GitHub's Joel Califa on keeping sight of your users

In his recent post “Subverted Design”, GitHub Senior Product Designer Joel Califa reminds us that if we gear too far to toward the latter, we can lose sight of our all-important users in the process.

Before linking up with GitHub, Joel was a product design manager at Digital Ocean. He’s also taught front-end web development at General Assembly and been a Design Mentor at Andreessen Horowitz.

I recently hosted Joel on our podcast to learn more about the root cause of subverted design, the shifting responsibilities of a designer, how to make a big impact with small design changes and much more. If you enjoy the conversation, check out more episodes of our podcast. You can subscribe to it on iTunes or grab the RSS feed in your player of choice.

What follows is a lightly edited transcript of our chat. Short on time? Here are five quick takeaways:

1. When designing for developers, who often build their own workarounds for less-than-ideal interfaces, be cognizant of and account for change aversion.
2. Designers should regularly search for “tiny wins” – small, standalone changes they can introduce that are lower on effort but high in impact. Often these are shortcuts that eliminate or go around unnecessary workflow steps.
3. In most cases, tiny wins aren’t features your customers are requesting, simply because they assume the steps you could eliminate are part of life. To find them, ask yourself, what about your most used flows might be frustrating? What regularly takes time or cognitive load?
4. Over the past decade, the perceived value of design has risen, and that’s a good thing. But one side effect has been a tendency to measure a designer’s worth based on how they meet business goals, rather than user needs.
5. To make sure you don’t lose sight of your users as you climb the career ladder, use your product all of the time. Test its limitations. Locate where the user frustrations might be.

Adam Risman: Joel, welcome to Inside Intercom. Can you give us a brief rundown of how you got to where you are today?

Joel Califa: When I was 12-13 I got into Photoshop, HTML and various programming languages as a hobby. I very quickly started freelancing, doing album covers and websites, and at 14, I built a tech company’s intranet.

After the Army, I joined a design agency in Israel, which is where I’m from. It was a great intro into how things are done at a professional level. I was a visual designer and did everything from waterfall-type UX to visual design to front end to back end. I learned a lot there, but the biggest thing I learned was that I wasn’t a good designer. I went back to school to fill in some foundational gaps I had in terms of typography, color theory and usability, and user experience became really interesting to me. That’s what ended up bringing me to New York, and I’ve been here ever since.

Initially during school, I started working at a Y Combinator startup that made outreach tools for nonprofits. That was really fun and super fascinating, but some point, as many startups do, it went through some tough times and we all had to find new gigs. Around that time, my portfolio was hosted on one of these old hosting sites like HostMonster or HostGator. It wasn’t super-fast and I felt while looking for a new job I had to make a really good impression. The thing I landed on was DigitalOcean, which let me put it on a cloud server. The experience there was so nice and full of potential I thought that I actually applied for a job there. I ended up working at DigitalOcean for the next three years, initially as a designer and then leading the organization for the next two. Designing and coding for my entire life and working at a developer tools company, all of it lined up to bring me to GitHub a half-year ago.

Adam: During your time at DigitalOcean you moved into management, and today at GitHub you’re now in a senior contributor role. Was that pendulum swing a conscious decision?

I was getting antsy about coming back to actual product design challenges.

Joel: It was. There’s a lot of things I really love about management – empowering people, building a strong organization, working across different departments and making sure decision-making is functioning well, etc. I’m a big nerd about all that stuff.

There were certain aspects of the job that burned me out over the last two years. I was not only ready for a break from that, but I was also getting antsy about coming back to actual product design challenges. After six months of GitHub, I’m so satisfied with the decision, because it’s super fun.

Adam: Those range of experiences really complement each other too. A lot of the really good individual contributors have had management experience in the past. At the same time, a lot of the best managers or people who very recently were at the coalface.

Joel: It’s given me more context for how organizations work. It’s made me a better individual contributor and leader than I was before. I know how to rally teams now. I know what to look out for on a higher level. It was definitely been useful experience.

The challenge of designing for developers

Adam: Today you’re designing at GitHub, which is unique because your typical user is a developer and that means they know GitHub inside and out. They are all essentially power users, and they probably don’t like it when you make changes to their workflow. How does your team think about things like change aversion and the cost of improvement?

