What Increases Cognitive Load?
The causes of cognitive load in a large-scale engineering organization.
This is the latest issue of my newsletter. Each week I cover the latest research and perspectives on developer productivity.
This week I read Cognitive Load Drivers in Large Scale Software Development by researchers at Lund University in Sweden. This study explored what causes cognitive load in an international corporation. While every organization is different, the cognitive load drivers identified in the study provide a helpful starting point for leaders who aim to make it easier for developers to ship software.
My summary of the paper
Software engineers manage a lot of information daily. While some of this is unavoidable and part of the job, certain processes and systems unnecessarily increase cognitive load. By improving these problems, organizations can reduce wasted time and improve their team's capacity to deliver. This study sought to identify the categories of cognitive load drivers.
For this paper, researchers conducted a four-stage case study. It started with a literature review, followed with two stages of interviews to identify factors that have an impact on the cognitive load of developers, and concluded with a thematic analysis of the findings. The case company was an international corporation with 1,000 employees, where developers work in cross-functional teams.
The study surfaced more than 15 cognitive load drivers. They’re summarized here and organized into three buckets: tools, information, and work process.
Tool-related drivers of cognitive load
The researchers identified three types of cognitive load associated with tools: intrinsic, delay related, and interaction.
Intrinsic aspects refer to problems with how the tools have been developed. These include lack of functionality, stability, integration, and ease of use. In other words, when tools lack necessary features, malfunction, or don't work well with other tools, it makes it harder for developers to work efficiently and increases their cognitive load.
Delay refers to when a tool or system doesn't respond quickly. This can cause users to lose concentration by having to wait. It can also include system downtime when the tool isn’t working at all.
Interaction refers to anything that makes a tool harder to use. That could include tools that are unintuitive, requiring unnecessary work, or inconsistent with how other systems are used.
Information-related drivers of cognitive load
The second main cause of cognitive load is the quality and availability of the information developers need. This includes several issues: incomplete or missing information, unreliable or outdated information, and a lack of traceability for tracking information over time.
Other problems related to the organization of information include:
Location: Difficulty finding information.
Retrieval: Difficulty retrieving information.
Distribution: Not knowing where to store information or who should receive it.
Structure: Poorly organized information can cause problems. For example, if test results are not saved for each test run, it becomes hard to find the correct information later.
Work process-related drivers of cognitive load
Work process issues add to cognitive load when there is a lack of support, which leads to wasted effort. This happens in two main ways: doing unnecessary work, and spending time figuring out how to use a specific tool or work within a certain team.
The primary drivers within this category are: lack of automation, ad hoc implementations related to both tools and processes, and not knowing how the organization or team works in certain situations. All of the aforementioned factors increase cognitive load.
Final thoughts
Though this study focused on pinpointing the causes of cognitive load in a single organization, leaders might find it a useful starting point for identifying and reducing friction within their own teams. As Manuel Pais recently put it: “Team cognitive load should help us reason about healthy and unhealthy types of workload in the context of a given team's mission. We know for a fact that this view has helped many teams and orgs improve their capacity to deliver on their mission and become healthier (less burnout, less needless stress) at the same time.”
I also asked Manuel if he had a comment about the findings from this paper. Here’s what he shared: “Not only do these cognitive load drivers apply to large software development, but they seem to apply to any knowledge or information-intensive team labor. Think of healthcare ICUs or complicated legal cases, for example. Maybe the tooling impact is lower but the ability to quickly find the right information or collaborate in real-time with the right people becomes critical. In our own research we found four clusters of cognitive load drivers: team characteristics, task characteristics, work environment & tools, and work processes & practices.”
Who’s hiring right now
See who’s hiring Developer Productivity roles right now. Find more open roles here.
AWS is hiring a Senior Data Scientist - Builder Insights | Seattle, New York
Citi is hiring a Director - Engineering Excellence | Texas
Costco is hiring a Platform Engineer | Seattle
GEICO is hiring a Staff Engineer - Developer Productivity | Maryland
Globalization Partners is hiring a Principal DevOps Engineer | Ireland
KnowBe4 is hiring a Software Engineer - Developer Experience | Florida
Sentry is hiring an Engineering Manager - DevEx | California
That’s it for this week. If you know someone who might like this issue, consider sharing it with them:
-Abi