The ballots have been cast, and the 2023/2024 version of Google's popular report on the state of DevOps within the software industry is available for teams to digest.
💡 We have released a concise summary of the entire 2023 Google Devops Report here if you want to begin w/ Cliff Notes of this article's subject
It is full of pictures, and can help to understand the full scope of the DORA report in five minutes or less.
After almost 10 years since the initial DORA report was published, you're probably familiar with its key measurements:
2023 software efficiency metrics as published by Google
In this article, we'll assume that you're broadly familiar with the idea of Google DORA Devops stats. We will dive deeper into two particular questions: 1) what to expect when getting your own stats, and 2) what specific areas of DORA did Google emphasize in their most recently published research?
linkThree practical takeaways for new DORA adopters
While GitClear offers around 100 different stats and measurements for teams, we consider ourselves relatively agnostic about the particular metrics that teams choose to memorialize as their KPIs. Talking to customers who rely on DORA to make decisions, these are the three empirical findings that we have heard (and seen) most often.
link1. DORA stats are noisy at a "monthly" interval (or less)
While there's little doubt that teams need to avoid change failures and long waits after bugfixing, the reality is that for many teams of 100 developers or less, these metrics are unlikely to collect more than 2-3 data points per month. The proof is baked into the the stats Google offers: Change Failure Rate was 10-15% for "High" and "Medium"-performing teams in 2023. Deploys happened, on average, weekly for teams of this size. Do the math:
Note that three of the four DORA metrics are based on frequency of deploy, so for the median software development team, which keeps a weekly release cadence, it takes several months to generate enough release data to begin looking for patterns. This doesn't mean the measurements aren't useful -- in the long-term, these devops stats do convey clear & usable signal on release hygiene. But for cases where a team/manager aspires to embrace "data-driven engineering," DORA needs to be combined with a couple other metrics to flesh out a full picture of performance.
link2. Focus on minimizing iteration interval
Like the popular Agile software methodology, the crux of Google DORA lies in reducing the time to iterate. In the case of DORA, the particular "unit of iteration" is "deploys." Google is correct to assert that increasing the rate of release correlates with numerous positive outcomes.
The "bad news" for most teams is that there is no single change that can shift a team from a long release cycle to a short release cycle. Operating at consistent high velocity requires minimizing tech debt and a commitment to principles that empower a dev team to do their best work, such as:
1. Clarity of next task | 2. Unencumbered time | 3. Aligned incentives |
Enterprise-sized teams are typically looking at a time frame more like "years" than "months" when it comes to accelerating their deployment velocity. That doesn't mean it's not worth pursuing, just that it requires persistent tenacity and great management to foster an environment where Google DORA stats climb to the "Elite" tier.
link3. Google DORA stats are (or should be) free to collect
GitClear, along with a handful of other developer analytics tools, all offer some level of Google DORA stats for free. As of Q3 2024, GitClear's free DORA plan is the only one that scales to commercial teams, and includes code review with the report.
You can read more about GitClear's free Google DORA reporting here.
linkGoogle's 2023 DORA points of emphasis
The 2023 State of DevOps Report, released by Google Cloud's DevOps Research and Assessment (DORA) team, provides an in-depth look into the current trends and best practices in the DevOps field. Here are the key highlights:
Focus on Documentation: High-quality documentation has emerged as a crucial factor in boosting overall technical performance and productivity. It not only supports effective software development but also enhances job satisfaction and reduces burnout.
User-Centric Development: Emphasizing user feedback during the development process significantly improves performance metrics, including continuous integration/continuous delivery (CI/CD) and overall reliability. This approach fosters a more satisfying and productive work environment.
Impact of Cloud Infrastructure: The flexibility offered by cloud infrastructure is essential for improving performance. However, the type of cloud (public vs. private) affects outcomes differently. Private clouds tend to enhance operational and team performance, while public clouds boost organizational and team performance but might introduce challenges in software delivery.
Burnout and Inclusion: The report highlights the higher risk of burnout among underrepresented groups in tech. Addressing inclusivity and ensuring supportive environments are critical for maintaining a healthy and productive workforce.
Technical Capabilities: Core practices such as CI/CD, trunk-based development, and rapid code reviews are more impactful on team performance than newer technologies like AI. These foundational capabilities are crucial for sustaining high performance.
Generative Culture: Building a generative culture, characterized by a high level of trust and collaboration, is linked to better innovation and job satisfaction. This cultural approach leads to more resilient and effective teams.
The report underscores the importance of these practices in driving successful software delivery and operational performance, offering valuable insights for organizations aiming to optimize their DevOps strategies (DORA | Get Better at Getting Better) (DORA | Get Better at Getting Better) (Tom Geraghty) (Semaphore) (Digital.ai).
For more detailed information, you can access the full report here.