The Future Of UX & Design

Dorothy Ryan, Jonathan Courtney, Michael Fitzgerald at UXDX EMEA. Video: https://youtu.be/CjxoUNo3BvE

Readable transcript: edited from the recording's captions for readability (fillers and false starts removed, punctuation and section headings added). Wording is the speaker's own. Timestamps are positions in the video. Names marked [?] could not be verified against the audio.

Panel introductions and where UX stands now

[00:00:06] Host: The guys have just joined us. Michael, Dorothy and John, maybe you could give us a little bit of background. Maybe we'll start with you there, Michael.

[00:00:18] Michael: Hi, my name is Mike Fitzgerald, from Interplay. Interplay is basically a tool that helps designers design with real code in the browser, so they can do realistic testing with real data, and responsive testing as well.

[00:00:32] Dorothy: Hello, I'm Dorothy Ryan. I work for a company called Asavie, and we make connectivity simple. We basically have a software platform for connecting mobile devices and the Internet of Things.

[00:00:48] John: Hi, I'm John Moriarty. I'm part of Fjord's Dublin team, a relatively new Dublin team. We're part of Accenture's new innovation center based in the Docklands. I'm an interaction design lead.

[00:01:01] Host: That's great. I suppose what we'll talk about first is the current situation for UX design. Jonathan, I'm going to ask you a question. You originally said that when you went into clients, you often had to explain what UX design is and why it's valuable. Where are we now? Is this working?

[00:01:28] Jonathan: Yeah. I think there's not a lot of having to explain anymore. I would say the majority of clients we go into now, the majority of companies, don't need convincing that design is something that they need, or that innovation is something that they need. It's just starting to get messy again, with a lot of different types of processes and terms. So it's all about trying to figure out what the problem is that they really want to solve. From their perspective, they usually come to us and say, "Hey, we need design sprints," or "We need UX," or "We need innovation, we need seven innovations right now." So we're trying to figure out, from that blob of things they're talking to us about, what it is that we could actually help with, and what the different services are that they actually need. Do they actually need UX design? Do they actually need these different things? So now it's not about convincing them to get in the door in the first place; it's about trying to focus them on the specific things that they actually might need.

[00:02:36] Host: Figuring out what the buzzwords really mean.

[00:02:39] Jonathan: Exactly, yeah.

Emerging technology at the Fjord innovation center

[00:02:39] Host: And maybe, John, I might ask you about innovation. Fjord has developed the center of innovation here on Hanover Quay. What sort of projects are you looking at today that you think will become a reality in the near future?

[00:03:02] John: The innovation center in general is there to explore emerging technologies, and we're working with Accenture and Accenture's clients from all over the world. The major domains we're looking at are blockchain, cybersecurity and artificial intelligence. We've been working with Microsoft to create an identity system for refugees based on blockchain, using a combination of blockchain and biometric data. That's something that was recently presented at the UN, and hopefully it's going to launch soon. That's one of the near-term things. A lot of the things we're looking at, though, are much further out. We're helping clients understand what these new technologies will mean for them and setting the roadmaps for them for the next few years.

Building a voice UX team at Alexa

[00:04:12] Host: Okay, that's interesting. It probably leads well into the whole idea of voice, and where voice fits in with it all. One of the questions I'd have is about the UX team in voice: how is it different from a traditional team that you might have, an interface design team or UX team?

[00:04:34] Noelle: At Alexa, we probably started this process, I don't know, five years ago, trying to think about how things are different for voice. We did not have a whole bunch of voice UI experts. Whether they existed or not, we didn't have them in-house. However, what we did have was a bunch of App Store people who had built apps for Kindle, who had built apps for Fire TV, and so we really looked at ways to reskill those people. I was actually at a great CTO luncheon, so nice to meet you guys, that was awesome, and they were talking about the cloud journey and reskilling your staff for the cloud versus old-school infrastructure. It was very similar for voice UI. We ended up making a dedicated effort to not just say, "Well, you're not an expert in voice, therefore we can't leverage your skills." We looked for artists, creatives, people that could show aptitude, and really developed that skill set. Today they're experts in the field, but they've only been doing it three years, and they started not knowing anything about it.

[00:05:40] So I think it's a lot about finding the people with the right aptitude, the right culture and level of creativity, and then really giving them a space to learn and be curious about this. We always say it's day one at Amazon; it's day one at Alexa, because we're just at the beginning. So give people a chance now, and within a few years they'll be your rock stars.

[00:05:58] Host: Do you think it's similar to what Jonathan was saying earlier, in that what you're looking for are problem solvers, people who can adapt themselves to solving the problems of working with new technologies and making sure that they work for users?

