent changes or lack of alignment harm productivity and morale.
The report shows that organisations that cultivate stable and supportive environments can achieve noticeably better outcomes. This involves creating a workplace culture that values developer input, provides clear direction, and minimises unnecessary disruptions to work.
Platform Engineering Offers A Boost But Not For Everyone
The study also examined the impact of platform engineering, the practice of creating self-service internal developer platforms to improve developer experience and overall productivity.
The data show that platform engineering done well improves both productivity and organisational performance and has found its biggest successes in large organisations. Smaller organisations might find these approaches challenging to implement and potentially counterproductive, with poorly implemented platforms being obstacles rather than actually helping productivity and the developer experience. The report also found dips in performance in organisations amid a platform engineering initiative. The researchers suggest that teams should consider these tradeoffs very carefully when choosing whether to invest in platform engineering capabilities.
As a counterpoint, in a post on LinkedIn Mark Panthofer suggests that this part of the DORA report does not tell the whole story; he positions platform engineering as a force multiplier:
"Beyond these key metrics lies a deeper layer of practice and strategy that many elite performers leverage: the platform engineering force multiplier"
Conclusions
In conclusion, the report's core message is that high levels of software delivery performance are achievable - with high-performing teams excelling across all four of the now traditional DORA key software delivery metrics (change lead time, deployment frequency, change fail percentage, and failed deployment recovery time), and organisations failing in any one of these areas are likely to be failing in all four. The report confirms that succeeding in these areas is not just a technical feat - but one where a supportive, innovative and stable organisational culture is needed.
Reflecting on Laura Tacho's write-up of the report, GitLab Field CTO Stephen Walters comments:
"The root cause is often that the tool is implemented in the same culture with the same working practices involving the same people. There is little to no recognition that the change lies beyond the tool"
- Stephen Walters (Field CTO GitLab)
An executive summary, and the full report is available on Google Cloud's website.
