Last week I attended the Amuse Conference here in Budapest. Great lineup of talks (among them two I really admire, Giles Colborne and Steve Portigal) with topics ranging from case studies until the more philosophical talk. One the best talks was by somebody outside the UX community, Bertalan Meskó did intensive, energizing, inspiring session on the future of healthcare.
This year two topics stood out for me: design of AI powered systems, and designing for ethical issues. AI system design has been slowly getting more important with recent big new products like Cortana, Google Assistant or Amazon’s Echo and the rise of chat bots definitely put the spotlight on conversational design and new interactions around these systems. Ethical issues are always a returning topic on design conferences, but recently more and more cases are popping up, where design should make the difference, Sara Wachter-Boettcher’s talk highlighted quite a lot of these.
In the intersection of these two topics are things like algorithmic cruelty, but as Giles Colborne put it: these are really the cruelty of the design system. In the near future if the designers won’t care the user’s will suffer, this won’t be an exaggeration any more.
What an interesting and challenging time to be working in product design.
Here are my notes of the talks, and you can watch most of them on the Amuse’s Ustream channel in a few days.
Giles Colborne: UX vs Artificial Intelligence
- UXers don’t write algorithms, but algorithms are the next generation of added value — currently design adds the most value. New apps with xhat and conversational interactions build on NLP and AI, but it all happens on the backend.
- What will be the next generation? Ambient interface, all around people, the environment adapts to people’s needs, a kind of environmental computing.
- So will designers be replaced by data scientist?
- There are four categories where this change will have an impact. 1st category: AI shortcuts user input, like putting in calories based on image recognition. 2nd category: pattern finding in complex data, by adding the user’s context, the user’s own data it can anticipate user needs. 3rd category: personal assistants. 4th category: coordinating complex environments.
- You need data: you should build on unique data only available to you (weather vs user’s location)
- You need algorithms: UXers need to work with engineers to do experiments on what will be valuable for customers. But for this, they need to understand algorithms, need to talk to data scientists.
- But prediction never matches reality, you can choose between high bias or high variance, you have to know what is your acceptable level of error.
- We will probably need suggestion etiquette to avoid blunders, random correlation, anything that breaks social rules or produces algorithmic cruelty. Algorithmic cruelty is the cruelty of the design team. You need to understand the human context.
- This is the next usability challenge. You have to know what people would get upset about. Smart solutions need to listen and empathize with the users, this is an interaction that can be designed.
- Human conversation is based on the common context, persons try to guess the context based on theory of mind. You need to design this into the chat bots.
- So strategy for intelligent service design: data, user need, data visualisation, sense of etiquette, shared understanding, algorithm design.
Paul Adams: Why the Next Generation of Startups Won’t Build Apps
- The internet connects people and information together, eventually everything connects. But it’s still a really young technology compared to for example printing press. It’s also changing news, commerce, communication and how we move in physical space.
- Interfaces work between customers and businesses. Earlier this was like a market stall or a shop, and with internet came the web sites. These are all destinations, where people have to go. We describe these with things like traffic and visitors. The same goes with apps.
- People are connected, live to connect with others since we are social.
- Destination is tricky, after websites now we have app stores, but people don’t need destinations, so we need people oriented systems.
- App stores do have advantages, like discovery, payments, hardware, but these are eroding you don’t need them on the web and social platforms any more, instead of destinations you would want organic product use.
- Instead of apps we need to design systems of services. This different than IA, as the services don’t necessary have screens. UXers need to be system thinkers. For example you need to be able to draw Twitter without doing screens.
- Everything you do today will be different tomorrow.
Laurissa Wolfram-Hvass: The Audience and Usability of Research
- We understand that research is important, but communicating it? It’s hard.
- Research is learning, without it we are just guessing, throwing darts randomly. We can also fail with research, but practice makes proficient.
- The bigger you get, the more you learn, but also to harder it gets to communicate the research, to spread the vision of who your users are.
