Source: ZDNet. Original research published by DevOps Research and Assessment (DORA) and Google Cloud, led by Dr. Nicole Forsgren.
Elite DevOps Teams Are Pulling Far Ahead of Their Peers
Adopting DevOps practices has become a standard ambition across the software industry, but a large-scale research study from DevOps Research and Assessment (DORA) and Google Cloud reveals that simply having a DevOps initiative is not enough. The organizations that are genuinely excelling at DevOps outperform their low-performing counterparts by staggering margins — not incremental ones.
The report, led by Dr. Nicole Forsgren and her research team, draws on six years of data collected from more than 31,000 professionals worldwide. It identifies roughly 20 percent of surveyed enterprises as “elite performers” — organizations where DevOps practices have matured to the point of delivering measurable, tangible business results.
Key Performance Metrics: Elite vs. Low Performers
Deployment Frequency: 208x More Frequent
Elite performers deploy on demand and routinely ship multiple times per day. The study calculated approximately 1,460 deployments per year for the highest performers (roughly four deploys per day × 365 days), compared to around seven deployments per year for low performers. That difference works out to a 208-times higher deployment frequency for elite teams.
Low performers, by contrast, typically managed between one and twelve deployments per month at best — far below the cadence required to respond quickly to market changes or customer feedback.
The researchers note that four deploys per day is actually a conservative benchmark. Companies such as Capital One have publicly reported deploying up to 50 times per day for a single product, while Amazon, Google, and Netflix deploy thousands of times daily when aggregated across the hundreds of services that make up their production environments.
Change Lead Time: 106x Faster
The study measures change lead time as the duration between a developer committing code and that code running successfully in production. Elite performers achieve lead times of less than one day. Low performers, by comparison, reported lead times ranging from one month to six months — a 106-times difference.
This gap has direct business implications. A team that can move from code commit to production in hours can respond to bugs, security vulnerabilities, or competitive opportunities in ways that a team working on a monthly release cycle simply cannot.
Why Larger Organizations Struggle
One of the study’s more striking findings concerns organizational size. For the first time in the research program’s history, the data showed that larger organizations — those with more than 5,000 employees — were more likely to be low performers than their smaller counterparts.
According to the DORA/Google team, the reasons include:
- Heavyweight processes and controls that slow approval cycles and limit team autonomy
- Tightly coupled architectures that make independent deployment of services difficult or impossible
- Organizational inertia that makes cultural change harder to sustain at scale
This does not mean large enterprises cannot achieve elite status — but it does suggest that architectural and organizational changes are prerequisites, not optional improvements.
What Elite DevOps Organizations Actually Do Differently
1. Start with Foundational Practices
The researchers recommend that organizations at any stage of their DevOps journey begin by establishing solid foundations:
- Version control for all production artifacts, not just application code
- Automated testing integrated into the delivery pipeline
- Monitoring and observability across services and infrastructure
- Clear change approval processes that are lightweight enough not to create bottlenecks
- A healthy engineering culture grounded in psychological safety and learning from failure
2. Identify and Attack Constraints Iteratively
The DORA team advocates a constraint-based improvement model applicable whether you are just starting a DevOps transformation or have been optimizing for years. The approach: identify what is currently holding your delivery pipeline back, focus resources on that constraint, resolve it, then find the next bottleneck. This iterative method avoids the trap of investing in improvements that do not address the actual limiting factor in throughput.
3. Empower Teams to Make Their Own Tool Decisions
Elite DevOps organizations grant their teams meaningful autonomy over tooling and implementation choices. The study found that empowered teams consistently choose tools that are both useful and usable — and that those tool choices measurably improve productivity. Top-down mandates that ignore the needs of individual teams tend to produce friction rather than acceleration.
4. Leverage Cloud Capabilities Fully
Cloud adoption alone is not a differentiator — 80 percent of survey respondents reported that their primary application or service was hosted on some form of cloud platform. What separates elite performers is how they use cloud, not whether they use it.
Using the NIST cloud computing framework as a reference, the researchers found — for the second consecutive year — that elite performers were 24 times more likely than low performers to execute on the full essential capabilities of cloud. Simply lifting and shifting workloads to a cloud provider does not replicate the benefits of cloud-native architecture, elastic scaling, and on-demand self-service provisioning.
Practical Takeaways for DevOps Teams
The DORA/Google research offers a clear message: DevOps maturity is not a binary state. There is a measurable, wide spectrum between low performers and elite performers, and the gap compounds over time as elite teams continue to improve while laggards maintain slow, manual processes.
If your organization is looking to move up that spectrum, the evidence points to a few high-leverage actions:
- Audit your current change lead time and deployment frequency honestly — these are the two metrics most directly connected to overall delivery performance.
- Invest in CI/CD pipeline automation before anything else, since manual handoffs are typically the largest source of lead time waste.
- Evaluate your architecture for tight coupling; monolithic systems are a structural barrier to frequent, independent deployments.
- Reduce approval gates that add days without meaningfully reducing risk — lightweight peer review and automated testing typically outperform heavyweight change advisory boards.
- Give teams genuine ownership of their tooling and deployment processes, with accountability for outcomes.
The gap between elite and low-performing DevOps organizations is not primarily a technology gap — it is a practices and culture gap. The good news is that both are learnable and changeable, regardless of where your organization sits today.
The statistics cited in this article are drawn from the DORA/Google Cloud State of DevOps research program, as reported by ZDNet in September 2019. The full report was authored by a team led by Dr. Nicole Forsgren.