When designing with data is dangerous

Sarah Winters, , User research

You know I am a big fan of designing with data. In fact, I practically come out in a rash without it.

But recently I was reminded how dangerous it can be to design with only data in mind.

There’s an absolute stack of metrics you can get from your content. How long people are on the page, which links they use, how long they spend on each section of the page, where they put their mouse etc.

You can also make assumptions about your audience. For example, if you are writing about palliative cancer care you are probably safe to assume most of your readers are going to be feeling emotion.

Assumptions are dangerous

We make assumptions about data all the time and act on them. Then we take those assumptions into the lab or into research and see if it is right. And that’s fine.

What’s not so fine is when those assumptions go unchecked.

We can’t assume because someone spent a long time on a page that they were engaged with it. Neither can we safely assume they didn’t understand it. We need to take a lot more than time on page into account.

Data in isolation is dangerous

If someone is on a page with 500 words and they take 2.5 mins and use the link at the end, then that’s good, right? They are using their scanning behaviour and using the primary action. Lovely.

What if they take 8 minutes on that page and use the link at the end? Is that good or bad? What if they don’t use the link – does that mean the call to action isn’t working?

It could mean a mountain of things. For a start – your audience: is English a second language? Does your primary target audience have reading problems? Does your audience need help to find/read/understand your information (eg assisted digital)? There are so many reasons. You can’t take any one piece of data and make accurate assumptions.

An example

So let’s take our cancer page. You want to tell carers what is going to happen to their loved one.

Your user need was:

As a partner to someone with terminal cancer,
I want to find out what will happen,
So that I can prepare my caring duties.

I’ve heard people say the audience will skip more words because of the emotional stress they are under. I’ve seen people under extreme emotional stress read every single word on a page because they are desperate not to miss anything. I’ve read a bunch of studies that support both sides of the argument.

The fact is a lot of these things don’t actually matter. When you are writing – for whatever subject – remember the basics. There’s a brilliant blog post about how people read on A list Apart by Jason Santa Maria.

Going right back to the beginning and understanding the basics of how humans read text can help you understand who is reading what and why. 80% of humans in the Western, left-to-right reading world read in exactly the same way. You will do the fixation, saccade thing mentioned in Jason’s blog post.

If you write to how people read, you are writing well. If you write to anything else, you are making your audience work harder than they need to.


Now, this is where you make a massive difference. You can help someone, give an authoritative air, put people off – whatever you want. This is where your skill comes in. You take the knowledge of how people read, add your skill as a writer, check your assumptions with testing and BOOM. Lovely content. 🙂

Don’t jump when one piece of data tells you to

I’m not saying abandon data – far from it. I love it and won’t do content design without it. I was just reminded of the perils of looking at data in isolation, not testing assumptions, and most importantly – forgetting we are writing for humans. Writing for the numbers can make us disconnect with our audience, the one thing we are not trying to do.

Sign up to our newsletter

Get content design insights sent straight to your inbox.

  • Choose what information you get: (required)