- Gathering info is half the battle, research becomes valuable, when it’s shared.
- The audience of research is the product team, goal should be to have influence with your research. So give information that matters to people who have questions. Be the person with the answers
- Even if the information is correct, it has to be digestible, suitable for everyone’s learning style. Different information is important for different people, like marketing, developers, designers, executives.
- Some of the useful formats to pack it up: immersive learning (taking part in sessions), usability lunches (everybody likes free food and they can watch sessions meanwhile), posters (personas, quotes, these spark interest), videos (of user visits), presentations (packaged research for like minded people), customer panels (users visiting onsite telling about their business needs), eating your own dogfood (have a resident superuser, like a diary study on steroids)
- Get as much mileage of your research as you can, by reusing a results in many different formats.
- Other people also did cool things: Mozilla personas, Facebook research museum, and live blogs of research trips.
Mike Atherton: UX is UI
- UX memes show what others think what we are doing. Some if it is true. This has an impact on hiring, UX/UI designer position mean companies are not sure what they need, and what they can get. Is it: UX is love, enhancement and magic?
- UX people are relationship engineers, people connecting people to people.
- How can be UX more valuable for business? That’s what product managers also do, but in an agile environment UX is hard, that’s why they need visual designers, the decisions are already made by the time the team starts on it, and it only need to look pretty.
- This is only making the stuff right, but what about making the right stuff? This is the shit flavored lollipop problem, no UX can fix a product like this.
- Value of UX must be bigger than the pain involving UX creates.
- The challenge is to find better ways to communicate the value of UX, especially on social media, better than simple memes.
Pamela Pavliscak: How to Design for Happiness
- Maybe emotion is not a problem to be solved. Happiness, optimizing for happiness is maybe not the key.
- Purpose and negative emotions are also part of the feelings.
- So going to happiness is maybe not through delight, as happiness means different things in different contexts, it has to be a balance between pleasure and purpose.
- This is the four zone model for happiness, personal and collective well-being: Purpose-Others: compassion (shared purpose, gratitude, generosity), Purpose-Self: transformation (golden stars, big picture, flow), Pleasure-Others: convinial (happy with others, like games, facetime, rituals, layered communication), Pleasure-Self: perceptive (sensory, opportunity to play).
- Pleasure and delight is in front of you, in you, around you.
- Purpose goes from details to big picture to really-really big picture.
- Principles to design for happiness: Bend time (to have mental whitespace), invite participation (discovery, synthesis, imagination), foster connections (intimate, tribal, textured), encourage generosity, cultivate creativity.
Janne Jul Jensen: Building a UX Department from Scratch: The LEGO UX Journey
- UX maturity models tell you where you are.
- Build competencies in hires and in internal people
- To integrate process in a process drive company, add yourself to the process steps.
- Creating a UX toolbox, enables projects to self help. Tools describe: how and when it should be used, what you get out, examples, dos and don’ts, illustrations, links to books etc.
- They also created a UX library with all the books they referred.
- Further vision is to have a UX academy, part of the onboarding of every employee, since everyone needs to own UX, and to build a UX mindset.
Steve Portigal: Doorbells, Danger, and Dead Batteries: User Research War Stories
- War stories, everybody has one. They are personal, how someone encounters a challenge Add your stories to your own inventory.
- But there is also survival bias, an optimistic view on how the world works, but sometimes it’s luck and not skill. Especially since you don’t see the failures. Let’s talk about those.
- Control is an illusion, as a researcher the things you don’t anticipate are sometimes the most valuable, but you don’t control the situations. It’s ok to walk away, call a time out, improvise, use the breaks that are given. The unexpected can reveal new truths.
- Those exasperating participants: they are all people, so might be unpleasant. Something might slip the recruiter or the screener, the user might be lying, but the researcher assumes honesty.
- A crisis in credibility: good recruiting is essential, the more you invest the more backup you need, but accept that failures can happen. If you cannot engage with the user, change the topic. Don’t blame if you meet a dishonest participant.
