Design is a Series of Overcorrections

I’ve been watching my fair share of WWDC sessions this week and one thing I kept noticing were times Apple engineers explained new features by taking subtle jabs at their old versions of those same features.

A great example of this are the new buttons on watchOS 7 that show up at the top and bottom of scrolling lists. So instead of being in Messages and force pressing the screen to get the “new message” button, they now have a button at the top of your list of conversations. They explained this is more discoverable for users, and that you should really update your app to do this because you’ll have happier customers.

The snarky reading of this is, “Apple is bad and they had a bad UI for years. This is an admission that they were bad all this time.” I don’t think this is a fair complaint.

Design is really just a series of overcorrections until you land on something great. That definitely happens in the design process for each feature/release, as you likely start by looking at the problems with the current solution and then make sweeping changes to those things. Those changes usually get filed down and aren’t quite as extreme when they get released to the public.

But even then, you are likely to overcorrect, or go too far in one direction. That just means you take the good stuff from the current release and, you guessed it, overcorrect back the other way and have to ease up your changes again. You keep doing this until you get to something that strikes a good balance.

And sometimes these changes take place over years, or decades. Look at iOS as a great example. iPhoneOS 1 was the first time most people used a personal touch screen device and lots of things in the UI had to be literally spelled out for you (“Swipe to Unlock”), and to make things feel tactile, rich textures were used, as were lots and lots of visible UI.

After 6 years of that design, iOS 7 switched everything up and ripped away tons of things that helped people get to grips with a touch-based phone. It wasn’t that the early versions of iOS were bad, it was that the needs of the market changed. The early versions of iOS trained people how to use a phone, and once that groundwork was laid (and popular design trends shifted) Apple had to change the look of everything to match people’s new needs. They of course overcorrected and iOS 7 swung too far in the “flat” direction, but they’ve iterated on it a lot and the latest version of iOS is a much more sensible mix of “clean design” and a UI that feels tactile.

Look at a designer’s computer and you’ll find it littered with drafts of designs, most of which are too extreme to ship. Maybe a search box is too large from when you wanted to push search. Maybe the design is too playful, which was a response to users saying your UI was stale. Maybe you stripped away too much UI chrome because your app was cluttered. There are tons of designs that never see the light of day, and most products make it to market with a bunch of overcorrections still in place. And because good design is a moving target in software, what was considered a good design one year could be considered an overcorrection just a year or two later.

The Rejection of Design, and a Reversion to Checkbox Comparisons

One of the things I see often from people who don’t like the look of Hey is “I can do all of this in Gmail with labels and blocking senders.”

My response would be, “sure, you could probably do a lot of this in Gmail as well, but that’s not the point.”

I think that the thing people don’t bring up in this conversation is design. The way Gmail and Hey are designed is totally different, and the apps push you into doing different things with your email. Gmail is a complicated service by default, with 4 dort of inboxes, numerous labels (you probably have 100 labels if you’ve used a bunch of third party apps in the past), and a UI that’s built to work for many millions of people because Gmail has become the default email service over the past decade and half.

Hey on the other hand is designed around a specific handful of workflows. It’s designed around a screener system that makes you manually allow each sender to access your email at all (I’ve blocked several spammy “can we write a sponsored post on your site” 🤮 emails already), which is a mental shift from Gmail’s in-by-default view. And then little things like removing the need to archive messages, options to create mini-todo lists for things you will reply to later, and more.

These design details and suggested workflows make Hey a different experience from Gmail, and while it will not appeal to everyone, I think that suggesting that “most of this can be replicated in Gmail,” is selling design short and reverts to checkbox comparisons that are not always useful.

Here’s another example: macOS and Windows. I would wager that you could get your work done on either platform, especially if you do a bunch on the web where the functionality is exactly the same. However, despite sharing almost entirely the same features, you likely have a strong opinion on which platform you prefer.