The gap between developers’ ideal and actual workdays
How developers allocate their time and where AI can help.
Welcome to the latest issue of Engineering Enablement, a weekly newsletter sharing research and perspectives on developer productivity.
This week I read Time Warp: The Gap Between Developers’ Ideal vs Actual Workweeks in an AI-Driven Era, a new paper from Microsoft researchers that explores the differences between how developers would like to allocate their time versus how they actually spend it. The study also examines how AI tools impact developer productivity and which tasks developers most want to automate.
My summary of the paper
For this study, the researchers surveyed 484 software developers to understand the differences between their ideal and actual workweeks. They identified 16 key activities that comprise a developer's workweek and measured both the time developers currently spend on each activity and how much time they would ideally allocate to them. Then they analyzed how deviations from the ideal workweek affect productivity and satisfaction, and examined which tasks developers most want to automate with AI.
The gap between ideal and actual workweeks
The study found significant differences between how developers actually spend their time versus how they'd like to. In their actual workweeks, developers spend the most time on:
Communication & meetings (12%)
Coding (11%)
Debugging (9%)
Architecting and designing systems (6%)
Code reviews (5%)
In their ideal workweeks, developers would prefer to allocate their time primarily to:
Coding (20%)
Architecting & designing systems (15%)
The largest gaps were found in communication (this includes meetings), addressing customer support tickets, and security & compliance, where developers want to spend less time. Conversely, developers want to spend more time on coding, architecting & designing new systems, and learning new technologies, while maintaining a more balanced distribution across the remaining activities.
Impact on productivity and satisfaction
The analysis showed that as the difference between ideal and actual workweeks increases, both productivity and satisfaction decrease.
The researchers measured this using the Spearman rank correlation coefficient and mean absolute error to quantify the difference between ideal and actual time allocation for each of the 16 identified key activities. They found that developers who felt very productive and satisfied had the highest correlation between their actual and ideal workweeks. And as the gap between actual and ideal workweeks widened, productivity and satisfaction declined.
They also identified specific activities that had statistically significant impacts on productivity and satisfaction:
Positive impact on productivity and satisfaction: Coding, documentation, refactoring, and learning new technologies
Negative impact: Development environment setup/maintenance and communication
AI tools and automation preferences
Developers who use AI tools daily reported the highest levels of productivity (83.7% reported being productive) and satisfaction (74.5% reported being satisfied). These percentages decreased as AI tool usage frequency declined.
When asked which tasks they'd most like to automate, developers' top responses were:
Documentation (creating, updating, and maintaining documentation)
Environment setup/maintenance (configuring development environments)
Testing (authoring, execution, and monitoring of tests)
Task tracking and backlog management
Security and compliance
One developer captured the documentation challenge: "I would love to see automated documentation of infrastructure and code. A lot of time is lost trying to understand the current resources that go into securing, building, and deploying our products and trying to understand how the different parts of our code base interact and depend on each other."
Final thoughts
This study may help leaders understand what kinds of work make developers more productive and satisfied, and what gets in the way. It also raises an important point about AI tools. Many tools today focus on helping with coding (like GitHub Copilot), but this research shows developers may benefit even more from tools that handle the non-coding tasks that slow them down—things like documentation, setting up environments, and dealing with security requirements.
The gap between ideal and actual workweeks represents an opportunity—both for increasing developer productivity and satisfaction, and for guiding the development of future AI tools that target activities developers find less satisfying but necessary for their work.
Who’s hiring right now
This week’s featured job openings. See more open roles here.
Adyen is hiring a Team Lead - Platform | Amsterdam
Scribd is hiring a Senior Manager - Developer Tooling | Remote (US, Canada)
Rippling is hiring a Director of Product Management - Platform | San Francisco
UKG is hiring a Director and Sr Director of Technical Program Management | Multiple locations
Snowflake is hiring a Director of Engineering - Test Framework | Bellevue and Menlo Park
Lyft is hiring an Engineering Manager - DevEx | Toronto
That’s it for this week. If you know someone who would enjoy this issue, share it with them:
-Abi