Google Cloud published the tenth annual Accelerate State of DevOps (DORA) report in late 2024, drawing on responses from 39,000 software professionals worldwide. The four core DORA metrics remain unchanged: deployment frequency, lead time for changes, change failure rate, and mean time to recovery. What changed in 2024 is the distribution of performance across them.
The elite tier (teams deploying on demand with lead times under one day, change failure rates around 5%, and recovery times under an hour) held at roughly 19% of respondents. The high-performer category, previously the largest tier, shrank from 31% to 22%. The low-performer category, characterised by deployment frequencies of monthly or slower and recovery times measured in days, grew from 17% to 25%. The middle is hollowing out. Elite teams are pulling further ahead. A significant share of organisations are falling further behind.
One of the most statistically unusual findings in the 2024 report was that medium-performing teams showed lower change failure rates than high performers. This is the first time this inversion has appeared in ten years of DORA data and the report's authors note it as an anomaly worth watching rather than a signal to act on immediately.
The 2024 report expanded its outcome framework beyond the four delivery metrics to seven broader organisational accomplishments: software delivery performance, individual productivity, team performance, organisational performance, product performance, job satisfaction, and burnout reduction. This is a meaningful expansion because it acknowledges what practitioners have known for years: delivery throughput is a lagging indicator. The leading indicators are in how engineers experience their work.
The finding on transformational leadership is particularly actionable. The report found that leaders who communicate a clear vision and actively support their teams measurably reduce burnout and improve performance across all seven outcome categories. This is not a soft finding; it held across company size, industry, and geography. The engineering leadership behaviours that correlate with elite performance are specific: providing direction without dictating solutions, removing organisational blockers before they affect delivery, and making it safe to raise architectural concerns before they become incidents.
On platform engineering, the report surfaced a nuance that many internal platform teams should hear: organisations that rolled out internal developer platforms saw temporary dips in delivery performance during the transition. The cause was insufficient investment in the user experience of the platform itself. Self-serve tooling and documentation that enables developers to unblock themselves outperformed platforms that required teams to raise support tickets. The lesson is that an internal platform is a product, not an infrastructure project. It needs a product manager, user research, and a feedback loop just as much as a customer-facing product does.
The SPACE productivity framework, developed by Microsoft Research and GitHub, covering Satisfaction, Performance, Activity, Communication, and Efficiency, provides a measurement model that complements DORA by capturing the dimensions DORA's four metrics do not reach. GitHub used it in their controlled Copilot study. It is the framework most aligned with the 2024 DORA report's expanded outcome model. Engineering organisations that want to understand not just how fast they deliver but why that changes over time should be measuring across SPACE dimensions alongside the four core DORA metrics.
Source: Google Cloud / DORA, "Accelerate State of DevOps 2024," dora.dev (approximately 39,000 respondents). Performance tier data and expanded outcomes sourced from the published report and the InfoQ analysis, November 2024.