968 was merely example to design, which is what we advocate.
So that’s really how the interesting thing for user experience conference to be in this type of environment. In fact, many of the chairs were designed for this hotel has since gone on to become classics like you’re sitting on some of these famous chairs and then in the conference room after the famous chairs and the chairs, one of the most famous ones is actually even from the San Francisco airport today. Even though I have to say that I think it’s better than this hotel, which is where it was designed for, but definitely a famous site. I want to talk about two different sort of main themes today, which relates to user experience, that it is something that is not mantic. It’s something it’s systematic. Good user experience comes from following methodology, from knowing about concepts and theories and principles and not wishing for it and throwing things at the wall in those type of ideas. At the same time, user experience is not really driven by technology. It’s run by focus on humans. And I think it’s quite often in the trade press that trade shows in other conferences and ours. There’s a lot of focus on the technology. And I really want to drive back to the focus of the people on the humans, because that’s what a user experience is about. But there has been a lot of technology which so it’s a little later some examples. There have been a lot of iterations of technology of this interface technology that we have been through in my career in many, many different ones. And what usually happens when we get a new technology is that it’s actually used really cool. Even the first decides to come out are confusing and terrible. And so there’s really often new fancy means that step backwards and usability. But then again, if it in fact is a technology, then eventually we learn how to use it and we get two steps forward. So new technology usually equates to two steps forward and one step back. But if you do the math, it is one step forward and it’s a new outcome for the users. But I just really detest why do we have to always have that step back, though, every time we do something new? And I think we ought to be better by now. But it’s like the lesson from so many years that every time there’s something new, you know, we have to step back. So that’s kind of one thing I want to campaign against, but I can easily annoyed at something that really does annoy me. And so I think we need to rely we know things about about what makes things easy. And by way of analogy, I want to talk about another field, which is astronomy and talk about what causes the solar eclipse.
And so this is a guy who kind of discovered that. But before him, the thought was solar eclipse was a punishment from the gods. You know, your king had entered zoos and as we saw in zoos, come down to some over your country. And this honestly was a really plausible explanation, because if you think about it, those ancient kings can always give their lifestyle and always done something to any of the gods. And so there was a sort eclipse kind of thing about, yeah, you know, last week the king did this terrible thing. And so that’s why. So it was a plausible explanation. But of course, the wrong explanation and it’s a little bit the same theme, as you said, often we kind of end up getting the wrong explanations. Well, Stevenson says he was the first person to predict the solar eclipse phase. I understand the mechanics of the principles of what what what happens, which is that as the moon goes in front of the sun, it casts a shadow on the earth.
And once you know that, you can critize the next one, which is what he did and it happened. Right. And similarly with user experience, once we know the principles of what causes things to be so difficult, we have a better time predicting or driving the side in an appropriate direction.
And that is another analogy that kind of matches with the solar eclipses, which is that today the astronomers can predict like to the 7th when the next solar eclipse will be many, many years in advance. And of course, the sales of Velentzas and those precise big telescopes at. No telescope, in fact. So his prediction was more light on this day and there’ll be a solar eclipse which which which happened so very similar to that user experience, a day we don’t have the ability to predict to date single decimal points, like exactly how much your conversion rate will go up if you make a certain design change. We don’t have any accuracy that astronomers know do because that’s a more advanced field since they started 2500 years ago. And we only started maybe maybe 50 years ago. I mean, I wish it wasn’t the first about 50 years ago, maybe people started studying this field. So it’s very, very new in comparison to astronomy. And so today we have a very rough course understanding of how it works. And so we can’t predict that accurately. That doesn’t mean we know nothing. So we’re beyond that stage of thinking that zoos as angry at us if our design fails.
That’s not why we do so. I want to talk a little bit about some of these things that we know about, about certain types of things we know. And so, you know that many people, when they give these sort of keynote speeches, they like to illustrate their examples with photographs that they download from Flickr, that sort of common thing these days. And I just don’t like to be like all other speakers. So instead, I’m going to illustrate my examples with old 19th century paintings from Denmark.
And so here’s my first painting.
And one, does this show and return the question? Because obviously it shows a bunch of press. And this is partly because, you know, before the photography became common, it was a number one job of an artist to paint things to look like what they were like. If you’re going to hire Rembrandt to paint your uncle, it better look like you’re right. So similarly here, if you’re going to have a painter painting the painting of grapes, they should look likely to look delicious and do what I want to eat that you see. Great. And that’s what he did. But beyond that, one of the other reasons you immediately see what this is, is simplicity. He didn’t do a lot of other stuff. And I think a lot of websites in particular, but also some of the interfaces today are not like his painting.
They’re like an angry fruit salad, you know, all kinds of things just bunched together, mushed up and not necessarily as appetizing.
And this one single, you should wait. So that’s one of the types of things we know about research. There’s a really broad, broad principle of simplicity.
If we do fewer things, people on both sides of it, let’s understand the things that we do. And so that’s one of what I mean by saying we know things. We do know that simplicity works. Now, let me show another another example of the painting here. We have a nice group of farmyard animals, but know Philipson. And this is a naturalist painting, which means that if we look at a dog, it looks like a duck, or if you look at one of the cows, looks like a cow. So far, so good. But what if we put the duck next to the cow? Man, that’s a big dog attack of the killer ducks.
Know it’s a movie that has not been made yet, but that’s like the poster for that movie. Right.
