On June 20th, 2017 I launched a color contrast tool called Shade, a small app that sits in your Mac’s menubar and yells at you if your color contrast isn’t good enough.
Here’s how things went (tl;dr they went okay!)
C’mon Medium, let’s work on that green.
Shaking Things Up
I left my role at Khan Academy in April of 2017 — the best job I’ve ever had (they’re hiring!) — to pursue some side projects and figure out my source of motivation. I packed my bags and headed roughly 1.2 miles north to move apartments in the same town in New Jersey where I had lived for the past few years (Khan Academy ❤s remote, did I mention they’re hiring?).
After spending a few weeks catching up on some old video games, spending time with friends and family, and reorienting myself with a life where I don’t have to worry about a video page breaking, I decided it was time to scratch an itch and get to coding. First order of business: the contrast tool I’ve always wanted.
—
Contrary to this section’s title, I’m a creature of habit and things generally stayed the same. I still woke up at 7:30. Still made the same cup of mediocre Folgers instant coffee half its weight in creamer. Still wore the same black Calvin Klein T-shirt I wore everyday (don’t worry I have like 20 of them). Still made an effort to leave my apartment, and still didn’t really know what I wanted to do.
But I didn’t need a plan, I just wanted to make something.
So for the next 4 weeks, I spent roughly an hour a day sitting in front of my text editor, cramming as many technologies as I could into a small menubar app.
Tech & Process
I’m not big on process, but my rules for this project (or, rather, “experiment”) were as follows:
- Make something quick and cheap
- Find out the hard way if people will give you money for it
So I spent my first day getting a simple React app and running (shout out to nwb for making this painless) in an electron menubar app (shout out to menubar for this one).
About an hour later, it was up and running. I had no idea if the tech would stand the test of time, and quite frankly I didn’t really care (spoiler alert: the setup today is pretty much the same and it works fine). Momentum was my priority.
I had a billion unknowns for this “experiment,” and cataloged them in a Trello board inspired by Rich Hickey’s Hammock Driven Development. I categorized my board into Things I Don’t Know and Things I Know, and my process went as follows:
- If an idea popped in my head (i.e. “How to remove dev dependencies from the build”) I would add it to Things I Don’t Know. More conceptual things, such as “Can this be a menubar app?” also went in this column.
- I would then continue my work for the hour, then take a long break (usually a walk — again, get out of your apartment).
- When the current task at hand, whatever it was, was done — I would reorganize my Things I Don’t Know and Things I Know. I’d either have an answer to questions from earlier (maybe through thinking or through code), or I would have uncovered more unknowns.
- Over time Things I Know grew, and Things I Don’t Know gradually shrunk. This was a good feeling, and I made sure to relish in it.
This went on for about three weeks, during which I had solicited some beta testers on twitter and sent out two waves of invites to try out the app. I also settled on a name and logo, but that stuff isn’t really important. Beta feedback went to, you guessed it, the Things I Don’t Know column**.**
This work was generally very relaxed. A couple gnarly bugs popped up here and there, but for the most part I worked sustainably (my #1 priority) at my desk or on the couch watching Arcus87 on Twitch7 — a cowboy from Arizona who streams himself attempting to beat NES games as quickly as possible (I’m not kidding, Arcus’s stream is amazing and so very relaxing). In total I worked in front of a computer for about 20 hours, but in reality I spent double or triple that if you take into account the times I “took my work home with me.”
Release
Eventually I got to a point where I felt I knew everything I needed to know, with the exception of “Would people actually pay me for this thing?” A card titled just that sat alone in Things I Don’t Know, and it was time to find out.
I spent the next week filling out some tax forms for Apple, put on some finishing touches, made a landing page, designed some screenshots for the app store, and a couple dozen other things I hadn’t considered. I didn’t use Trello for this, but that wasn’t important. Momentum momentum momentum.
On the morning of June 20th, 2017, I got the email from Apple that Shade was ready for sale (I priced the app at $4.99 USD), and sent out the following tweet before hopping in my parents’ pool with my older brother.
Response, Reflections, and Beyond
Compared to previous projects I’ve “released,” the response for Shade has been…okay. Hacker News and Reddit didn’t take the bait, Product Hunt offered only a mild presence near the top, and word has spread primarily through my network (thanks to a few wonderful folks who shared it beyond my initial tweet ❤). Sales-wise, nothing extraordinary, but something I’m still very much proud of.
I’m no growth expert, but I believe this is what’s known as “Great Potential For Hockey-Stick Growth.”
To my nonexistent investors, sales for Shade have “increased in total at a promising yet declining rate” and “Q4 projections put us in the 8 or 9 figures.”
After one week I’ve moved $258 worth of product, which amounts to $170 for me after you take out Apple’s cut. Enough to cover the fees I paid to establish my LLC, but not enough to cover Apple’s license or any of my other expenses.
But this is money that I made. Not money I was given for showing up at work, money that people gave to me for a thing I created.
That’s a very cool feeling, and I’m trying my best to not forget that.
—
On an additional personal note, it just feels good to have a thing out there that people are using. Answering support emails, prioritizing bugs and features for my next version of Shade, and figuring out ways to get this in front of more people are all exciting challenges that I haven’t a damn clue how to solve.
There were plenty of moments where I wasn’t sure if Shade would sell, or if I even had the guts to charge money for this thing instead of selling it for GitHub stars, but I tried to keep true to my word: Make something quick and find out the hard way if people would pay for it. Turns out at least 50 people would.
Above all else, I’m having a lot of fun. If at some point I can use this app or some other to help cover some of my expenses, even better. For now, I’ll just keep playing around with ideas, and try repeating and refining this process on a couple other tricks I have up my sleeve.
—
Feel free to follow me on twitter where I mostly tweet bad jokes and complain about technology. Please consider buying Shade, my rent is very expensive.