For the last decade DORA has put out an annual State of DevOps report, looking at how different practices impact software delivery performance. The 2023 report reinforced some of the things I’ve been thinking for a long time about code reviews - mainly that faster reviews on smaller code changes have an outsized impact on team performance. But they also found a fascinating relationship between faster code review turnarounds, developer happiness, and burnout.

A primer on DORA metrics

If you’re already familiar with the core DORA metrics jump ahead to the next section, but if you haven’t heard of them before or would like a quick refresher here’s an overview.

DORA looks at four key metrics to measure the likelihood of success and efficiency of software engineering teams:

• Change Lead Time - The time it takes for a code change to go from being committed to being deployed
• Deployment Frequency - How often code changes are pushed to production
• Change Failure Rate - How often a change deployed to production leads to a failure or bug that needs to immediately be addressed.
• Failed Deployment Recovery Time - How long it takes to recover when a deployment fails

There’s no standardised way to measure these metrics, but tracking them on your team over time can help identify areas where you can improve and see where you’re making progress.

Smaller PRs are more efficient

Even if there isn't a universal way to measure the DORA metrics, the 2023 report found that one of the best ways to improve all of your metrics is to make sure you're working with smaller code changes and PRs.

Smaller pull requests:

• Are faster to review and can be merged and deployed faster, reducing Change Lead Time
• Are easier to deploy consistently or continuously, increasing Deployment Frequency
• Are narrowly focused, making bugs easier to catch during reviews and testing, reducing Change Failure Rate
• Are less complex and easier to fix, reducing the Failed Deployment Recovery Time

Faster reviews lead to better performance - and less burnout

Cutting down the size of your PRs can help improve performance and speeding up your code review process overall can be a force multiplier on top.

DORA found that “teams with shorter code review times have 50% better software delivery performance”. Across a number of technical capabilities and processes (AI usage, continuous integration, loosely coupled architecture, and trunk-based development) code review speed was the biggest factor in driving better software delivery performance and better operational performance.

But performance isn’t the only area where DORA saw an impact from faster reviews.

Faster reviews led to a substantial reduction in developer burnout and a moderate increase to job satisfaction.

Improving the review process isn’t just about making development more effective and more efficient - it’s also about creating an environment that is better to work in.

In a world where the average review still takes more than a day, there’s a lot of work to do to improve code review processes - but the benefits are clear.

---

At Sourcery we’re trying to make the code review process faster and more effective by providing automatic reviews for every code change. Try it out for GitHub, or sign up for our waitlist for other platforms.