Improving Developer Experience Drives Profitability
Developers are first in DevOps in name only. In reality, the first decade or so of DevOps transformations was largely focused on operations. It did allow the ops teams to get more involved in the development process. And you can probably see over some silos. But a lack of DevOps consideration around developer experience met with exponentially increasing code complexity and staff cuts, meaning developer burnout continues to rise.
If the definition of DevOps is to remove friction between silos, then most DevOps transformations have been slowed to a crawl.
“Friction is abundant, the development lifecycle is riddled with red tape, and successful delivery of code to production is a frustratingly infrequent event,” reads the newest developer experience research DevEx in Action. This paper defines developer experience, or DevEx, as how developers feel about, think about and value their work.
As we’ve previously covered, DevEx metrics consider:
- Flow state – the ability to get in the zone with limited interruptions.
- Feedback loops – the time it takes for a developer’s code to be reviewed and work to be delivered, along with the quality of the feedback.
- Cognitive load – the amount of information devs need to retain to accomplish something.
In these tighter economic times, it can be hard to justify the added expense of a new strategy focused on developers. Except, it’s now proven that a focus on measuring and improving developer experience can enhance productivity and profitability for individual developers, teams and whole organizations. Here’s what you can take back to your DevOps leadership to advocate on behalf of your developers’ experience.
What the Latest DevEx Research Has Found
Led by Microsoft Research’s Nicole Forsgren, co-founder of DORA and co-author ofthe SPACE framework, the team at the DX developer experience platform recently surveyed 219 developers at their customer organizations. Of these respondents, 78% were at tech-native companies — as opposed to legacy orgs like banking, insurance and healthcare that have become pseudo-tech companies — while 91% were at organizations of 500 or more employees.
It’s the latest study to analyze the statistical relationships between DevEx factors and outcomes at the individual developer, team, and organizational levels.
Through 18 questions, the researchers proved that:
- A focus on enabling more flow state positively impacts developer, team and organizational outcomes.
- Tighter feedback loops influence team-based outcomes.
- An effort to reduce cognitive load positively affects developer, team and organizational outcomes.
Overall, the latest DevEx research found that the ability to go deep on interesting work provides the biggest potential impact on developer experience, which in turn drives productivity and innovative problem-solving. It also uncovered some cross-organizational areas for improvement, like meeting-free days and putting in place more efficient approval systems.
Developers that achieve that deep work feel 50% more productive, while those that find their work interesting are 30% more productive, the study found. These are the results that most align with one of the underlying goals of DevOps which is to limit repeat work through automation. It also echoes 2023’s focus on unlocking developer productivity by increasing developers’ focus on and speed to delivering differential value to the end user.
Trying to convince DevOps leadership this is worth the investment? Interruptions have a cost. It’s found that for every break in flow, it takes an average of 15 minutes and three seconds to get back on task. At the hourly rate of the average developer, time is definitely money.
Can Platform Engineering and GenAI Help?
Two more of the hot 2023 trends that align with this research are artificial intelligence — particularly generative AI — and platform engineering.
The paper found that devs who have a higher understanding of the codebase see themselves as 42% more productive, which is a great generative AI use case, like generated, interactive documentation and getting started guides. This is not meant to replace mentorship and pair programming, but as a way to help developers get up to speed on increasingly complex and distributed systems.
The 2023 StackOverflow Developer Survey found that 63% of respondents spend at least half an hour a day looking for answers — this context switching is a huge cause of interrupted flow. This is another opportunity for generative AI, especially when chatbots are trained on internal documentation and processes that are then embedded alongside the code. Switching between a penumbra of tools will hopefully soon be a thing of the past.
The paper also found that sensible work processes and easy-to-use tooling made developers feel 50% more innovative. This harkens to 2023’s other big trend platform engineering, which looks to standardize and automate repetitive, tedious developer work, especially in trying to release to the cloud. An internal developer platform or portal combines developer tools into a single flow, behind a single pane of glass. This creation of guardrails guides the developers on an easy-to-follow pathway then opens up more time and energy to dedicate to creative problem-solving and deep work.
In addition, since technical debt is a huge barrier to faster feedback and often a cause of friction and frustration for developer workflows, the paper, in its citing of work design theory, considers that technical debt and code quality have a strong impact on team outcomes. This can be an argument to take to DevOps leadership that investing in the reduction of technical debt and progressing cloud migration can have a strong impact on productivity and profitability at the team level.
Generative AI is already impacting code quality, which will in turn impact team-based outcomes. GenAI helps developers create more code faster, which in turn increases productivity. However, recent research out of GitClear found a decrease in code quality due to an over-reliance on GitHub Copilot.
More doesn’t equal better, that research found. Generative AI’s response is based on the probability of being accepted, not on its accuracy. It also, so far, does not consider the context of your overall complex code base. This makes it challenging for long-term maintainability and it could indeed create more technical debt. Add to that the valid DevSecOps concern that teams can’t keep up with this speed of code creation, which in turn could create more vulnerabilities that interrupt developers’ deep work. So approach GenAI with caution, but consider it within the context of your individual developers, your teams, and your organization as a whole.
Why Is DevEx So Hot Right Now?
“The focus on developers and just the cost center that comes with their salaries and tooling, developer experience started to surface as a way to be more efficient with those tools and developers in their workflows,” Troy Gray, account executive at DX, told The New Stack. He speaks with 10 to 12 different organizations a day, having conversations with DevOps leadership about how to improve developer experiences which typically divulges into how to measure it. After all engineering is a science, which means you can’t improve what you can’t measure.
Certainly, the economic climate has triggered a rise in interest in DevEx over the last year or two. “The biggest kind of push and trend is that teams are looking to be more efficient,” Gray said. “And the way to do that is to identify what is slowing teams down, what are the bottlenecks that they’re experiencing, and tools and workflows.”
By employing DevEx measurements, he said, orgs are not only able to identify bottlenecks but prioritize which have a greater impact on developer experience.
“Now there’s data behind it. All of a sudden, there’s actually a way to prioritize how big of a bottleneck that is, rather than just ad-hoc conversations or one-on-ones that go from developer to an engineering manager, engineering manager to a director, to a VP,” Gray said. “It’s really left to that communication path to understand: Is this something we should prioritize?”
The developer experience at each organization is as different as each tech stack and team. This is why any developer experience effort should be grounded not just in data but a combination of data and developer surveys. But, when those two sources conflict, which should you trust?
“I see leaders getting stuck with ‘My developers are telling me X, but my metrics are telling me Y’.” Abi Noda recently wrote on LinkedIn. “Your developers are always right.”
This is why the DX platform, Gray explained, always leaves an open form next to the very quantifiable votes, so devs can leave any explanation or further info as they see fit.
In the same post, Noda cited Amazon’s Jeff Bezos, who has been known to have said, “The thing I have noticed is when the anecdotes and the data disagree, the anecdotes are usually right. There’s something wrong with the way you are measuring it.”
When in doubt always trust your humans. Developer experience is supposed to be about them, after all. They will more often than not know their bottlenecks and barriers to deep work. So start a conversation with your devs today.
