CEO and founder Tobi Lütke on UX at Shopify
How UX plays a role in everything we do at Shopify
The Inside Shopify UX podcast is back for season two… both audibly and now, visually. Listen with your favorite podcast app or watch new episodes on YouTube.
Our host, UX Director Lola Oyelayo-Pearson, will be chatting with leaders and crafters on the Shopify UX team and will be asking questions like ‘has UX killed visual design?’ and ‘Is design an inclusive practice?’.
If you’re low on time or prefer reading, continue for the TL;DL.
Watch on YouTube and Spotify, or subscribe wherever you listen to podcasts.
When Shopify founder and CEO Tobi Lütke first started out, it was the era of the full stack web developer. At that time it was just called “doing things on the internet,” versus the robust design and development practices we cultivate in the industry now. Growing up in Germany in the eighties and nineties, he was always surrounded by good design — they had Dieter Rams designs in their homes. Tobi described: “It was a time of terrible haircuts, but also really good product design.”
A Toronto-Raptors superfan of UX
Tobi is an engineer and self proclaims that he’s “extremely bad at designing things”. But he’s a big fan of UX. He says, “I consider myself sort of a Toronto-Raptors superfan of UX. I jump around when I see something amazing. It’s just such an incredible discipline because, the most powerful thing that we’ve been working on in the last half century is computing and software, but none of it can be utilized to the benefit of people without actually making it sensible and approachable.”
Balancing design
How do you balance designing for the user but also have a strong opinion? Tobi thinks that people end up overemphasizing the quantifiable. Being highly utilitarian can be a very reductionist way of looking at things. He shares an example, “The biggest exports of the German economy certainly are the cars and I do think they are good examples of things that are not purely utilitarian. Logistically from A to B, you can get everywhere a lot cheaper. […]
“I think the reason the cars that we are now picturing in our minds, why they do so well is because everyone who works on them knows this. These are not people cosplaying car builders. These are people who have an extraordinarily strong sense of who they are building for and they tend to be okay with not building for everyone.”
He goes on to say if you’re creating something and no one hates it, that may be a sign that you’ve designed something that ends up being middle of the road and you may not have fully committed to a particular vision.
Narrowing in, how do we balance design at Shopify? Lola notes that because Polaris is so robust and powerful, Shopify is seen as an innovator in design systems. We have openly shared both the thought process behind developing Polaris and even specific components. Now there are a bunch of places on the internet that look like Shopify because other teams have borrowed it. Open design is important to further the industry, but does that become a burden for us as we continue trying to innovate?
Tobi responds that Shopify is very big software: it has over two and half million lines of code in TypeScript — just for the UX. For this reason alone, we get an incredible amount of value out of a design system. He says, “Design systems are incredible at raising the floor. Where they can be problematic is if they lower the ceiling.”
“Design systems are incredible at raising the floor. They can be problematic if they lower the ceiling.”
Design systems can feel like a set of constraints, even though they aren’t. Though, it is a fine line that requires constant adjustment. And that line is not due to restrictions in code; it’s related to the design culture at your organization. Tobi says teams should be asking, “What do you believe this system should do for you? Does it help you get started? Or is it a setup that pushes conformity and sameness?” Your team should be having a constant conversation about your design system.
If you want to hear more about whether design systems stifle creativity, listen to episode 2 of the podcast Did UX kill visual design? with Yesenia Perez-Cruz and Roy Stanfield.
Getting jazzy with constraints
Tobi believes the best teams tend to resemble jazz bands. One person sets the key and the pace, and everyone else brings everything they know about their instruments. From there it’s fluid and changes depending on the style of music. He also believes constraints can be good if they facilitate the opportunity for more creativity. For instance, a constraint specific to the jazz analogy is time. If the band decides they’re going to play for an hour, that’s a hard constraint that influences what they create and how they sound.
“I actually learned a lot from the UX world and jazz,” Tobi said, “I think one thing that really helps in the UX world is also learning engineering principles, because every discipline is figuring out a lot of important things which then become relevant to them. So one thing I always push, and it always horrifies people, is deleting code, like many very important features in Shopify.”
Tobi explains it’s not just subtracting code but it’s also when building net new features. He’s asked the team to start the day with nothing and ship implementable features in 24 hours.
“This sounds surprising that this is an approach, but it led to probably some of the best code we’ve ever written. […] Almost all programming is not actually writing code. You’re actually building a model of a solution. You’re actually exploring the problem set deeply by trying to reflect it as a model in code. And then as an idea about UX and so on.”
“So 90% of work on a project is actually building up off your mental model using your creativity to find novel solutions to the same problem.” And UX might be even better. Tobi says, “Your deliverable at the end of the day is something that everyone can reason about rather than just other engineers.”
“Ninety percent of work on a project is actually building up off your mental model using your creativity to find novel solutions to the same problem.”
An early investment in UX
Shopify wasn’t the first e-commerce software. It was designed for new brands to get their start, which was a novelty at the time when most software was designed around the big brands who already had money. Shopify made starting a business accessible to people that have never run one before, people that wanted to make their side hustles a reality. And that means it had to be built with really good UX. Tobi says, “We have to make something that’s approachable to the people who want to reach for independence during the lunch break.”
