How to build and care for your own sassy, quirky, funny interface robot
I run a robot store of sorts. People often come to me with their web things and say, “Make it talk”. Usually they also have some ideas about what it should sound like:
- “Make it friendly and happy. It should have lots of energy!”
- “Make it sound like your best friend. We want people to trust it.”
- “Make it quirky and fun. Like Slack or Mailchimp!”
I spend my time writing content that goes into interfaces. So when people ask me to make their web things talk, they’re usually looking for a solution that brings the content in their UI — in-product help, buttons, error messages, blank states, forms, sign-up flows — to life.
I’m generally impressed with any company that’s concerned enough with their interface content to know they need help. Too often, interface content is seen as something that can be added at the last minute or not at all (ex: auto-generated “error 1259a” type messages). It’s usually well-meaning marketers, designers or engineers who are tasked with filling in empty boxes with words that hopefully help more than they hinder.
It’s curious to me though that these requests I get are so often concerned primarily with personality, zing and panache instead of with the business of adding clarity or improving the usability of interface content. There are only a handful of companies that are succeeding at personality-driven interface. Actually, I can think of two: Mailchimp and Slack. The problem is that both companies are so adept at their content game that people don’t seem to realize a few core truths:
- They have a strategy
- They have mastered the table stakes of UI content
- They work hard to maintain their smart, clever content robot
Strategy
You don’t need someone with the job title of strategist to have a design strategy, but you do need people who regularly discuss the whys of what you’re doing and who are tireless in relating the answers to the needs of your users.
Here are some examples of basic strategic design thinking questions you should be asking before even thinking of injecting zing into your UI content:
- Why does our UI need more personality? Who or what is it in service of?
- How can our concept of personality be carried across the visual and content elements of interactions? How will designers and writers work together to create a cohesive robot?
- How do we think personality will impact trust?
- How will we design for the edge cases where personality may feel disruptive or lack empathy?
Mailchimp is a great example of a company that has thought strategically about how and when to inject humor into their experience. In Voice and Tone they help writers consider how a customer might be feeling to help them respond accordingly.
Here’s how Mailchimp tells writers to handle error messages:
- Offer a solution or next step.
- Be straightforward. Explain what’s going on right away.
- Be calm. Don’t use exclamation points or alarming words like “alert” or “immediately.”
- Be serious. Don’t joke around with people who are frustrated.
Twitter has been getting in on the personality content game lately, but I’m going to guess that they haven’t thought about it strategically. Instead of avoiding humor in their error messages, it’s one of the few places they’ve decided to inject zingers:
As Sara Wachter-Boettcher pointed out, there are a number of scenarios where this kind of content could feel tone deaf or even insulting to the person receiving it:
Whoever wrote this error message for Twitter undoubtedly had good intentions and meant to add some levity to a potentially annoying experience. Instead, the joke felt out of place in the context of Twitter’s other UI content and at worst, tone-deaf and inappropriate.
If you want a robot with a big personality living in your UI, you need to be prepared to do a lot of thinking about the “whys”. Why do you want this? Why is it right for your product and your customers? Only then can you come up with a strategy for how to train it to react accordingly.
Table Stakes
Most companies still don’t care enough about their interface content. When faced with a limited budget, almost every startup in history will employ a designer before a person who has expertise in UI content. Many big companies with huge budgets still don’t take interface content seriously enough to employ experts.
I’ve thought about this a lot and I think that the reason companies struggle to justify investing in their interface language is because they don’t recognize the expertise required to craft these tiny but powerful pieces of content. When it’s done well, it’s almost invisible and easily taken for granted. Not everyone can make a beautiful mock or knows how to create something from nothing out of code. But most people can write a few words, at least enough to get by. So that’s what they do. They get by.
As a result, apps and websites are littered with confusing UI content and behind-the-scenes, people struggle to make words fit into tiny little boxes and bubbles. What’s called a tag in one place is a label in another. We ask for information we don’t need. We throw error messages that are deeply confusing. If we’re lucky enough to have really committed users, they try to get by too.
“In business, table stakes are the minimum entry requirement for a market or business arrangement”. In UI content, table stakes are making sure that your content is clear enough that people:
- Understand what’s happening enough to trust you (this includes things like what data you’re collecting and how it will be used, and basic security around things like payment flows)
- Are able to successfully complete the core flows in your UI
- Can find what they’re looking for
- Find value in what you’re offering (if your core customer isn’t coming back for more, you have a problem and it isn’t going to be solved by an infusion of personality or a bigger marketing budget)
The last thing a company with messy UI content needs is a zingy robot piping up in surprising and weird places to further confuse and alienate customers.
Hard Work
I’m going to go back to Mailchimp for a minute because Kate Kiefer Lee has been speaking and writing about this stuff for a long time. Developing a quirky presence in your UI content is neither quick nor simple and it’s risky business. You risk annoying customers who don’t happen to like your undeniable personality. You risk coming across as insincere or unkind. If you’re a global brand, you risk cultural insensitivity. If you have mostly white, male Western folks working on your product, you risk reflecting that homogeneity out to your customers.
This isn’t to say that you shouldn’t do it anyway if you’ve made a strategic decision that it’s the right thing for your customers and your business. But prepare for it to be hard work and prepare to make lots of mistakes. It’s not about sprinkling some funnies across your 404 page or your error messages. It’s about thinking long and hard about who you are as a company and what your customers need from you in each interaction. Importantly, it’s about drawing a line so that it’s clear where jazz hands are inappropriate.
Most humans learn to modulate their behavior to the context they’re in. Through years of social education and experimentation, we’ve learned what makes people happy, what puts them at ease and what does the opposite. We usually try to avoid doing things like:
- Walking into a library or study hall and yelling “I LOVE TO EAT HOTDOGS”
- Finding someone in distress and responding with arm pit farting noises
- Trying to give a face-first hug to a terrified, growling, fang-showing pit bull
Just like you’ve learned how to navigate the infinity of social situations in your world, your chatty, fun-loving interface robot will need to take into account all the variables of your product and all the edge cases that represent the humans you’re building things for. It needs to learn how to behave and that means knowing when to shut up so that people can go about the business that they came to you for in the first place. Every time your product expands or changes, your robot needs to be re-trained, cared for and improved. After all this work, despite being the best robot trainer in the world, you may find that it just doesn’t work in your product and then you’ll need to kill that voice you invested so much in.
Robot ❤
If you’ve got a sound design strategy that points at the need for more razzle-dazzle in your UI content, if your existing content is usable and clear, and if you’re prepared to do the work, yours may be one of the very few products that is made better with big personality UI content. Go for it! Heck, you can even hire me to help.
If not, take heart in the knowledge that the thing you’re building helps people do something that matters to them. And that’s more important than all the chatty, jokey, quirky robot content in the world.
Robots are cool though.
I was inspired to write this post after reading What happens next with conversational UIs by Cennydd Bowles.