Let’s Stop Talking About Human Error: It’s Your Product


What if I said that there was no such thing as human error, only poorly designed products?

Ok, park your pitchfork. Let’s approach this slowly, starting with Emma.

It’s pay day in the small firm and hope is running high. Emma, who otherwise enjoys kayaking and reading sci-fi, manages the one-person finance office. She oversees collections, salaries and leave. On this special day she walks into the open-plan office area, but something is wrong. Those aren’t payslips in her hand, and that’s not a smile. The area goes quiet.

A few months earlier the firm moved from one payroll system to another. The sales guy for the new application assured Emma that the transfer of established profiles, calculations and tables would be seamless. “It will almost configure itself,” he had grinned. Emma had years of experience with similar platforms and was confident. Pay day arrived and the platform became self-aware (like something in one of her books) and started making big, generous decisions. So now Emma is going to spend half the day walking around the office handing out notifications of debt stapled to payslips, and the other half crying in her office. Not a good day. And the general sentiment in the office? “Emma screwed up, and now it’s our problem. And she’s got years of experience. It can’t be complicated!”

 When Things go Wrong

We’ve grown accustomed to blaming people when things go wrong. Some fool used permanent markers on the whiteboard! Somebody poured water in where the coffee beans are meant to go! Some crazy person wiped the profiles off the digital mixing desk and now the drummer sounds like he’s playing in the room next door – on crockery!

We seldom consider a poor outcome to be the result of a failed interaction between a thing and a person, where the responsibility for the outcome is actually shared.

Products and interfaces are designed for unpredictable humans. We interact with these products, and as we interact we make decisions based upon the actions which are afforded to us and the feedback which the product provides. All this, as we press ever forward towards our important goal, whatever that may be. Our decisions and progress as we interact are also influenced by our knowledge and experience. Every human-product interaction is exactly that: an interaction involving two-way communication, and a path of discovery for the human. Your product is either the inhibitor or the enabler of my success.

But how can anyone suggest that there’s no such thing as human error?

I can see you’re still eyeing the pitchfork, so let’s quickly address this. Humans make errors, yes. They look in the wrong places for what they think they need, or get distracted by things which send them down the wrong rabbit hole, or try to twist the thingie which they should be pressing. The product team receives isolated user feedback and is often quick to jab back with technical categorisations like PEBCAK and ID-10-T. All this against chants of “I can’t believe they did that?! I would never have done that!” You know who you are.

Your User Universe

The trouble is, you shouldn’t be reacting to cases of individual user feedback unless they represent your user universe. Feedback from isolated (often outspoken) individuals is anecdotal at best. You should, however, be paying attention to meaningful test samples of typical users. Aggregated user feedback. This group, representing your users, will either succeed or fail to reach their goals, on aggregate.

And the spoiler alert? If you’ve tested meaningfully and your typical users are failing, it’s not human error.

Let’s stop putting the blame in the wrong place. Interfaces should be designed to help humans achieve their goals, but also to avoid doing stupid things. Alan Cooper refers to “error-inducing design”, which is design that basically encourages error because it fails to consider human limitations.

Don Norman, in his book “The Design of Everyday Things”, suggests that we should stop talking about human error and start talking about poor communication. It should rather be thought of as a collaboration, like between two humans, where one person’s poor communication would not typically be labelled an error. We expect humans to adapt abnormally to the peculiar demands of inanimate objects, including giving precise instructions. Humans are bad at this, but when they fail to meet inhuman requirements then it’s called error, or even stupidity. Just ask Emma.

Consider a radical resolution: Let’s stop talking about human error and start talking about poor communication. Will you join me?



Antony Adelaar


Johannesburg-based product guy and maker, currently in the customer experience (CX) software world. More thoughts and observations on my blog, 'Bye, Product', on Medium. Come say hi on Twitter (@antadelaar) where I'm also likely to tweet about tech, startups, running, parenthood, songwriting and coffee. Especially the coffee.

  • Ross Pope

    Love this. A really nice start to raising awareness throughout product teams

    • Antony

      Thanks for the comment, Ross. All the best.

  • GonzoI

    I get what you’re trying to accomplish, but this is basics of responsible software design, not a “forget what you know” revelation. What you’re talking about is coding to minimize human error, not a lack of human error. I can make number fields only take numbers in a reasonable range and I can give text fields masks and error checking to make sure that what’s entered makes sense, but that can’t stop human error from entering the name of someone they were just talking to instead of the name of a new hire they were supposed to be entering.

  • Product Mgmt Blog

    Excellent article. One way I like to avoid as many mistakes as possible is by utilizing this unique method: http://spectechular.walkme.com/whats-your-ux-combo/