Key Benchmarks and Considerations for High-Performing Developer Teams
August 24, 2022

Michael Stahnke
CircleCI

In software delivery, there is no question that speed is important. When software teams move fast, good things happen and business value is delivered more frequently.

But, speed comes with a tradeoff: complexity.

As this complexity grows, how can engineering teams succeed?

After analyzing millions of workflows from more than 50,000 organizations across the world, I've outlined some ways teams can start optimizing their software delivery for high performance.

Identify and Meet These 4 Benchmarks

To help teams optimize their software operations for efficiency, CircleCI's latest State of Software Delivery Report examined more than two years of data from over a quarter of a billion workflows, representing more than 50,000 organizations, to gain insights into the DevOps practices used by software teams globally. As a result, the research identified four key benchmarks that the most successful engineering teams routinely meet:

Throughput: Prioritize being in a state of deploy-readiness state most or all of the time, rather than the number of workflows run.

Duration: Reach workflow durations between five to ten minutes on average.

Mean Time to Recovery: Recover from any failed runs by fixing or reverting in under an hour.

Success Rate: Achieve success rates above 90% for the default branch of an application.

Every software team is different. However, the software delivery patterns observed on our platform, especially the data points from top delivery teams, show key similarities that suggest valuable benchmarks for teams to use as goals.

Now let's break down what these four benchmarks really mean.

The number of workflow runs matters less than being in a deploy-ready state most, if not all, of the time. Rather than the number of workflow runs, the most successful teams prioritize being deploy-ready.

The second item that teams should focus on is Duration, which is the time it takes for a workflow to run. Most successful teams achieve workflow durations of five to ten minutes on average.

Third, Mean Time to Recovery describes what it takes for a workflow to become successful again after a failure has occurred. The data shows teams that recover from failed runs in under an hour are the most resilient.

And finally, Success Rate, which is the number of successful runs divided by the total number of runs over a period of time. The most successful engineering teams achieve success rates above 90%.

Prioritize Team Structure and Culture

Prioritizing team structure and culture is essential to improving software delivery metrics. While the ideal team structure and culture will vary depending on the organizational goals, keeping developers in flow is essential to keeping them as productive as possible. That means scheduling meetings at times that don't conflict with peak productivity hours, which the data shows is between 6 a.m. and 7 a.m. PR on Wednesdays.

It is equally important to determine the number of people on your team. Three out of four of our key metrics show a correlation between larger team size and better engineering performance. The research shows the ideal number of code contributors to aim for is between five and twenty, depending on your team's goals, the scope of your responsibilities, as well as other variables. A larger team is also the best way to avoid burnout, and during a time when developer talent is coveted is especially important to consider.

Test, Test, Test

Regardless of your team size, teams prioritizing test-driven development (TDD) can confidently rely on their tooling during market swings, seasonal fluctuations, and times of uncertainty — such as the pandemic. TDD helps companies ensure bad code gets resolved and that organizations can remain safe and resilient.

TDD includes extensive testing, quality checks, and systems that prevent bad code from being put into production. For example, if bad code gets written into your pipeline, TDD can act as a fail-safe when headcount is low. It's the key to preventing bad code from being put into production and staying competitive, regardless of team size.

Great software delivery is a constant loop, not a linear process. The goal for developer teams isn't to make updates to your application, but to constantly innovate on your software while preventing the introduction of faulty changes. Great developer teams that meet the benchmarks outlined above are key to helping businesses differentiate from their competitors and deliver digital products to consumers as fast as the market demands and as often as it changes.

Michael Stahnke is VP of Platform at CircleCI
Share this

Industry News

September 29, 2022

CloudBees announced the acquisition of ReleaseIQ to expand the company’s DevSecOps capabilities, empowering customers with a low-code, end-to-end release orchestration and visibility solution.

September 29, 2022

SmartBear continues expanding its commitment to the Atlassian Marketplace, adding Bugsnag for Jira and SwaggerHub Integration for Confluence.

Bugsnag developers monitoring application stability and documenting in Jira no longer need to interrupt their workflow to access the app. Developers working in SwaggerHub can use the macro to push API definitions and changes directly to other teams and business stakeholders that work within Confluence. By increasing the presence of SmartBear tools on the Atlassian Marketplace, the company continues meeting developers where they are.

September 29, 2022

Ox Security exited stealth today with $34M in funding led by Evolution Equity Partners, Team8, and M12, Microsoft's venture fund, with participation from Rain Capital.

September 29, 2022

cnvrg.io announced that the new Intel Developer Cloud is now available via the cnvrg.io Metacloud platform, providing a fully integrated software and hardware solution.

September 28, 2022

Kong introduced a number of new performance, security and extensibility features across its entire product portfolio, including major new releases of Kong Gateway, Kong Konnect, Kong Mesh, Kong Insomnia and Kong Ingress Controller, as well as new projects from the Kong Incubator.

September 28, 2022

BroadPeak Partners announced the availability of the new K3 API Connector.

September 28, 2022

Aqua Security announced a new end-to-end software supply chain security solution.

September 27, 2022

DevOps Institute will host SKILup Festival in Singapore on November 15, 2022.

September 27, 2022

Delinea announced the latest release of DevOps Secrets Vault, its high-speed vault for DevOps and DevSecOps teams.

September 27, 2022

The Apptainer community announced version 1.1.0 of the popular container system for secure, high-performance computing (HPC). Improvements in the new version provide a smaller attack surface for production deployments while offering features that improve and simplify the user experience.

September 26, 2022

Secure Code Warrior unveiled Coding Labs, a new mechanism that allows developers to more easily move from learning to applying secure coding knowledge, leading to fewer vulnerabilities in code.

September 26, 2022

ActiveState announced the availability of the ActiveState Artifact Repository.

September 26, 2022

Split Software announced the availability of its Feature Data Platform in the Microsoft Azure Marketplace.

September 22, 2022

Katalon announced the launch of the Katalon Platform, a modern and comprehensive software quality management platform that enables teams of any size to easily and efficiently test, launch, and optimize apps, products, and software.

September 22, 2022

StackHawk announced its Deeper API Security Test Coverage release.