Joel: That’s a great question. Developers are indeed the biggest power users I’ve ever met. They’ve essentially been my users for the past 3.5 years (including DigitalOcean). Something I can definitely validate for you is that power users hate change. I remember I changed the sidebar at DigitalOcean to a top navigation in order to unlock a lot of things that we wanted to do, and everyone went crazy. This small change really bothered so many people, because they were used to clicking things in that specific place and they were so averse that to that change. The reason for that is that developers find ways to make even the worst interface really efficient for them. They get used to things and they work around things and they build hacks for things. Then, any change you introduce, even a really positive change, is going to throw that efficient flow off balance.

Developers find ways to make even the worst interface really efficient for them.

Adam: They’re basically turning weaknesses into strengths. That’s got to be such a complex part of your user testing process.

Joel: It is. People like absolutely horrendous stuff. It’s something that’s hard to deal with, especially because of the variety of workflows that exist out there. Different engineers have created very different flows and have built very different systems for themselves within GitHub. Accounting for all of that is something that we need to do, but is also kind of impossible to do.

We don’t really have a framework for this at GitHub, but I think we try to take more of a common sense approach, which just means we take into account all of the context. How representative is this specific power user for the entire user base that we want to serve? We have 20 million engineers and usually the power users are not the majority of them. How disruptive is a change going to be initially? Is that going to be offset by the changes potential? Lately we’ve been biasing towards introducing these difficult changes rather than ignoring them or punting on them.

The value of designing tiny wins

Adam: Not all the changes your teams have been introducing have been big or difficult. In fact, you wrote recently about a couple of “tiny wins” your design team shipped. These weren’t necessarily things that your users were asking for, but based on the social proof, your users loved them. The impact of these changes vastly outweighed the effort. So, how would you define those these tiny wins, and when and why should design teams be seeking those out?

Joel: Tiny wins are few things. They’re standalone, which means they provide their own value and they aren’t just part of a larger project. They’re not just a tiny thing that requires you to see all these other things to see what it amounts to. On their own they are valuable. They’re low effort, which means it’s straightforward and takes a short amount of time to build these features. They’re high impact, and that means that they effect very high volume flows in a positive way. Things that your users use every day. Even if they save just a couple of seconds, they can create this cumulative impact with every use of that flow.

A useful way to think about these is as shortcuts. They get rid of existing steps, or they go around them. These can be physical or mental steps, like just having to think about or click something. Then, they make doing the thing easier. A few examples are Netflix skip intro button or Chrome’s icon for which tab is currently making noise, which lets you shortcut clicking through every single one of your tabs before you find it. My work at GitHub was turning vague icon into an arrow or changing the favicons to be dynamic so people know what’s happening inside a tab.

Ever get confused between your “base” and “compare” branches? Shipped a tiny @github change today that will hopefully clear things up. pic.twitter.com/acOddhxLDs— Joel Califa (@notdetails) December 4, 2017

There are few reasons wins are really awesome. The first one is that company momentum breeds trust. When you see these frequent changes that make your life better, that translates into the company actually caring about you, and in turn it can make you care about the company and breed some loyalty. You can see that from the love fest that comes through on Twitter after any one of these pieces ships. People are like “Nobel Prize to the Netflix skip intro button.” They really love that. It’s a seemingly small frustration. Everyone was thinking, “This is part of life”, but they now understand it doesn’t have to be. They’re feeling the emotions from that.

Adam: That “it’s a part of life” angle is interesting, because a lot of these opportunities aren’t what you would get as common feature request. What’s your advice for identifying tiny wins?

Joel: That’s totally true. Any of the tiny wins I listed aren’t things that were very clearly needed. No one had asked for them. It’s like that Ford faster horses quote. If you ask people what they need, they’ll say faster horses. They won’t say a car, because they can’t imagine it.

Company momentum breeds trust.

I don’t have like a silver bullet solution, but in the blog post I suggested a few questions to ask as you put together a list of potential tiny wins to build. It’s more of an introspective process rather than going out and trying to figure things out. Once you have a list, you can validate that and see if these things are real.

What are your product’s most frequently used flows? Think about that. What about those flows might be frustrating? What regularly takes time or cognitive load? This could be an extra click or maybe an ambiguous component. How frustrating do you think these moments are, and what’s the sum of time or frustration that fixing each of these is going to save? Is it going to be noticeable? Are people going to enjoy it? Will it create that joy? Will it be shareable? Asking these questions can give you that first list and then you can validate it and figure out what you want to do with it.