- We could define a danger scale for perils: mental discomfort, physical discomfort, potential danger, real danger. But our job is to get out of the comfort zone to learn.
- Don’t go alone, always have a partner. Let your company know where you are. Know your limits, you are also participating in the research, and not just observing. Debrief also your feelings, your reaction is also data.
- To research is to be human.
Russ Unger: The 3 Cs of Design: Charters, Critique, and Culture
- Had to suck to be a UX leader, to get here.
- Where is the manual for leading a design team? Just wing it.
- Good for people to leave the office to do weird stuff to bring back the weirdness to do awesome stuff.
- The team charter: it’s a unifying plan for rules of engagement, goals, objectives, kind of like a persona for your team. It needs to include focus areas, team purpose, commitment to each other, perception of the team, growth and improvement opportunities. Sign it! This starts defining the team’s culture.
- Critique: how to make this into culture? Have critique leads, having teams of critique buddies. They have weekly meetings, ad-hoc meetings, 1on1 critique meetings and team critiques. You can do critique on everything, like even writing emails
- Culture: lead by example.
- Don’t encourage heroic efforts, not fair for the others and yourself. Give time people to do weird things
- Allow people to raise their hand, and ask for help. You should fix things together.
- Sometimes the biggest thing you can give is free time.
- Give the permission to say “No”. The sentence for your team members to tell: “This is a great thing, but I’m working on an other nice thing, talk to my supervisor first.”
- Set up people for success.
- Hack the org chart and make people leaders.
- Mitigate design surprises. Big reveals are amateur, talk to your peers as soon as possible.
- Give the chance for people to be awesome.
- Slow down. And plan appropriately.
- Give it 5 minutes. To think and react.
- Will this matter in 5 years? Then let it go.
- Servant leadership means ok to let people graduate.
- Be a shit umbrella.
Adrian Zumbrunnen: Designing Decisions
- If you stop making design decisions, your users suffer.
- The way we ask questions, defines the answers we get.
- How can we make users do better decisions? And make them feel better on the choices they make?
- Number of options: there are too many, give people FOMO. How much choice is optimal? Generally less choice means more focus.That’s why when thinking on MVP you shouldn’t think in number of features, but in your product. It isn’t called the Minimum Viable Set of Features after all.
- Placebo choices: the sense of control change how people feel about waiting and choosing. The value of options is defined by the surroundings, and adding choices makes people less decisive.
- Moments matter: delightful moments relieve anxiety or potential bad moments. The right moment changes how people think about an interface.
- Friction: usually design is about removing friction to create flow, but this also comes with a cost as it turns off critical thinking, for example 1-click buying button. Friction can also improve UX, like when ATMs eject the card first. Adding intentional waiting times can show that the software is indeed working. Friction can make people more deliberate giving them the mental space to think.
- Default behaviors: sometimes it’s not obvious what’s default what’s an option.
- An interaction is like a sentence, it’s finished by a period, animations can help in this.
Basak Haznedaroglu: Crafting Future-Proof Experiences and Smart Interactions
- A future proof designer’s manifest.
- Speak the language, personality is the new UX. There are differences to what people do based on language and culture. Don Norman said: “Design is conversation between the designer and the user.” Interaction is dialog-driven.
- Craft the context: personalized contextual experiences.
- Be relevant: find genuine problems to solve
- Strive for effortless: user interface designers design the interface. User experience designers avoid the interface if possible.
- Think like an industrial designer
- When you create a product, also create the philosophy for yourself and the users also notice that.
Hilary Johnson: 3, 2, 1 Product Launch 🚀🚀🚀
- Working on the startup Impossible that is a marketplace for altruism, a collaboration between Pivotal and the startup. The goal was to have shared understanding, validate assumptions.
- Starting with a framing exercise, to understand the space, based on user research. They defined the product goals based on this. The metrics best describing the real world impact they want to achieve. They also wanted to shoot for native experience to meet the user expectations.
