The downside of delight
Sara Wachter-Boettcher explains why priotising delight in your design process means letting your users down
Sara Wachter-Boettcher on why creating delight should never be your priority
I took a friend from out of town to a restaurant in my neighbourhood the other day. She has coeliac disease – the real-deal autoimmune disorder that makes gluten attack her small intestine. Eating out is stressful; some places assume she’s just avoiding carbs, or that a little bit won’t hurt. They’re wrong.
My friend hates bothering staff, so she doesn’t go to new places much. But here she is, across the country, trusting me to lead her right. I’m anxious to find a good meal that won’t trigger her illness. She asks the host a bunch of questions: What’s safe? Are you sure about the sauces? Can you prep my food separately? They’re patient. They’re confident. They’re kind.
We’re told there’s no gluten-free dessert, but we don’t mind. Then, out of nowhere, a dish of homemade ice cream appears. It was delicious. But it wasn’t what made that
meal memorable. It was everything leading up to it: all the small ways my friend’s needs – needs most of us don’t have – were treated as valid and worthwhile.
That’s the thing about delight: it’s a wonderful feeling, but a terrible goal, because you’ll only ever get there if everything else has already gone well. Restaurants figure this out fast, or they fail. But on the web, we’re still stuck on it: Inspire and delight! Delight is the key to UX! Add delight to your designs!
It sounds good. Who wouldn’t want to leave their users grinning ear to ear? But the reality is less rosy. By laser-focusing on delight, we’re failing real people.
When delight fails
Recently, JAMA Internal Medicine released a study ( netm.ag/study-282) that looked at how smartphone AIs like Siri and Cortana handled crisis situations, such as a user asking for help with domestic violence or rape. The short answer is they didn’t. In fact, they often cracked a joke instead of attempting to look up the information.
I’m sure the engineers at Apple meant well when they loaded Siri up with jokes: they wanted to make it fun and personable. But that design direction only works when things are going well. If a user is frustrated or under stress, it feels insensitive, and it’s not helpful.
Once you start looking for misplaced ‘delight’, you find it everywhere, from the cutesy Facebook feature that wants you to ‘See what your year looked like!’ – and then shows you a picture of a dead loved one – to Microsoft’s chatbot Tay, which started out having light conversations with teens, and was immediately trained by trolls to harass women and praise Hitler.
Just an edge case?
When I wrote about that JAMA study a couple of months ago, more than a few people told me that no one would use Siri that way – and anyone who does is an idiot and deserves what they get. We can do better than blame our users.
Karen McGrane has said, “You don’t get to decide which device people use to access the internet: they do.” This is the same. You might not understand why someone would use Siri during a crisis, but that doesn’t matter. Apple gave them ‘an intelligent assistant that helps you get things done just by asking’, and they’re using it.
The real problem is that Apple never thought about tasks like these. Across our industry, design teams are trained to focus on best-case scenarios; happy users achieving desired outcomes. That’s normal. Of course we want people to have a wonderful time with our products. But sometimes, they won’t. They might be depressed. They might be in pain. They might be having a full-fledged ugly-crying public meltdown.
In other words, they’re human. Our job isn’t to judge them or push them away. It’s to help them get things done – and treat them with respect while we do so. In my book with Eric Meyer, Design
for Real Life, we call scenarios like these ‘stress cases’. A stress case is something you put at the centre of your work – a lens you use to see whether your decisions hold up under pressure.
Try it out. Whenever you find yourself imagining happy users contentedly clicking around, pause. Ask yourself, ‘What if someone in the opposite scenario used this product?’ Would it work, or would it make them feel unwelcome and unsupported? It’s not complicated or resource-intensive. All it takes is a little less ‘delight’ , and a little more compassion.
Delight is a wonderful feeling but a terrible goal, because you’ll only ever get there if everything else has already gone well