Highlights from the 2024 DORA State of DevOps Report

Key findings and takeaways from this year’s State of DevOps report.

I look forward to DORA’s report every year. There’s always something unexpected in the data, and this year’s no different. If you’re pressed for time, here are the top takeaways:

  • AI boosts individual productivity, but hurts software delivery
  • Throughput and quality metrics are moving independently
  • Overall software delivery performance is slightly down from last year
  • Invest in systems that enable developer independence (e.g., documentation, self-serve platforms)
  • Developer platforms may slow delivery but boost individual performance
  • AI adds risk—but not due to poor code quality

AI’s mixed impact on delivery and productivity

For the second year in a row, DORA’s research shows that AI tooling correlates with worsened software delivery performance. I was particularly curious about this, as last year’s report also showed findings that bucked broader industry trends.

The reason isn’t as straightforward as “AI code is garbage.” While 39.2% of respondents indicated distrust in AI-generated code, the real issue is that batch size tends to increase when AI is used in coding. DORA’s data has consistently shown that larger changesets introduce risk. AI simply makes it easier to write more code, and the more code in a single batch, the higher the risk.

Tradeoffs and operational stability

Interestingly, the rise in AI adoption has led to a surprising increase in operational stability, as many teams prioritize AI projects despite competing demands. This stabilization is a welcome side effect for organizations often working under shifting priorities.

However, the AI story is full of tradeoffs. One counterintuitive finding is that AI tooling reduces time spent on meaningful work, but doesn’t decrease time spent on toilsome tasks like meetings, busywork, and admin work. This makes sense though: the most common use case for AI is assisting with coding tasks. I’ve never met a developer who wanted to spend less time coding. We get time back because our meaningful work gets completed faster – not because we can get the robots to do the unsavory parts of our jobs.

Documentation is one of the biggest growth areas for AI in terms of potential impact. DORA’s research has consistently shown a link between documentation quality and performance. AI could accelerate improvements here, though it’s still unclear if it will generate better documentation or simply make existing documentation easier to navigate. DORA estimates that if AI adoption increases by just 25%, we could see a 7.5% boost in documentation quality—a significant impact.

If AI adoption increases in engineering

Software delivery performance and quality metrics

The report also dives into the four key metrics of software delivery performance—deployment frequency, lead time to change, change failure rate, and time to recover. An interesting shift this year: for the first time, the medium performance cluster has a lower change failure rate than the high performance cluster. This is unusual, as the four key metrics have typically moved in tandem.

This year, DORA had to make a choice: either rank high performers based on more frequent deployments with higher failure rates, or prioritize lower failure rates with fewer deployments. Ultimately, they chose the former. Even more interesting, both clusters report failure recovery times of less than a day, a notable shift.

There’s also been a significant shift in cluster distribution since 2023. The elite cluster has remained stable, but the high performance cluster has shrunk from 31% of respondents last year to just 22% this year. The low cluster, meanwhile, has grown from 17% to 25%. This isn’t due to higher standards; in fact, the low cluster performed worse in both deployment frequency and lead time to change, while showing slight improvements in failure recovery.

Platform engineering’s role in productivity and control

Aside from AI and software delivery, this year’s report took a closer look at platform engineering, developer experience, and transformational leadership. The data shows that internal platforms improve both individual productivity and team performance, though they can sometimes slow overall throughput. Organizations using platforms typically deliver software faster and perform better operationally, despite potential tradeoffs.

This aligns with an interesting insight from the DORA community. As James Brookbank noted, “we rarely see companies doing platform engineering primarily for productivity.” So, while platforms may improve governance and security, they also generally boost productivity—though with tradeoffs, of course.

Prioritizing user-centric design in platform engineering

A key factor in successful platform engineering is user-centricity: understanding what internal developers need and designing tools that meet those needs. This approach, often called a “platform-as-product” mindset, emphasizes making developers self-sufficient. Developers should be able to get their work done without relying on enabling teams.

User-centricity was also a recurring theme in discussions about developer experience. This year, DORA incorporated qualitative interviews alongside quantitative data, allowing respondents to share how user-centered practices impact their work and the value they derive from it.

The impact of transformational leadership

Finally, the report highlights transformational leadership as a key driver of high performance. Leaders with a clear vision who actively support their teams reduce burnout, boost job satisfaction, and improve performance across the board—at the team, product, and organizational levels. These may seem like basic traits, but they have a measurable impact on outcomes.

Applying the data

The DORA report isn’t just a benchmarking exercise—it’s a guide for continuous improvement. Performance clusters like elite, high, mid, and low aren’t static categories. They shift each year based on respondent data, underscoring that these benchmarks are dynamic and should be revisited regularly.

I see high performance as a horizon to chase, not a finish line to cross. DORA’s research is built to help you get better at getting better. These insights are a roadmap for ongoing improvement, not a one-time assessment.

For a deeper dive into the DORA metrics and other developer productivity frameworks, check out my guide here.


Published
October 29, 2024

Get started

Want to explore more?

See the DX platform in action.

Get a demo