Writing UX Copy Like a Human
In my personal experience as a product designer, UX writing is one of those things I’ve always been expected to do, but I’ve never had any training for it. That means I’ve written plenty of really bad copy in my time (sorry, past employers).
So far, the two ways I’ve been able to see meaningful progress in this area are:
- Working with people who are better writers and receiving critique from them.
- Reading books and articles on the subject such as this InVision blog article.
After years of effort, there are a couple of things that work for me when I am analyzing UX writing and I want to share them with you.
Trick #1: The Human Test
The name of this is inspired by a John Mulaney standup bit called, “The Robot Test” (please watch it on YouTube if you haven't seen it, it’s gold).
In the bit, Mulaney makes fun of the experience of dealing with CAPTCHAs (by the way, it stands for, “Completely Automated Public Turing test to tell Computers and Humans Apart” 🤯 ). He personifies what the interface is telling the person on the other end — it goes like this:
Joke Setup:
“The world is run by robots, and we spend most of our day telling them that we’re not a robot… just to look at our own stuff!”
Human:
“May I see my stuff please…”
Computer (Robot):
“[Grunts] I smell a robot! Prove, prove, prove… prove to me you’re not a robot!”
Look at these curvy letters! Much curvier than most letters, wouldn’t you say?
No robot could ever read these.
You look mortal if ye be… You, look! And then you type what you think you see!
Is it an ‘E’? Or is it a 3? That’s up to ye.
The passwords have passed, you’ve correctly guessed, but now it’s time for the Robot Test!
I’ve devised a question no robot could ever answer… Which of these pictures does not have a stop sign in it?”
Now of course this example is an embellishment of the idea here, but the idea of The Human Test is to take the message your trying to write in the context of an interface, and say it the same way you would say it to a person.
Let’s see this idea in action.
This is a pretty standard app experience. The designer chose a dialog for the message because it is required to continue. The context behind it (what the user wants) is present to keep them oriented and motivated to accept the terms…
Now let’s imagine telling our friend this same information:
“Hey Jack, How’s it going? I know this is annoying, but I need you to re-accept this privacy policy thing before you move on, it’s just some corporate BS we all have to do. Thanks, man.”
Now let’s compare that sentence to the one in the interface:
“Hey, Jack. Privacy policy and terms of use. To use this app please accept privacy policy and terms of use. Read full privacy policy and terms of use.”
Now I want you to literally read that out loud and imagine having to say this to someone you know in person.
Feels pretty weird, right?
We know the writer of this copy did what most designers do. They took the conventional pattern of a dialog and filled in the blanks as 3 separate parts rather than designing a single, fluid experience.
They didn’t bother to think about these words outside of the context of the interface. The interface is part of the equation here, but starting with the human version and then converting it to the interface leads to a better solution. Let’s try it:
Full-Human: “Hey Jack, How’s it going? I know this is annoying, but I need you to re-accept this privacy policy thing before you move on, it’s just some corporate BS we all have to do. Thanks, man.”
VS:
Full-Robot: “Hey, Jack. Privacy policy and terms of use. To use this app please accept privacy policy and terms of use. Read full privacy policy and terms of use.”
Middle Ground:
“Hey Jack, we’re sorry to ask but we need you to accept our updated Terms of use. Here is the new agreement.” [Accept & Continue]
As you can see, the sweet spot lies somewhere in the middle between robot and human, because that’s exactly what a user interface is — the thing between the human and the robot.
Notice my willingness to ignore the traditional dialog structure. I believe that form follows function, so I’m not going to add a title just because most dialogs have them. I only care if the content requires it. I’m attempting to avoid redundancy by making the link to the agreement part of the message itself. I would make, “Here is the new agreement” a clickable link. Now notice the CTA. “I accept” is not human. “Accept & Continue” is more outcome-focused, therefore more in line with what the user wants (they wanna get past this damn thing), and therefore more human.
There are oftentimes legal requirements with these things that influence the design and I’m not trying to pretend that sometimes you can’t make copy more human for legal reasons, but even then let the compliance team be the one to force your hand and use your human-first copy to at least nicely remind them of the drawback to their ways.
Trick #2: Avoid Redundancy
It’s common to overlook redundancy in UX copy because we fail to question common UI elements.
Let’s take another look at what we see in this interface without the dialog form to disguise the problems:
“Privacy policy and terms of use. To use this app please accept privacy policy and terms of use. Read full privacy policy and terms of use.” [I Accept]
It says, “Privacy policy and terms of use” 3 times. This makes me sad.
No wonder people don’t read! People don’t read because most UX copy doesn’t read well. What you see above is an example of cognitive overload. It’s just a bunch of words that don’t flow together and say the same thing repeatedly but in slightly different ways.
By removing the Title + Paragraph format and just having one sentence we are liberated from the first repeated sentence. By moving the link to the agreement into the paragraph we are liberated from the second repeat.
If your title is merely the same message as your content in fewer words, consider dropping it.
This is especially true with push notification copy so beware!
To point out some potential pitfalls of my own work here, I want to mention the risk of apologizing to your users. It may not be wise here to say, “We’re sorry to ask…” It frames the thing they’re doing as bad which you might not want to do, I could imagine my team’s feedback on this right now lol. You could also make an argument against “Hey, Jack” as it’s somewhat unnecessary and mostly there as a less-painful way to support the title+paragraph structure. I’m not a fan of, “Here is the new agreement…” as it reminds me of “Click here…” so I might work that a bit more. But you get the idea!
In Conclusion
If you’re ever struggling with your UX copy, start by writing it out just as you would say it to your best mate over a round of drinks. Start from the human end and work your way closer to the happy medium and only go legalese when forced to, making sure to shame compliance along the way ;)
Be ruthless about avoiding redundancy in your writing. The best designs are the ones that have removed the fluff and kept only the absolute necessities.
Don’t let the form of standardized UI elements dictate your message — form follows function!
Have fun, don’t stress, and love your coworkers. This isn’t life or death, it’s just design and we’re so lucky to be in this game… I might disagree with all of this in six months from now anyways.
Thanks for reading!
