Justin Duke, founder at Buttondown

Insights:

What’s your relationship with visual design?

Ha, I feel like “relationship” can be such a loaded and interesting word. My brain immediately jumped to two things:

Those of us who do creative work, we get into it because we have good taste. But there is this gap. For the first couple years you make stuff, it’s just not that good. It’s trying to be good, it has potential, but it’s not. But your taste, the thing that got you into the game, is still killer. And your taste is why your work disappoints you.

Ira phrases this as if one aspect of the gap (taste) is relatively static whereas the other is dynamic, and your skill slowly closes the chasm between the two. My experience is sort of the opposite: I feel like I consistently strengthen my taste without making much headway on the actual execution side of things.

I’ve tried to lean into this as much as possible. Even if I almost never practice the metis of visual design for its own sake—I am always designing in service of some larger business function—I try to use my own taste as the gatekeeper as much as possible. This is an advantage of having total control: I can ship whatever I want to in theory, but I only want to ship things that feel correct to me. So often when I am in the mode of visual design it feels like a minimization problem rather than an optimization problem: how do I build this (whether this is a button or a flow or a set of patterns) in a way that disgusts and disappoints me the least?

Why did the word “toxic” spring to mind?

As I touched on towards the end there—“good design” often feels to me like a chase towards something out of reach as opposed to a craft that I can get monotonically better at. Running a company, I get so many quantitative or at least qualitative signals as to how marketing is going or engineering is going—with design, it feels like a very adversarial process with myself to not just improve as a practitioner but as an observer, which feels so painful sometimes!

How do you think your taste has improved over time?

In a phrase, proximity to users. I design software for a target audience who is… persnickety, to put it kindly—and I mean that in a good way—unlike a lot of traditional application designers who spend a lot of time designing for use cases and preferences outside their own, the majority of Buttondown’s users prefer software to look and feel the exact same way I do.

The first few times I received churn notifications from users not because the pricing was too high or because there was a painful bug, but because they didn’t like the way the thing felt, that lit a fire under me. On my best days, my personal taste feels like a process by which I can collate and metastasize thousands of voices and takes from my users on my own, because the feedback loop I get from them is so quick and tight.

How much of the feel of Buttondown is down to visual design?

A huge amount, often in ways that I think customers might not be able to verbalize. One of the things that I really try to lean into with Buttondown’s visual language is to prime them for the overall philosophy of the product, especially in contrast to related tools.

There’s an emphasis on austerity and consistency and (as much as I hate the word) minimalism that informs users’ successful relationship with the tool. When done right, my visual language tells folks “Buttondown is not going to be a rich playground or a canvas for you. There are other, better, options for that. But it will be simple and ergonomic for the things that it’s good at.” You can do that in copy, of course (and I do!)—but I think people respond to it more immediately and subconsciously by seeing what the tool is trying to convey.

What process did you go through to learn how to communicate that through visual design?

Ha, I think the word “process” implies some level of deliberate metis or artisanship which would probably overstate how thoughtful I was about the whole thing. I think the two things that really come to mind are:

  1. Just a huge level of osmosis. I think I pulled a lot of my growth and learning from the well-designed tools around me that communicated what I wanted to communicate.
  2. After a certain point, I realized I needed to shift from expanding my palette to contracting it: rather than thinking about how to use bespoke signifiers and visual elements to communicate a state or a surface, think about how I could actually reduce the options I had to communicate the same thing.

One of the striking things about the Buttondown website is the limited use of colour. Is there a reason you did that?

I probably err on the side of “I want color to overcommunicate rather than undercommunicate”. There are a lot of apps and sites I love that take advantage of a wide palette, but both in terms of Buttondown’s marketing site and the core application I want all focus to be on the prose, and color is such an easy way to distract the eye from the written word.

Have you found other ways to distract people less?

When I was a little less mature, I think my answer would have been density of content; focus on sparse content so that any given screen has a small amount of information for a user to process and upon which to act. Now, I think the answer is more about diversity of content. Users tend to respect a very dense visual field so long as that density is consistent and easily navigable, and the user testing I do shows that people tend to get more distracted when the ’noise’ comes from inconsistency rather than from volume itself.

Can you give a practical example of how you handle that diversity of content?

Reworking Buttondown’s writing interface (wherein it went from a pretty busy screen with a number of affordances to what is, in its base state, a big text area with a bunch of buttons) is probably the most recent one.

The main insight that drove this redesign was there’s a lot of information and configurability in sending out an email, but surfacing all of that immediately to a user breaks them out of how they spend the vast majority of their time. It was incorrectly optimizing for the long tail of user interactions at the expense of the 90% of the time they spend simply writing and editing.

Are there any uncommon principles that guide your visual design work?

We obliquely touched on this a little with regards to use of color, but I think the idea of ’designing for joy’ is misguided in the context of how I often hear it deployed (an emphasis on thinking that joy in its own right should supercede the joy that a user feels upon using and internalizing a useful visual frame.)

Beyond that—and Dieter Rams may hate me for this!—I constantly try to be as un-innovative as possible. I don’t think that should necessarily be a global principle, but I try for my designs to be extremely boring and unsurprising. I think if I need to teach a user how to interpret or how to understand something for the first time, either in a visual sense or an interactive one, it is probably a sign that I have erred somewhere along the way.

Have you found any drawbacks to the “boring and unsurprising” approach?

I think the hardest part is that “boring and unsurprising” to me is vastly different than for the average user, and in particular for the average new user. I have three big biases, and I have to dispel them all in slightly different ways:

  1. I am exposed to the innards of the data model and the backing functionality
  2. I am exposed to the full universe of the product
  3. I have spent a long time using the past iterations of the product

All of these pull me away from sort of the ’tabula rasa’ perspective, and I try to remedy that with as many blind walkthroughs as possible.

Is there anything else about visual design that we haven’t covered, but that you think is important?

I think the single biggest thing that has accelerated my design practice is an increased willingness to look outside the confines of ’visual design’ qua ’visual design’ to find inspiration and feedback, both from a consumptive and a productive standpoint.

The visual design of objects that I don’t think we necessarily associate with ’classical’ visual design—say, the HUD for a first person shooter or the steps for a LEGO instruction manual—contain an entire industry’s worth of lessons in terms of how to elegantly present background information or how to occlude user-generated state.

anthonyhobday.com/sideprojects/visualdesigninterviews/