- Focusing on core features meant to: learn to say no, consider anti personas, have a creative outlet (enable creative flow by doing an icebox to park ideas).
- Agile transformation: rapid response, rapid feedback, have objective decisions based on user research. Agile rituals like retros, standups help in get things going.
- Choose a tech-stack means to get the tech team involved early, and this needs to fit the product goals. They chose Ionic that meant accelerated design and development.
- Building a brand design along the product design, especially if it’s key to the product, as research revealed. Test your brand, it’s a collaborative design process and make time to delight.
Samuel Hulick: Growing Your Userbase with Better Onboarding
- Start your design where your users start using. Like in videogames, in Super Mario Brothers the first level is perfect, while the end screen is lame.
- Don’t do tool tips in general. It blocks UI instead of letting it get used, throws everything at the user. Add one recommendation at a time. And just assume people don’t read tooltips.
- Metrics that matter: AARRR, especially for UX activation and retention. Churn lowers these. Activation: 15% conversation is good. Retention: shoot for 95% / month, less is showing problems.
- A place for genuine help is empty states.
- Progress feedback is great for showing onboard status. Trick: don’t start with zero: the endowed progress effect, by giving progress by default helps with the conversion.
- For set up use the built in interactions: like a todo list having onboarding todo elements.
- Product should be part of the product experience, onboarding helps the product work in the user’s life making it more awesome.
- People don’t buy products, but better versions of themselves. Onboarding needs to bring the users to that better version.
- Providing success states, congratulating on the progress made gives back.
- Teams missing on onboarding? It’s the Conway’s law: “products are organised how the teams producing it are organizes). Silos between marketing and development break onboarding. But for example UX and Growth team can spot problems in the funnel.
Sara Wachter-Boettcher: Design for Real Life
- Design is never neutral, since it’s made by people who embed their culture. So design for diversity, and you need to test with all types of people (like skin color).
- Everyone has blind spots, because we picture our audience in a very narrow way.
- Machines learn from us. We chose what we teach them.
- When Siri doesn’t understand “I was raped”, or makes jokes… These are not edge cases, something that doesn’t matter. These are stress cases.
- Real life is not always a friendly, positive experience. We may think of bots talking like humans as a fun or delightful experience. But humans are not always fun either.
Bertalan Mesko, MD, PhD: Science Fiction Helps Design the Future of Healthcare
- Science fiction provides the daily inspiration for the medical futurist. Good sci-fi pushes humanity forward. Sci-fi is real already.
- Now the real problem of healthcare is the lack of design. Two things are missing: hardware and software.
- Physicians are working with obsolete technology. They don’t adapt for fear of losing the human touch.
- There is too much medical knowledge already, it grows faster then exponential, you cannot read everything. IBM’s Watson can read them all and help in diagnosis, a human and a computer working together.
- The future of healthcare is: accessible, personalized, preventive, augmented, humanistic. Via 3d printed prosthetics, drones, VR, designed implants and drugs, and patient design. Is this an optimistic future? This future is already here.
- We need to add empathy to technology, but how to prepare for it?
- We decide how we use technology: it should extend and connect us. Also ethical issues to be considered: would you share your genetic data with your employer? With your government?
- Good healthcare design needs exponential thinking and have a feeling for other industries (and watch sci-fi movies).
- Healthcare is coming home, how to make the home products a system?
- Watching and reading sci-fi helps us discuss technological advancement before it happens.
- Patients will lead. They should all be hackers. Everything will be customized around them. But we need to do this for ourselves as patients, no one else will do it for us.
- Upgrade your health! Quantify your health and experiment how you feel better. Analyzing sleep patterns, scoring the daily health is easy already.
- Making information out of data is easy, machines can do it. But knowledge out of information needs to be designed, and with time knowledge will turn into wisdom.
Thanks for organizing Amuse 2016 Zsuzsanna Kovacs and Zoltan Kollin (and the whole team), see you next year!