Human error: an important ingredient in great designs
To err is human; to forgive, divine — Alexander Pope
Some of us designers like building systems. We think things through and we pay attention to details. We make sure things work and we always ;) cater to our users. But when we design, we have this fantastic idea that everybody will behave in the same perfect way when interacting with our system.
You go here, swipe that, press this and voila. You purchased your thing!
Then, we user-test our system and observe how it is actually used. We correct and adjust and test it again only to find out users will still use it based on their perception and not our plan. And deviating from the ideal cases always will come with unpredictability, untreated edge cases, and errors.
It is hard for some of us designers to accept that it is actually impossible to design a system that isn’t susceptible to human error. Everybody makes mistakes. Even our users!
Human error is still a primary cause contributing factor in disasters and accidents in industries as diverse as medicine, nuclear power or space exploration.
Let’s go over the types of errors we make, what causes them, what are the strategies for correcting errors and last but not least how to design for errors.
Disclaimer: the term system is used loosely to describe digital products, flows, actual systems, etc.
How many types of error are there?
As someone who studied human error as close as possible, James Reason, Professor Emeritus at the University of Manchester, divided human errors into two categories: Slips and Mistakes.
Slips — when the intention and the action don’t match
- Action-based slips: You unlock your phone to call mom but because of conditioning, you jump on Instagram instead.
- Memory lapse slips: You forget your keys hanging in the door lock when you leave home.
Mistakes — when the goal and the outcome of the actions are wrongly evaluated
- Rule-based mistakes: The doctor interprets an illness correctly but prescribes the wrong medication.
- Knowledge-based mistakes: You frantically press the enter button when the computer is stuck, wrongly assuming it will unblock the computer.
- Memory-lapse mistakes: You forget to finish the task in progress because you got distracted and carried away by an Instagram notification.
To remember:
Most everyday errors are slips. Slips are the result of automatic, subconscious processes. You plan to take one action and you find yourself taking another.
Mistakes are sometimes referred to as errors of intention or errors of planning and they result from conscious deliberation. You plan to achieve something but consciously take the wrong steps.
When researching for this article, I found this PDF published on UK’s Health and Safety Executive website with more examples and details.
Why do errors happen?
That’s right, you guessed, human errors happen for many reasons. 🙋🏻♂️
Selected from the large pool of contributing factors, here are the most influential ones:
- Decision-making biases
- Working under stress
- Working under un-natural conditions
Let’s briefly zoom in a bit on each of these.
1. Decision-making biases
For the purpose of this section, I selected three of the most common biases that influence human error so you can make an idea of what they entail:
- The focusing effect: which is our tendency to place too much importance on one aspect of an event. Tunnel vision might sound more familiar ;)
- Illusory correlation: when people inaccurately perceive a relationship between two unrelated events.
- Automation bias: the tendency to depend excessively on automated systems which can lead to erroneous automated information overriding correct decisions.
Of course, as humans, we are predisposed to a huge number of biases. There is a huge list on Wikipedia where you can find more.
If I would to give you some advice, at the beginning of your project, you should create a list of biases you assume will influence your users’ decisions and validate or invalidate them during user-testing.
2. Working under stress
The Yerkes-Dodson Law
Back in 1908, Robert Yerkes and John Dodson, two psychologists, postulated the relationship between stress (called arousal in the field of psychology) and performance and it can be summarised as:
Little stress (up to one point) can help you perform a task because it heightens awareness, but when the stress is too high, performance decreases.
Of course, research shows that the optimal amount of stress/arousal depends on the difficulty of the task.
- Low-stress, low-difficulty equals boredom.
- Little-stress, high-difficulty equals better performance thus less human error on the task compared to:
- High-stress, high-difficulty which leads to poorer performance and an increase in human error.
As a best practice when you build systems, you should take into account the stress level of the environment in which our system is used because it might drastically influence the occurrence of human error.