In the 15 years since Shopify’s official launch (17 if you count Snowdevil), there are going to be some things we’ve shipped that may not be the best designs, or were just plain ugly. But Lola wants to know what they are. After skirting the question, Tobi gets specific and points to the first iteration of shopify.com and his business cards. He encourages Shopifolk to look back and see where Shopify has been (via Wayback Machine), so we can understand how far we’ve come and that there’s always room for growth.
Our CEO and UX superfan has always been involved in design at Shopify. He tells a story about working with co-founder and former Chief Design Officer, Daniel Weinand. Tobi shares, “Daniel asked me to stop cosplaying being a UX designer. That I should just do the programming.”
Before design systems existed and before the days of code reviews and pull requests, they agreed to a special CSS class that signified where Daniel should go in and “fix” Tobi’s designs before they went into production.
Even now with a team of over 500 UXers, Tobi is still very present in the design process at Shopify. He gives feedback, attends review meetings, and hangs out in UX Slack channels. Lola asks Tobi if the team would be able to ship something he doesn’t like but does the job well.
Tobi says yes, absolutely, and continues, “My contribution is I’ve seen all 17 years of this company. I know the conversations leading up to the conversations. So I can help discover what’s behind things and I know the merchants really well. I talked with tens of thousands of them over the years. I know what they tell me about the early days. I have empathy because I sell myself — I’m a user of Shopify. So the thing I tend to give the most feedback on is when I see category error around the empathy and mental state that users might be in.”
A specific example Tobi points to is a recent exploration of charts. He tries to poke holes in designers’ use of the color red. Red is a great solution in the context of destructive operations, but if used incorrectly, it can make merchants worry when they shouldn’t. In the week after Black Friday and Cyber Monday, a merchant will probably see a week over week reduction in sales. In this case using red to emphasize the decline could alarm merchants unnecessarily. Using gray for that expected dip and potentially using green for the growth they saw over the holiday could be a more empathetic choice.
Kiss, marry, kill: UX edition
Now for a game! Of course, we’re playing the work-friendly version of the popular game: kiss, marry, kill. Lola shares that to play the game Tobi must tell us which product he’d kiss (a fleeting fancy), marry (start a long-term relationship), or kill (well, you know). His choices are Twitter, Discord, and Reddit.
Tobi’s answer? Kiss Discord, marry Reddit, and kill Twitter.
He explains, “Hmm. I kill Twitter because Twitter I have the strongest emotion to. I marry Reddit just because of the wonderful frivolousness and crazy that’s there. Honestly, here’s the thing, these are like Rorschach tests, right? My filter bubble on Reddit, it’s just very different from my filter bubble on Twitter. I can’t decide if Discord is great UX or not. It’s sort of like this hot mess that somehow always has these moments of delight in it.”
Lola’s answer? Kiss Reddit, marry Discord, and kill Twitter.
Lola explains her answers, “My answer to that is I would also kill Twitter just because it has this impermanence where people kind of basically shit statements and then run off and create controversy. I’d probably kiss Reddit just because I feel like it’s great, but I find it’s kind of like the precursor to what Discord is, which is much more community-oriented, deeper conversations, and higher quality sources. If I get something from Discord, I feel like I could probably trust it a bit more than Reddit and a heck of a lot more than anything I got off Twitter.”
She goes on to speak to the design of each three, “Twitter is probably the friendliest in terms of the user experience, but I have found that Discord and Reddit have been highly intuitive and that’s something that’s my holy grail of design — you look at it and you can have all sorts of opinions, but everything you think you want to do is extremely intuitive. And so you’re basically learning with every single click, every single action does exactly what you expect. And I kind of feel like Discord is nailing that.”
Is delight controversial?
When diving into Discord’s UX decisions, Tobi mentions that he thinks delight should be the goal when it comes to designing interfaces. Lola counters that delight is a controversial term in UX now.
She elaborates saying that designers are wary that delight comes off as frivolous and they ask, “If you start with delight, are you actually problem-solving first?” Tobi says it depends on where you put delight in your experience and to only add it after the utilitarian values are finalized. He says “It’s almost like a display of mastery over the task, that can only be achieved by the teams that so fully understood it, that they ended up being able to spend the little bit of extra time — just to outperform, but even reasonably well-versed people in the tool of would ever expect of the team.”
Tobi adds that delight is that cherry on top after everything else. It’s after you’ve found the venn diagram of all the stakeholder requirements of an interface and find the perfect spot in the middle — that’s where delight goes.
It’s a great time to be a designer
At the end of their conversation, Tobi explains why it’s a great time to be a designer at Shopify. He says, “It’s a really fun company. You’re building awesome stuff and it’s a fantastic place to spend a couple of years surrounded by amazing people like Lola and team, working on really important things that help millions of small businesses reach for independence and make huge contributions to local economies.”
(Join us.)
Extra credit
Want to dig in more on the topics discussed in this episode?
- Lola Oyelayo-Pearson
- Tobi Lütke
- Polaris design system
- “Physics of Passion: The Koolaid Point” by Kathy Sierra
- Wayback Machine Internet Archive
- Kano model: a theory for product development and customer satisfaction developed in the 1980s by Professor Noriaki Kano, which classifies customer preferences into five categories.
- Crafting Interpreters by Robert Nystrom
- A Pattern Language by Christopher W. Alexander, Sara Ishikawa, and Murray Silverstein; with Max Jacobson, Ingrid Fiksdahl-King, and Shlomo Angel
- Read the full transcript.
