Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News Embracing Complexity and Emergence in Organisations

Embracing Complexity and Emergence in Organisations

Focusing on the actual emerging organisation and the work people are doing can make a difference in embracing complexity and dealing with it a bit better. Psychological safety is critical for people giving feedback without fearing retribution or negative consequences.

Fred Hebert spoke about embracing complexity at QCon New York 2023.

A key assumption is that key elements of the organisational structure emerge from decisions made, trade-offs, goal conflicts, and every day interaction across teams, Hebert argued. As groups attempt to meet individual objectives with their own incomplete views of the whole, the organisation shifts in a dynamic and often uncoordinated manner, he explained:

Various unexpected relationships exist within an organisation, and are relied upon to effectively make things work. Attempting to reorganise based on an idealised organisational chart may remove very few of the key pressures driving the existing emerging structure, and they’ll stick around.

Each organisation has its own context and evolves in a slightly different ecosystem with varied dynamics, Hebert said. Almost any behaviour can be successful for a while given the right environment. But most behaviours that are based on a desire for collaboration, communicating better, and with reciprocity in mind seemed to do better in a way that felt more sustainable, especially when being challenged by new circumstances, Hebert added.

Psychological safety is critical when dealing with complexity, Hebert said. The only way you can get into that decision-making process that people have is for them to be able to report what they find challenging, risky, difficult, or even easy. This happens when people can do that without fearing retribution or negative consequences, Hebert mentioned. That feedback making it to you also depends on them trusting that giving that information will lead to positive outcomes and will be acted on, not just that nothing bad will happen, he said.

When something happens you don’t see reflected in your existing metrics, this is a sign that events and interactions might be more complex than you planned, or that your overall system—which is likely dynamic and rapidly changing—shifted around already, Herbert said. Listen to these little feelings of dissonance and dive into them, he concluded.

InfoQ interviewed Fred Hebert about dealing with complexity with psychological safety and feedback loops.

InfoQ: What can we do to increase psychological safety and trust?

Fred Hebert: They’re far easier to lose than to earn. If you’re in a position where they exist, maintaining them should be one of your top priorities whenever you make a decision.

Trust and safety go both ways; if you want people to trust you more, also ask yourself if you trust them as well. How would you delegate decision-making to these folks? Listen to the information you get from them. Ask yourself what type of support trustworthy people deserve when things go bad, and offer it. These are patterns that take a long time to establish and can potentially be baked into the structure of your organisation and its hierarchy as well.

You have to first respect people and their efforts, and understand where they’re coming from. Think really hard about their situation before changing their work – chances are they thought about it as much as you did already, but we all work within different contexts. I’m sorry I don’t have a better recipe.

InfoQ: How do feedback loops help to deal with complexity?

Hebert: A shorter feedback loop means you can better understand the impact actions and changes have on parts of a system. Some actions are necessarily slow burners—like finding out that making it possible to spin up new microservices super fast indirectly discourages collaboration on existing services because that takes more time—but a lot of other useful signals are far more straightforward and easily available; they just need our attention.

One of my favorite ones is people looking at dashboards for a different service than the one they’re trying to debug because the faulty service’s dashboards aren’t trusted. There’s a lot to dig into that behavior!

About the Author

Rate this Article