[00:06:17] Noelle: Yeah. We actually have 14 leadership principles that we model ourselves after, and I won't give them all to you, but they're all awesome. All of them are important to us, and we actually interview just for these. Functionally, if your resume says it, we'll ask you a couple of questions, great. But really, invent and simplify is one of them. We want to know that you can invent. We're not asking you to invent the Fire Phone; we're asking what you invent on a daily basis, something as simple as a script that means you don't have to do something manually anymore. At Amazon we expect people to invent every single day, even if it's a small, tiny thing or a really big thing. So we're looking for problem solvers, but those that can invent, that can innovate on behalf of a customer, think through what a customer wants to do: not just what they're telling us, but what we think they might tell us a year from now, three years from now.

[00:07:09] Host: When we start looking at things like future trends, and we'll get on to other technologies, but just sticking with voice, what do you think the future trends for voice are in the next five years, say?

[00:07:22] Noelle: Some of you were with me this morning. Our future trend at Amazon is really voice everywhere. I anticipate very quickly you'll start to see voice, and we already have a product, someone here has it, in the car: not just the way it is today, but in a way that you can speak to the cloud and get access to any information. Voice everywhere: in your kitchen, in the bathroom, in your workplace, being able to schedule a meeting with your voice, being able to navigate conference room scheduling with your voice. I know it sounds miraculous, but who knows, it might happen. Again, we're not trying to invent a whole new way of doing things. We're trying to look at what people do every single day and figure out a way to make it easier and voice-available. So there are years and years of that kind of work that we have to do.

Connected devices, data and design tools

[00:08:13] Host: Okay, maybe I'll turn to Dorothy, because I know you've been working a lot in machine-to-machine product development. What impact do you think that will have on UX, even in the very near future?

[00:08:35] Dorothy: With technology, everything is connected. What we work on is the enabling of the connection piece. The basic user experience that people are looking for is that it works and it's secure. Having a secure connection is a really important part of smart homes, smart cities, smart cars and the industrial Internet of Things. There are so many amazing things that people can do, and some of them are quite boring really. You can connect bins to tell when your bins are full, so you know when they can be collected. Clock-in solutions for HR systems, things like wastewater pipes measuring the quality of our water. There are just so many things that come under the large umbrella of the Internet of Things.

[00:09:49] It's a huge stack, from the connection piece and your devices right up to your cloud and your analytics. So there's a user experience at every layer, and it's really important. We realize that it's not just about the security. The value is the data that you collect and how you manage that data, and user experience and design are really important when it comes to figuring out what to do with that and how we can innovate using it.

[00:10:26] Host: Michael, I might ask you: with all these new technologies and new ways of interacting, where do you think design tools will come into play?

[00:10:38] Michael: I think with the design tools, because of the different mediums we have to deal with now and the different interactions we can do, where before it was just different device sizes, now there are all these different interactions. I think we really need to have tools that will let you design in the medium that you're designing for. For too long we've been playing with images of things, whereas we should be playing with the real things, in the medium they're designed for. I think that's the only way forward, so that you get a real sense of how the user is interacting with the experience. Tools have helped us get through the development cycle so we can get that experience quicker, but I think we need to pull that back as far as the design tools, so we can see what it's really like when we're designing.

Designing conversations without a screen

[00:11:21] Host: Okay, and maybe I'll move on now to some of the challenges that we're facing, particularly with all of these new ways of interacting and new technologies. Maybe I'll ask Noelle. Obviously there are lots of complexities and challenges in designing a voice UI. One of the key challenges I see is presenting, say, a long list of options. When users are dealing with a visual UI, they can scan it and pick whatever they want. How do you deal with those kinds of things? I know you spoke about multimodal quite a bit earlier on.

[00:12:10] Noelle: We have to completely rethink how we're looking at UX when it comes to voice, because there is no screen. Even though we have devices with screens today at Amazon, we don't build for that. It's a huge challenge to satisfy this different way of thinking, because all of us think in screens, whether we like it or not. I describe things by going, "Oh, go here and then go here." With Alexa it's much more difficult. So we always encourage that the first thing you should do if you're thinking about building a skill is literally go into a room with a big whiteboard, or maybe a bunch of cards on a table, and you point to someone and say, "You're Alexa and you're the customer. Go. What are you going to do?" And you document that conversation. Then you have that conversation at least 10 more times with different types of people, with men, with women, with older people, with younger people, to get that huge level of diversity.