Adam: Is a big part of that validation process keeping things in scope? One thing we talk about at Intercom is that there really are no small changes. There’s trickle-down effects for all these things.

Joel: For sure. I include as many relevant people as possible in the process. Get your engineers in there early to help you scope these potential tiny wins. I’ve definitely worked on a ton of projects that end up becoming much, much larger and longer than they should have been. I think when you start on a feature like this, you can make a conscious decision not to pursue it if there’s many known unknowns. Then, if you decide to pursue it and you encounter a huge unknown unknown midway through, then you should ask yourself whether it’s worth continuing. The point of this isn’t to build this no matter what. It’s to build it because it’s easy to build and you get that ROI.

How design becomes subverted

Adam: One of the reasons we were really excited to talk with you was your recent post, Subverted Design. What was your thesis for the piece?

Joel: If you look at our roles a decade ago, a UX designer’s role was pretty clear. They removed friction and complexity between a user and their goal. They clarified systems. It was an interesting role but compared to today, it was fairly small in scope. You are responsible for this one part of the product development process. We weren’t super important – no one felt like they needed a designer in order to make it, which was a shame as a designer.

Over the past decade, design’s perceived value has been steadily rising, probably as a result of companies like Apple or Airbnb that were very vocal about the importance of design or companies like Facebook investing very heavily in design. The entire industry has shifted gears. Today, you can look at almost any industry and you’ll see a company that’s differentiated by their design.

With that shift in the perceived importance of our field and our importance as individual designers, there’s also been a shift in what’s expected of us. The new narrative in the industry is what makes a designer valuable is their pragmatism and their ability to achieve company goals. Today if you see a designer who’s very idealistic, the “cliché designer”, one that pushes for things to be perfect all the time for their users, they’re usually treated more as a junior designer. That wasn’t always the case. Every designer was a bit headstrong and really into their users, and it created some positive friction between designers and other stakeholders.

Today, what we see as senior designers are people who understand the business.

Today, what we see as senior designers are people who understand the business. They care about adoption and conversion and generally they’re really amenable and easy to work with, which on the face of it is a great thing. It’s easy to buy into it because there’s a positive feedback loop. When you start thinking of prioritizing business needs and speaking the same language as other stakeholders in your company’s leadership, you get treated with more and more respect. In that respect, it feels really good and it’s generally an indicator of you being on the right track. It can come with promotions. It can come with things that are associated with good feelings.

Prioritizing business value, which comes with all of this prestige and aligns with the entire industry and what it sees as important, how do you not choose that over prioritizing users and “the experience”, which is something that frames you as inexperienced or childish? More and more designers, including myself, increasingly have centered their work around company goals. Thinking about those generally ends up resulting in good outcomes for the designers.

Adam: Designers have been advocating for this seat at the table for so long, but it sounds like you’re saying the pendulum has swung too far in some ways.

Joel: Totally. I still cared about people using my product, but my perspective and responsibilities had shifted. It’s ironic, because you’re right, we have been fighting for a seat at the table for very long and for people to get design and understand what the tools can do for them. Now that they do and design is more important than ever, we’re using these skill sets and tools to benefit businesses rather than our original stakeholders, which are the users.

Adam: Originally this post started as a Twitter thread. Was there an actual moment that caused you to put pen to paper, or was this again just a gradual itch in the back of your head?

Joel: It’s been there for a while. Especially with the current government and the current climate in the world, a lot of people in tech are doing a lot of introspection this past year, about whether tech is good. The thing that ended up triggering it was finding out about the faith of the Facebook deactivation page. Basically, when you try and deactivate your Facebook account it shows you photos of a bunch of your friends, people you’ve interacted with recently, and tells you “Hey, Dana will miss you. John will miss you.” It says you won’t be able to get in touch with them or any of your other friends, which to me is going a step too far. That’s kind of emotional blackmail.

Joel’s experience with Facebook’s deactivation page. Read his full post on subverted design here.

Seeing that and understanding the alternatives of what they could’ve done just resulted in a Twitter rant. I don’t have much of a filter on Twitter, so it’s easy for things that I’m thinking to kind of end up on there. I ended up writing about it more thoroughly because of how much it resonated with everyone in the community. Twitter threads tend to disappear into the ether and I wanted to immortalize it somewhere.

