// Insights

How Amazon Builds Remarkable Products

Sean Johnson
Founding Partner

Chris Durbin is a product designer for Amazon. In this discussion we discuss what she learned leading the design for the Apple Watch app, why Amazon's culture of writing is the key to its success, and why Amazon believes the phrase "Minimum Viable Product" should be outlawed.

Manifold: You've done a bunch of things in your career. When someone comes up to today at a networking event or a party and asks you what you do, what do you say?

Chris: It depends on the audience. I usually keep it simple and say "I like to make stuff at Amazon, and I watch people use the stuff I make, and then I go back and make it better."

Manifold: You're mostly known for the user experience side, which has evolved quite a bit over the years. It used to be websites, and then it was mobile, and now it's almost device-agnostic, especially at Amazon. How do you as a UX professional navigate that, designing and optimizing for every different type of form factor?

Chris: Once nice thing about Amazon is all the internal talks and events and just opportunities to immerse yourself in another team's world. But any time you approach design at Amazon, you have to remember it's worldwide. You always have to be thinking, "I'm building this right now for my audience today, but in one to two years we will have to scale." And scaling at Amazon includes Alexa devices.

I usually start mobile first. I think about if someone has a phone and an Android phone that's a couple of years old, how are they going to use it?

But all the while I also have to think about the distilled experience. I have to ask myself if someone is using an Echo device and they want to use my product, what are the prompts and things they will say? So you have to really be crisp in what you want the person to do and how you keep them moving through the flow.

Manifold: Does that lead you to focus and remove anything extraneous? Or do you still visualize a full featured product, and as you're working on identify use cases for a voice interface?

Chris: I like to start with the customer and work backwards. I try to design as robust as system as possible. I go through every use case, and then leave it up to either the product manager or developer to say either we can't do this or I'm glad you thought about this, but we're not going to do this right now. So there's an organic push and pull.

There's the concept of the MVP or Minimum Viable Product. It's a bad word at Amazon. If anyone says, "We have an MVP we want to show you," it's frowned upon. We can't ship something that's just viable anymore because the landscape is so competitive.

We say MRP or "Minimum Remarkable Product." It has to be memorable. It has to be lovable. Making it easy is good, but those are the table stakes now. Now it's about making this memorable.

When people think of Amazon, they usually think of customer experience. People talk about how they love Amazon. You put the logo in front of them and their mood and emotions change. I have to bottle that up. It's my job to push the PM and engineer to build the best thing possible.

Manifold: If you were giving advice to a team that didn't have the resources of Amazon, but wanted to hold themselves to the same standard remarkable products, are there principles or patterns that would help them figure out what qualifies as remarkable and not just simply viable?

Chris: It starts with your vision statement. Before you think about the tech or devices this is going to be on, you have to boil it down to a very simple straightforward statement.

At Amazon, before you can get any traction behind an idea, you have to be able to codify what you want to do for the customer in a one-page document. And from there. If it's got legs, that grows into a six-page document which goes more into the implementation how it might work. You have to be willing to change your vision based on what you learned in that process. If you never change your vision statement, then you're not really doing any work. You're just pushing your same vision forward.

When I worked in the concept lab we had five basic questions we asked:

  • Who is the customer? If you don't know who you're building for and who you're not building for the vision statement is going to be way too broad.
  • What is the customer problem? You need to be able to prioritize? This is the person that I want to make a thing for. And this is the first thing that I'm going help them solve in their lives. There might be secondary and tertiary problems that they have. But you just have to focus on one when you start.
  • Is the most important customer benefit clear? Again you have to focus and make sure you understand which benefit you're focusing on.
  • Do you know what customers need or want? Have you done your research or have some data have some signals?
  • What does the experience look like? When you're starting out just because people are very visual, we started with prototyping first. Answering what the experience looks like does not mean it can't change. It actually should change.

It also doesn't mean throw some quick sketches together. Rather, it means thinking through the entire customer lifecycle. So when I am brand-new and when I'm onboarding, how do you talk to me? What's the messaging? Now I've been a customer for three months. That's a lot different ––what are the things that you're doing for me now? What about one year? Five years?

Manifold: Are those prototypes being created at the same time as you're fleshing out that document? Or did you need to have that six pager down first?

Chris: It depends on the team. What I saw was the tenure of the team matters. Newer teams, if you show them something shiny they love it. They'll go right for it. But someone tenured at Amazon is going to ask a very Amazonian question. "What are the dogs not barking?" You've told me all the great things about this product, but what are the risks? What is your contingency plan?

Manifold: The press release and the FAQ document are probably the things that Amazon most famous for in terms of their product development process. What other aspects of the process are unique relative to what you've seen in other organizations?

Chris: There's not a fear. We know this is really hard work, and we know it's not going to always be right. We just want people to push everything as far as they can, and know that if this whole thing blows up, it's not going to reflect badly on you.

The other thing is going back to the big bold vision statements. My favorite vision statement is for the original Kindle. "Every book ever printed in any language, all available in less than 60 seconds." Most companies wouldn't be that bold.

Manifold: One of the things I know you did while you were in the concept lab was the Apple Watch app. What that process was like?