[00:13:13] Because it's just different. Before, we didn't care: anyone could push the button, anyone could drop down a list and find what they needed. Now I have to figure out not only what they want, but what they are going to say to represent what they want, which we call intent, which is not the same, and it's extremely difficult to model. It's a human thing. So I always tell people, "Yep, sorry, you're going to actually have to communicate in a room with your voice, because we're building voice stuff now." It's much different from what most people, especially UX designers, are familiar with. But the skills that do that single act of work at the beginning are exponentially more successful than the skills that are just coded and deployed to production. You'll know them when you experience those skills; they are very clearly not user experience driven.

Machine learning and new problem-solving

[00:14:05] Host: Maybe on machine learning, going back to yourself, Dorothy: how do you think that will influence the way we design?

[00:14:30] Dorothy: We have to gain from the data. When we have data that we didn't previously have, we can solve problems in different ways, in new ways; we can innovate. I suppose everybody's heard of how Viagra was discovered by accident, through looking at the trends in the data and the side effect that showed up in the trends. There are all sorts of areas like that where all this data exists. The problems are the same; we're just going to be able to leverage new methods of learning from them.

Who owns user experience in an organization

[00:15:14] Host: Maybe moving on from that, to look at the role of UX and design within organizations, and I'll open this out to anybody who'd like to answer. I think it's come up a lot over the day: how to integrate teams. There's been diversification within organizations, or distributing it around organizations. Who do you think will ultimately end up with ownership of user experience within organizations? Will it be a design function at all? Will it be a product manager function? Will it be the organization? Anybody have any thoughts on that?

[00:15:56] Dorothy: I think it requires the entire team to be thinking of the customer, the user. Business goals are important, but development and design are working together as a team. It's not just, "Oh, I'm the designer, I need to do this." It's how we can leverage everybody's skills in order to make the best possible choices for our customers. And I think once we have that thinking at a higher level, as John was saying earlier, at a strategy level, that's important: to try to get that level of thinking at the higher management level, so that it trickles down and everybody is singing off the same hymn sheet, everybody is working towards the same goal.

[00:16:52] John: Just to expand on that, I think like any good design, it's at its best when it gets out of the way. As designers, if we're going into organizations, helping them to set up the capabilities and helping our clients bring design into their organization, I think we're successful when we can take a step back and say yes. And if there's no clear owner, even better. If it's something that pervades the whole culture, something that's deeply embedded in a culture, and it's not really clear where design sits or where the role of creative problem-solving sits, that's a good place to be. I think that's the whole idea of the UX maturity model: if you become a mature organization, it's throughout the entire structure of the company.

[00:17:43] Noelle: I'll go. I just wanted to say that we were not perfect in the placement or creation of our UX team. It was not even really a priority. We were very focused on the product and getting it to our customers as quickly as possible, but it was never too late. So I just want to reiterate that no matter how advanced or mature your organization is, it's never too late to figure out where UX fits in. I do agree that pervasiveness would be ideal, but in our organization, someone literally had to go, "Okay, so who's designing this?" and have someone take ownership of that, and it really became a priority conversation in our organization and in our company. And again, we're a bunch of companies at Amazon. Retail has got that down, they do that all the time, but at Alexa we were an infant, a new startup if you will. So we really had to take a step back about a year into our massive growth and go, "All right, we need to do a better job at this." Our customers were great about telling us that we needed to do a better job. But it was never too late. We ended up hiring a new director, a role that didn't exist when we created the company, and we now have an entire team of engineers dedicated to that, and it's super important to us. I just wanted to say it's never too late to do that, and it's ridiculously important, at least to the work that we're doing.

Big data, hunches and design decisions

[00:19:11] Host: Generally, I do think it ties in very much with the whole strategy thing, and understanding business is the next stage for UXers. But what do you think the role of big data is, in terms of how people are designing and how it fits in with the organization?

[00:19:36] Jonathan: Why do I get that question? I think there's always this battle between quantitative data and qualitative data when people are building products, and I think it always really depends on what you're actually trying to solve and what you're actually trying to build. If you're looking at something like big data, you also need people to analyze it and understand it, and be able to take that and translate it into product terms or into user problems and all this sort of stuff. There was this thought that big data was going to be able to answer all these really big, important design questions. But when I'm in these larger organizations, or these large, famous tech companies that are always talking about big data, when you actually look at what products are getting out there and being successful, it's often based on a hunch, combined with what actually happens on the market, and then iteration.

[00:20:35] So maybe once companies start getting much, much better at not just collecting data and putting it into buckets and having it there, but once they can analyze it, find trends and be able to say, "Okay, we need these four products, we're missing these four products over here," maybe then it's going to change. But I think right now, today, it's still very messy, almost in a nice way. It's still very human, these decisions that are being made.