Adam: A lot of these ideas start out innocent in the beginning. Where do you think they lose their way? I don’t think a lot of these designers are going into these projects with bad intentions. Is it more that the goals are misshapen?

Joel: This stuff is insidious. It’s not something that’s super obvious. It’s baked into the culture, and honestly, it’s baked into capitalism. The entire system runs on businesses thriving. Incentive structures run on businesses thriving. Even when you aren’t doing stuff on purpose, it ends up getting shifted towards a certain direction.

I think you’re right. A lot of features start out being very innocent and even as good features. A great example of this is Facebook safety check. Do you remember that? Safety check started as this super wonderful and actually very pure feature that I think was started bottom-up as a hackathon. After accidents or natural disasters like earthquakes, if Facebook detects that you’re in a certain area, you can mark yourself as “safe” to set your family and friend’s mind at ease. The original goal there was awesome. It was to calm people down and tell them “Hey, even if I don’t have reception, just so you know, I’m okay. You don’t have to freak out.”

A while back there was a terrorist attack in New York, which was dubbed by Facebook as “the violent incident in Manhattan”. I received a notification from Facebook that said, “Hey, George and John and Kelly want to know if you’re okay.” These specific friends of mine, they are asking if you’re okay because I’m in New York, even though most of those people know I’m not in that area usually.

I talked to one of them afterwards, because that notification seemed really odd considering I’ve known this feature before. They told me that the way it worked was Facebook told them that I was in the area and then prompted them to message me like, “Hey would you like to ask Joel if he’s okay because he’s in that area?” It’s a really subtle shift. Instead of calming my friends down now, which was the original intention, Facebook was now literally telling them, “Hey, this terrible thing happened and Joel might be there. Do you want to check?”

They might not even be thinking about me being there, but Facebook is telling them I might be dead right now. All of this stuff ties back to business needs. That decision wasn’t made in a vacuum. Everything Facebook does is tied to engagement and people using the system more frequently. A result the original intention of what was a very well-meaning feature is being undermined. That’s sad. We might barely notice that we’re doing it, but we are doing it and the results are real.

Adam: As someone who’s been in both the leadership role, where I’ve assumed you’ve to answer to business needs and come up with solutions that fit those as well as user needs, how do you advise designers when it comes to staying close to their users as they move into their leadership role?

It’s important to actually use your products as much as you can.

Joel: Keep talking to them is one thing. Also, check yourself. When you’re in a company it’s really easy to drink that company’s Kool-Aid. It’s really easy to rationalize decisions. I’m sure those Facebook designers had good intentions at heart. It’s also important to actually use your products as much as you can. So, I have a friend who works at Deliveroo, which is a delivery service out of London. She actually rides around sometimes to better empathize with her users. She delivers food to use her own app and see what frustrations might be there.

Similarly, ClassPass doesn’t provide their employees with unlimited classes. They give them a very small package. If they want to actually use the product, it needs to be worth it, which makes you actually inspect and explore different ways that the pricing model can function. It makes you actually look at the thing versus just taking for granted that it’s fine, which makes you less biased and maybe less forgiving about things that can be hostile to users.

Adam: I know this is a problem with no easy answers and what we’re looking for more is a broader conversation. So, where would you like to see this conversation go next?

Joel: I’m a big fan of Tristan Harris’s idea of time well spent. He has an organization now that’s trying to push companies into more ethical decision-making. I would like tech at large to start introspecting more. I wrote the blog post about designers because I am a designer and that’s what I can speak to. I’m not comfortable about telling people what to do. I wouldn’t feel comfortable about saying, “PMs need to change their roles in this in this way.” A a designer, I think I can be the catalyst to some kind of change, but honestly, I can’t change things myself. As much as we now feel that we are the center of the world in this industry, we’re just a small part of a machine. In order to have real change, you need that entire machine to start shifting what it cares about.

Adam: I think a open discussion from all parties is the first place to start. Joel, this has been great. Where can our listeners go to read more of your insights or generally keep up with your work, what’s going on with you?

Joel: You can go to my blog. I’ve been trying to write very high quality materials so it usually comes out like once a month. Last month was pretty special with me writing three blog posts.

When I write blog posts, a thing I’ve started doing is sending out additional material in my newsletter. So, if you’re interested in that stuff, feel free to sign up. Then, all of this stuff always start out on Twitter. So, just follow me there. You’ll probably get most of it.