Chris: Imagine if you can before the time before everyone had a smartwatch. We didn't even know if people would care that they could use the app on their watch. We didn't even know if watches were going to last or if it was going to be a fad, or where it was on the hype cycle.But again, this being Amazon, they said who cares if this device doesn't go anywhere. If it does, we want to make that bet.

One thing I learned was the importance of function-driven design. Being as concise as possible while trying to make an experience make sense on the wearer. With Amazon, you can theoretically buy millions of products. You could buy an Xbox on the watch. You could buy a bicycle on the watch. What's the catalog that we should put on this device?

I started with research, and I went out to customers homes and showed them storyboards. I tried to figure out how customers would envision using this device, both early adopters and people who didn't have watches yet.

I did tons of storyboard sketches for different user stories that we imagined might be successful. Let's say you go into a grocery store and you have a list and you just want to know where in that store your items are. I made a storyboard for that scenario. When you are at home, or you're in the car and you need to do a task, what types of tasks would you want to do, both shopping and non-shopping related?

That's how we decided on our initial set of use cases are initial catalog. And then from the design process from there is actually very simple because Apple does a very good job with all their documentation. The bulk of it is revising copy, because most of that experience is just text.

Manifold: If you were to go work for a company that doesn't have the reputation for Innovation that Amazon does and you were tasked with trying to help them make things happen, how would you how would you try to go about doing that?

Chris: Writing more documents. I had never written a document like these before I joined Amazon. Now it's second nature. The amount of crap you can cut through when you can just write your idea down and pass it around and hone the idea is fantastic.

I'd also focus on the culture of giving candid feedback in a way that's constructive and direct. Sometimes you'll get a PDF and you look through it and not really want to engage or give feedback or even look at it. But at Amazon it's your job to say help identify the holes in this argument, to help fill gaps.

One amazing part of this culture where I don't have to wait until my idea is refined enough to share. I can share a very very unrefined idea and know that I can share it with my inner circle and they'll fill the gaps and the more people that I share it with the more quickly I'll be able to rewrite my document and get it really crisp and ready for the next phase.

Manifold: Now you're at Twitch. What are you doing for them?

Chris: I've mostly been studying what makes Gamers different as an audience. I wanted to understand what makes gamers unique, and how do we design for them. I'm calling it "gameful design", which is different from gamification. We're not building for users or customers. We're building for "players."

It's important to make the experience feel more game like. It's a departure from function over form because with games, they reward free-spirited exploration. So how do we do that within the prime experience?

Manifold: The whole E-sports world has blown up. For people who are confused about it - watching other people play video games - how would you describe it to them? What is responsible for the rapid growth?

Chris: I think it's the age of the population. If you were to ask someone if they played football as a kid or played sports, that's probably part of the appeal of watching sports. Now you have people who have grown up playing more video games than sports. It's that same sort of nostalgia. So just like you'd go to a Super Bowl party, it's like that. I also think that with E-sports, there's a level of polish. The graphics are amazing and they look beautiful, and there's an extra layer of imagination. In games, you could be on any map in any world anywhere and it's just beautiful and imaginative and immersive and even if you don't even understand everything that's going on, the commentators are there bringing you along, making you feel excited.

Manifold: Facebook and other sites have gotten a fair amount of bad press in the last year around people becoming addicted to their devices, and how some of these sites are sort of complicit in that. And they're leveraging a lot of the same kinds of game mechanics that video games have gotten really good at. When you're designing an interface that is not a game, and you have the opportunity to leverage similar types of mechanics to build a habit loop or tap into a psychological trigger, how do you think about that? What do you think is responsible use of those mechanics?

Chris: This is an awesome question, because from my point of view as a designer I have this intrinsic motivation to make things to the best of my ability. I want to make things that people want to use. And in order to tap into those emotions, it's sort of in my duty to use those loops and in some way shape or form at some point.

So how do you use them responsibly? There's one worldview that says people are going to be addicted to some product's feedback loop. It might as well be mine. But that's not ethical necessarily.

There's also this issue around privacy and data. Right now the dominant opinion is it should be easy for me to opt into license agreements and make it as easy as possible. But I'm not sure if they should be designed that way. It should be a difficult process and the fact that. I should know exactly what I'm opting into, especially as we come upon this age of machine learning and data democratization, where our data is now being used in many more ways than we ever thought possible when we opted in.

Manifold: Any other technologies you're particularly excited about?

Chris: Machine learning and AI. I want better design tools. Autodesk has an amazing tool where, if I want to design a motorcycle and I want it to be optimized for speed not durability, and these are the materials I want to use and these are my constraints, it can recommend all the ways I can build this and now I'm curating. It dramatically speeds up the design process.

Manifold: Are you in the camp that says that you can't automate creativity. Or are you in the camp that says this is going to take my job away from me in 10 years?

Chris: You can't automate creativity. Creativity comes from the juxtaposition of two opposing unrelated ideas. Machine learning works by teaching it all the things that are related to each other. Here's all the trees. It doesn't do a great job of creating new or interesting or creative things. I can't say "make this bike pretty" and have it know what that means. So I think there's going to still be the human element of creativity and creating things that are emotional.