Feedback Loops

Feedback loops are processes where designers use a system’s outputs as inputs to find cause-and-effect relationships within it. Some systems (e.g., the environment) have many feedback loops, and the effects of human actions can take decades to show. In complex systems, feedback loops can hide causal links and problems.

“The human mind is not designed to understand the complexities of all these systems.”

— Don Norman, “Grand Old Man of User Experience”

See why it takes careful insight to work with feedback loops.

The Long, Winding Webs of Feedback Loops

Cognitive science and usability engineering expert Don Norman identifies 21st century design as the way to address the many complex problems that people face across the world. But the human brain has problems understanding the sheer complexity of many systems. We’re surrounded by interconnected systems that feed back and forth between one another extensively and often imperceptibly. And we’re used to looking for simple answers to understand why certain events happen, preferring nice, direct case-and-effect chains because they make sense. For example, if you click “Buy it now”, you get your desired item.

However, nothing happens in a vacuum. Our planet itself is a giant, extremely complex system that contains many, many subsystems, and human activities impact these, ourselves and other things often in hard-to-see, indirect ways. The feedback we want to find to address the right problem isn’t always clear, since the environment where the many factors are interacting—for example, to cause climate change—is a system of entangled, “moving” parts that touch, and are touched by, many other factors elsewhere. As designers, we approach what Norman terms complex socio-technical systems, which, like wicked problems, are:

• Difficult to define.

• Complex systems.

• Difficult to know how to approach.

• Difficult to know whether a solution has worked.

  What compounds the challenge is that:

• Many systems have several feedback loops. When you take an action, result A might not appear but instead impact something else you can’t see (result B) and then affect other things. The first result you notice (e.g., result P) might appear only after a threshold is crossed. Meanwhile, the outputs/effects in between these might have impacted other parts of other systems and produce results that only appear much later elsewhere, or eventually return and influence other things in the system you’re concerned with. Or one might appear once another tipping point has been reached and your system shows a new symptom.

  The human brain isn’t designed to follow such non-linear cause-and-effect relationships. A system can have multiple, circular inputs and outputs, and we can’t always appreciate which factors affect which other factors or how. Many remain invisible.

• There could be a long delay between the feedback loop’s inputs and outputs. Many years might pass before the effects (or outputs) of a feedback loop manifest, whether the input (cause) is large or small. For example, one new factory causes smog that year, but combined with other polluting agents in the region, its emissions will raise temperatures via the greenhouse effect to affect rainfall, farming, marine life, etc., 20 years later. Meanwhile, the issue also might well have become politicized as powerful lobby groups constantly block remedial measures.

© Michelle Leman, CC0   

How to Work with Feedback Loops

Use humanity-centered design to get the best vantage point to understand the outputs and inputs of a system’s feedback loop. Specifically, leverage the approach’s four principles and:

• Use people-centered design to understand the world through your target population’s eyes, what they understand the cause-effect relationships are, etc.

• Solve the right problem, after deep analysis and (e.g.) using the 5 Whys approach to work your way back from as many effects to causes as possible. Note: sometimes the feeding back will actually be feeding forward, since systems work in complex ways. So, if you isolate one apparent “root cause”, you may uncover another series of causes and effects behind it.

• See everything as a system, and use systems thinking. “Societal” and “technical” are terms to always consider. Remember that effects can be far-reaching in the most unpredictable directions.

• Take small and simple steps towards sustainable solutions. Specifically, use incrementalism:

  • Big problems demand big solutions; big solutions are too expensive, disruptive and failure-prone, though. Be pragmatic; “go small".

  • Once you understand the people you want to help, their situation’s realities and what their environment lets them do, wait for an opportunity to do something small but positive. If it works well, you can repeat/duplicate or improve it. If it fails, it’s still a positive experience as you’ll have learned something

  • Small steps will also be more likely to win the community’s support.

  • Success breeds success. If a small step leads to more victories, you’ll win even more community support.

  • Small steps taken at the right time can lead to the “best solution possible” at any future point—in contrast to a “big fix” taking (e.g.) 10 years, when the whole situation, including the nature of the problem will have changed.

Remember, everything depends on something else happening somewhere else.

© Daniel Skrok and Interaction Design Foundation, CC-BY-SA 3.0

Learn More about Feedback Loops

Take our 21st Century Design course.

Read insights about feedback loops in this thought-provoking UX Collective piece.

Here’s a fascinating and compelling UX Planet piece that involves feedback loops.