[00:21:06] Host: So do you think that data at the moment is just something to assess design decisions that you've made, and provide metrics and those beloved statistics that management teams love, to justify stuff?

[00:21:21] Jonathan: It's hard to say. Maybe you can answer that, but for us, I know data is often used as a metric to calm down the stakeholders, so they don't feel like it's such a hunch and that we're not going to market spending seven million on this thing that might not actually work out. It's also sometimes a way to get rid of the responsibility of the product team, because they can say, "Well, the data actually told us that we should launch this thing that has now lost three billion dollars, so let's just blame it on that." For me, I think it should be used as something that helps you. Right now, while we're still struggling to understand what users really want, and we still don't have a good way of predicting whether something is going to be successful or not, I think it should just be considered what it is: a help, not the thing that's making the decisions. Potentially, if used correctly, it could. Maybe five years from now everyone will be out of a job on every product level, because there'll just be the data, and there'll be an AI that says, "We need to make an app that's like Tinder, but when you swipe to the right, I don't know, a poo emoji comes up, and that's going to make us three billion." And then we just sit down watching these products coming out and using them. I think that's going to happen. I think you're going to say the same thing, right?

[00:22:46] Noelle: I concur. From a different perspective, at Amazon, if you look at our homepage, which you probably don't go to often, but if you went to Amazon.com's homepage, it's made up of about two hundred services, and each service is developed by a pizza team of about eight or nine engineers. But it's completely driven by data. For example, many of you have probably heard about the recommendations engine, where you go in and it's, "Oh look, your friends bought this." That's all data that we are mining and tuning algorithms for every single day. So it's a very specific thing. Anyone who's in product, who has product websites like that, can instantly use their data to determine what their customers might want. We are very clear about it at Amazon: we love your data, we love your shopping experience, because we use that to make a better experience for you, and our customers are now actually asking us to do that. It is a huge part of what makes that retail site work.

[00:23:52] So there are a lot of strategic discussions I'm in that mirror exactly what you said, and then there's the meat of retail organizations, which are using data every day to train AI models to build functionality that's readily available for customers. AI and machine learning and big data are such big, swathe-y words. It's useful to dive deeper into that and focus on how it's used in a niche, because it's hard to talk about at the industry level. You did a great job, because poo is a good reference. Poo emoji.

Wrangling data and using it for the ordinary

[00:24:31] John: It's interesting, because as an industry, our clients for the last few years have been asking about all these things. We find that a lot of what we're doing in Fjord and Accenture [?] is working with clients and companies to help them understand what to do with these big, massive droves of data. There are a few things we've found that are maybe worth being aware of for designers. On data-led design projects, we often spend up to eighty percent of our time just wrangling that data into shape, getting it into a usable format that you can start to extract insights from. It never arrives to us like that. The data is often seen as this magic tap that you just turn on and all these insights flow out, and that's not the case. So I think it's a misconception sometimes, and it oversimplifies the amount of work that goes into getting this big, messy tangle into some sort of shape that you can start to have a conversation about.

[00:25:42] Host: Is the question really whether we are collecting the right data at the outset?

[00:25:51] Dorothy: Just on that point: we tend to think the data's going to be so exciting, we're going to learn so much from it. Well, we can use it to support really boring things as well. We use machine learning to suggest what kind of IP address range we can give to our customers when they're setting up their Internet of Things projects and connectivity. It may seem like a small, insignificant thing, but it means that they can do something that's normally really, really hard.

[00:26:30] Michael: I actually think that some of the time we're thinking about the data from our customer experience: we can enhance the customer experience with the data. But there's also the designer experience. I think we can use a lot of data from our previous user tests and releases; we can learn from that. And we can eventually use machine learning and AI to predict how a user test is going to go on our designs before we test them. So the minute you finish your design, you can click submit and get some kind of indication of how it will perform before you do the real user test.

[00:27:05] Host: Is there even a case where, rather than using data to influence the design, the design should be to collect data and basically create more value within a product?

[00:27:21] Michael: Machine learning is just a tool. It's an extra asset for us. It's not something threatening that's going to take away our job; it's actually something that's going to help us do a better job.

Methodologies: AI as a teammate and design ops

[00:27:31] Host: Okay. The last thing I was going to talk about for the future of UX is the methodologies that we're using. We hear there are new processes, lean, agile, and I actually want to get John into this, because he knows quite a bit about different development processes. What methodologies do you think are surfacing that will further enhance the role of UX?

