In June 2020 37signals released HEY into the world, the culmination of 2+ years of explorations and intensely focused work.

It goes without saying, but I’ll say it anyway: inventing a new product from scratch is one hell of a challenge. It’s the toughest thing you’ll ever do as a product team. There are a million reasons why it won’t work and zero guarantees that it will, so the whole project is a massive gamble. You just have to buckle up and trust that you’ll figure things out.

Since HEY made a big splash on arrival, I thought it’d be fun to share the backstory of how we ended up reinventing email. Because we certainly didn’t start by wanting to reinvent email. (That sounds hard and intractable!)

For the past several years, even pre-dating the work on HEY, our team noticed some ongoing communication problems at our company. They all revolved around external email: things like going back and forth with lawyers, working with benefits providers, dealing with press outlets, accountants, and stuff like that.

It was too hard to keep the right people in the loop, because we’d have lots of email threads scattered around with different participants, and none of them had the complete story. It was too hard to get caught up on a project that someone else was working on with a vendor. And too hard to hand work off when someone went on vacation or parental leave. At best, you’d end up with a steaming hot pile of forwarded emails, and it was all a chaotic mess.

So we began thinking about how we might solve these problems with a new product.

At first, we were considering making a successor to the company's older CRM product, Highrise, but it doesn’t do nearly enough to cover all the scenarios we wanted to solve.

And thus, HEY started its life as a prototype for a possible Highrise 2. It was originally focused around work problems — not personal email problems.

In May of 2018, we kicked off the first work on a prototype, which was code-named Haystack. We were intent on addressing real-life situations right away, so I asked our Head of People Ops to find the gnarliest vendor email threads she’d been involved with recently. She forwarded all the emails to me, along with any internal chats she had about those emails. Using that source material, I painstakingly reconstructed all the conversations into seed data for our prototype. That way we’d have some real stuff to judge our ideas against.

We started by designing how these internal/external threads might look and work, then gradually worked our way out to the Inbox view, and finally filled in various supplemental ideas around the edges. We focused specifically on email pain points and discarded any notion of how email “usually works” in favor of the way we would ideally want it to work, while also considering what we could realistically foresee building.

After a few months of prototyping, it looked like this.

Thread view

Thread with comments

Early Inbox

Early Screener (Bouncer)

All Files view

Contact page concept

Having explored lots of different examples and core ideas (including the nascent version of The Screener, the All Files view, and contact pages), we felt like we had something promising.

Building it

By early 2019 we started the process of making HEY real. At first, we used the prototype’s general skeleton as a foundation to start wiring things up, but we continued to hone the app’s visual design.

Tightening up, adding features, and doing all the hard work

Designing with a team where everyone has autonomy is a bit like playing in a band. Someone seeds an idea, and then someone else builds on it and takes it somewhere else, then those changes get promoted into little systems, and over time the best ideas persist through all the changes. Harmonies prevail.

You can see that taking place in the screenshots above, but it really picked up steam from mid-2019 on, when the whole company started working in earnest. For a while, there was a lot of tug-of-war design and development between the web and mobile teams, where all of us bent and stretched the system for what we needed on our respective platforms. Since HEY is a hybrid web app, we ended up building a ton of new tech to optimize for UI patterns that would work equally well on a phone or a laptop.

By early 2020 we had worked out most of the fundamentals. We had Reply Later and Set Aside, and we had The Feed, which at that time was called “Slowbox”.

From then on, the remaining work was a recurring cycle of polishing what we already had and using the product ourselves to figure out what was still missing. For example, in January I noticed I was having trouble filling the gobs of junky capitalism emails I constantly receive, so I pitched the idea of a special place just for that, and the Paper Trail was born.

We spent the remaining months adding all the many, many, many things you need to release something to the outside world, like billing, trials, onboarding, exports, cancellation, security settings, 2FA, password resets, contact imports, forwarding settings, and a seemingly endless parade of email rendering challenges.

All of that infrastructure work is an amazing achievement for our small team. Creating a reliable, robust, functional email service is no joke. We obviously had big ideas and wanted to do something new and inventive, but we also had to do something old and unsexy: make the thing actually work as well as products like Gmail. Those platforms have been around for 15 years and have massive amounts of engineering talent backing them up—and we’re just a handful of people working our tails off to make the best thing we could. It took the full force of every single person at 37signals to make this happen.