3. Working under un-natural conditions
Knowing the conditions in which people will use our system may drastically influence our designs. The tradeoff between error correction mechanisms and un-natural working conditions should be proportional and should be partially observed from user-testing.
Some scenarios our users might find themselves in when working:
- Staying alert for prolonged periods of time
- Constantly providing accurate, and minute details
- Having to work in environments prone to interferences
- Working with large amounts of diversified data
How do people solve problems?
It might be of no surprise, but several patterns of problem-solving and error correction emerged when studying people executing usability tests.
The important ones you should know about and take into account (when making assumptions about your users) are:
Systematic explorations
This is an optimal pattern of behavior that leads to the highest probability of successful task completion. People try all the available options in a particular order until they found the one they need.
- Example: Your task is to open the “Feedback” app on a phone with a new operating system that represents apps only through icons. You will proceed to open every app from top to bottom until you open the requested one.
Trial and error
This is another viable technique, even though it is considered less effective or less efficient than the systematic exploration technique. People randomly try different options until they find the one they need.
- Example: Your task is to take a screenshot on a new phone. You will randomly try combinations of buttons and gestures in order to complete your task.
Rigid explorations
People assume the way to solve a task. They proceed to repeatedly use the same action or process over and over even though the outcome is never achieved yet they still believe it will work.
- Example: Your task is to play your favorite song on a music app. You will open the app and press the big play icon, restarting and re-doing the same process, even though it never worked to begin with.
Further reading:
The smart people at the University of Amsterdam published this very interesting paper that is definitely worth your time if you are curious to know more.
Designing for human error
We now know how many types of errors there are, why they happen and how people usually correct them. Human error might be inevitable, but it need not be catastrophic.
Let’s see some design techniques you can use when designing for human error.
1. Preventing errors
It is a good design practice to avoid human error before it can even happen. This transforms your regular old system into one that gives your users a sense of security and stability.
Techniques you can use:
- Adding constraints: you will expose a credit card number text field to human error way less often if you allow it to receive only numerical characters.
- Being consistent: having the confirmation button of your modal window on the right side at all times makes it less prone to human error because it builds a specific mental model of your system.
- Using clear metaphors: the trash can icon means delete for most users because it’s a distinct and recognizable metaphor, making the chance of a misunderstanding very small.
2. Forgiving
Having a way to correct your errors is a reassuring thing and allows you to explore, navigate and learn any system without any anxiety or fears of doing something wrong.
Techniques you can use:
- Good affordances: buttons should look like things that can be clicked or tapped.
- Undo capabilities: destructive actions are susceptible to human error and the system should allow them to be reversed in case of a slip or when the intent of the user changes.
- History log: your system should keep an action-log so that your users can identify and correct any potential human error.
3. Explaining
As designers, we know people don’t read screens/pages of text word by word. They scan. That doesn’t mean we shouldn’t use explanatory text at all. No communication is bad communication and when we’re dealing with human error, that can prove itself to be a delimiting factor.
Techniques you can use:
- Confirmation messages: the actions users are going to take (along with its outcomes) should be explained in short, plain and clear language.
- Warnings: your system should clearly indicate a potential misbehaving with the help of signs, alerts, and prompts.
- Help sections: easy-accessible, well structured, dedicated FAQ/help pages or screens, user manuals or usage instruction in the form of plaques/leaflets.
Bonus techniques
- Reducing the cognitive load: by having a simple, easy to parse design or by reducing visual or audio stimuli altogether(this also reduces stress).
- Offering support: by designing dedicated onboardings, offering hints, providing a customer support line, video tutorials, or training in the form of courses.
In conclusion
The reality is that something always goes wrong: a colleague makes a mistake in shared-file, or a company releases an app that has too many bugs, or an engineer builds a system that is unusable because he or she doesn’t understand their user’s needs.
With all this, we should aspire to better understand human nature and how to utilize its strengths and correct its weaknesses.
If you enjoyed this article, I would appreciate if you’d share it with your co-workers and friends. Hold the clap button as much as you like 👏 👏 👏