[00:28:03] John: There are a number of different buckets there, but just to look at artificial intelligence and machine learning, that's something we've been thinking about a bit as a way that's directly influencing our UX process at the moment. We found it really useful to think about the machine, the AI or whatever's behind a product or service you're developing, as a user itself, as a teammate. It's a really good analogy to think about how this function and capability will fit into whatever it is you're building. When you think about it as a teammate: what kind of qualities do you want in a teammate, and what qualities do you want it to bring? Humanizing it in that context helps you understand what role it might play. That even extends to things like giving the AI user stories as part of our roadmap and our planning process, where it's saying, "Okay, as an AI I want to do X so I can achieve Y." Slotting it into existing processes, we found, is a really useful way to make that work and make it fit in a context that makes sense.

[00:29:06] Michael: Definitely. I think in the last couple of years we've seen design ops. Design ops can really help streamline the process of development, from code directly to the user, as quickly as possible. But I think it's really going a bit further back, for some reason: DevOps stops at that design step, which I think is pretty ridiculous. I think the new tools coming out will help the designer actually release code directly to production. So it's not just a developer going straight through continuous integration, but designers going straight to production. Similarly, Salesforce is using tokens to allow designers to make little style changes, but I think we can enhance that and allow designers to do more stuff quicker, so you don't have the bottleneck of a developer in the way.

Design sprints versus design thinking

[00:29:48] Host: Okay. And Jonathan, I've heard you've worked a lot with Jake Knapp from Google Ventures on the design sprints. Do you think that these processes will influence design processes?

[00:30:06] Jonathan: I think that something like the design sprint, for anyone who doesn't know it, is just a five-day process that Google Ventures put together for quickly validating products. What's really interesting about it is that it's almost a cold, logical approach to design that at first feels super weird, feels very anti-design. But I think building these almost logical, cold systems around something that's extremely creative is a really great way to increase the productivity of design. One problem that I see, and I know this is going slightly off topic, but in terms of methodologies, is that companies try to integrate extremely difficult things like design thinking. It's like, "We're a giant old company with 30 billion in revenue, and we now want to get really good at design." They call in a company, and they try to integrate design thinking as an overarching thing: "We're going to be more creative. We're going to give everyone design thinking workshops, and then everyone's going to be more creative." But as soon as that's done, people don't really know where to start.

[00:31:23] They have all of these individual tools, they have the general theories: "Okay, we've got to be creative, we've got to make prototypes, we've got to mix our ideas and do backflips." But these companies are like, "Okay, but how do we actually design stuff?" And then they end up having to call agencies in again. I think things like the design sprint are a simple enough step-by-step recipe that product teams would be able to figure out how to integrate something like that much, much quicker and much more successfully than these broad, theory-heavy things like design thinking. So over the next few years, and I'm already seeing this in Silicon Valley as well, companies that were relying very heavily on very creative people, spending a lot of time trying to hire creative people and trying to poach creative leads from company to company, are now becoming a bit more self-sufficient by teaching simpler, more logical, more cold design methods that are almost producing creativity and producing product. I'm a big believer in the design sprint just from a cold, logical point of view. I also like design thinking, but I just really don't like it when companies say, "Yeah, we just did our design thinking workshop, so now we're going to be able to make really good products."

[00:32:43] Host: Do you think that with things like design sprints and agile, where they're very lean and focused, there is any danger that the overarching thing, whether it be a design construct or something that's actually holding all of these components in place, is somehow forgotten, and people start focusing on the individual components?

[00:33:14] Jonathan: As opposed to, you mean, one big design philosophy or something like that?

[00:33:18] Host: Or more, I was thinking, like a patchwork quilt, where you have a lot of individual innovations and focuses, and that would not only be in design, it could be in research, it could be in development, and then all of a sudden you're left with something that doesn't hang together.

[00:33:43] Jonathan: I totally understand what you mean. But there's almost no company that I know that even does design in a consistent way in the first place. So even just getting them to do one thing right would already be a big win for about 90% of these larger corporations. For me, I'm less worried about having a consistent user experience and design, the bigger-picture thing with writing the values on the toilet door and stuff like this. I'm more concerned about: let's prove that this one team can release a product faster than that team, and let's see if that starts spreading throughout the company. I'm sort of resigned to the fact that having a consistent design system throughout an entire company is kind of impossible. At almost every one of these large corporations, or even successful Silicon Valley-style companies, you talk to every single team and they have their own way of doing things. I think you just have to try to get your small wins. I used to really try to say, "We're going to change everything, we're going to get everyone on board, and we're all going to be doing this stuff," and now I'm like, no, not a chance. It's a bit of a negative answer, but it's the truth.

[00:34:50] Host: Great. Well, I think we've actually gone way over time for the day, so we'll probably finish up there. I'd like to thank Jonathan, Noelle, John, Dorothy and Michael. I'm terrible with names.