And but no, that’s not actually the tag of the killing ducks, that it’s just physically, graphically bigger because in the foreground. Right. And so this is an example of a painfully technical repertoire, which means put an object in the foreground that pushes to the bank the rest of the picture that creates a sense of death, even though obviously a painting is a two dimensional object. We don’t know that, but it creates a sense of depth in the painting. And so this is similar to like a more specific design going forward as opposed to this broad principle of simplicity. So the equivalent web design would be things like if you’re going to do a search to make it a box and type in box, type in fields, and you put that search box in the upper right hand corner, it would be locally on the left hand corner. They follow a specific design guidelines. People will be better able to use your site, will focus more on your content, your product offerings, your message, and not be confused about how to operate this website. So now we know about this idea of putting something in the foreground, but that just put it duck in the forefront. A complete deficit painting now. So what is this group doing in the lower left hand of this picture? And that’s where Kazmaier is in the foreground that pushes the better pictures more of the back and creates a sense of depth, which you can see this from down. Exactly. From the beginning, though, it’s just a flat picture.
So, you know, these guys do work and they can use a lot of different context. Now, be a very significant naturalist, paintings where they try to give a Three-Dimensional. To satisfy this reckless logic now, this painting, on the other hand, doesn’t have that it doesn’t have that dark in the foreground. But if you look at this, we should learn a little bit. You’ll notice I say three stripes of color that’s painted, three stripes of color, three dimensional effect at all. It’s the sky, the land and the water. And yet we can interpret this repeatedly in the as naturally as the other paintings. We can tell the skies above the land and the detail of the waters in front of the land. But how can we tell that? Because honestly, if you look at the graphics, the water would be below the land and the sky could be next to the land or something like that. That could be next. The sky. Whatever the reason we can interpret this painting, despite the depth of these 3D effects, is because another U.S. principle, which is consistency of the sort of things we taste in the real world. And so you guys open up in nature. You’ve seen these fields and the sky above the land and things like that many times. And so, you know that when you see something like this, the sky is above the land, you know, that way up and the water is in front of the land this way. And so because of our expectations, as long as they follow our efforts felt like, you know, very modern painting. That’s weird. You know, we can immediately understand what what it shows here. We go back and will A be the oldest painting start being very realistic style. And so this is a painting from London. You can see the position, Paul’s Cathedral here, the monument to the Fire of London, and you can see the smack in the middle of the painting, the London Bridge. Now, those who live from London, they say now, well, that’s not how the London Bridge looks. And it’s it’s about how the London Bridge looks today. The bridge in this painting is today in Arizona, in United States as a tourist attraction. It was a great bridge to the U.S. They built up back up in Arizona. But back when the Grand Canyon was painted, this painting in 1874, the bridge was still there. And there’s been a London bridge to that exact spot since the days of the Roman Empire. So almost 2000 years has been London Bridge has been there in the same spot. And those two thousand years maybe had different sides of the bridge, many, many different engineering technologies used to construct ever bigger and more solid bridges for 2000 years, but always in the same place. And the reason for that is that that’s a place where you want to cross the river and that is a place we are different. There are no roads go through and all of that. So that’s convenient to have a bridge there. And this is my example of a more kind of broad principle, again, of user experience, which is the persistent of infrastructure and of social organization. And so if you’re designing something for this system, big organization like maybe a big city or a big company or something like that, you have to figure out a way that that organization works and its existing systems and infrastructures and way of way of working in the way of thinking, because if you’re trying to disrupt too much of that, you make quite likely fail. Or conversely, if you’re designing something completely new, we have to kind of be aware of our responsibility because we’re often creating some of those infrastructure and thoughts that could be there long after our specific design for individual screen design. They are long gone. But the infrastructure that the systems of the processes that we build up, they can be they can remain for a very long time, maybe not 2000 years for the case of your next project. But you never know. And and so think about it, this one that has studied these things more carefully when we do this type of design. Well, I want to talk about a number of different user interface technologies that we’ve been through. And they for follow this kind of lifecycle that I’ve put in here. And so we start from the very top and the circle is called the magic bullet.
That is that a new technology. People initially are very enthusiastic and thrilled and happy about this new technology being the solution to all our problems. And if only we could speak to the computer, there will be no usability problems or the only things by 20 or 30 years. If only we have the graph Windows graphical user interface, there would be no usability problems. And of course, we know this is truly the opposite is true, because when we move to a new user interface technology used with the first design for launch technology driven designs by Pulecio and Fouzia, I think about this new technology as they forget about the humans. And so often, I mean, the iPad was one of the worst examples of this. Almost all of the first launch apps for the iPad were completely wacky and weird and. People who would potentially people could use them and but this is true for so many other technologies as well. So we get test promising all the different. The sign of the things that happens is I call that the US panic, which is that this was supposed to be so good and so bad and I was wrong. And so now people like us come in to say, OK, research, why don’t we go to some studies and figure out and after that we move to the next stage, which is only now we know what not to do best and figure out what’s worse and what to do. And so we research no additional, more useful design options. All of that will increase relative to best practice guidelines.
And once we know how to decide this technology right now, we are here because this takes a few years to go through this cycle. But I want to be like it’s now it’s old technology, but at least we know how to do it, how to use it, how design for it, how to make goods for good products, use high usability, great user experience with that generation of user interface technology. And then so is the time to go through the cycle because it’s going to be some other new thing that comes out. And what I would really like is we like bypasses kind of right hand side of this child and not have to have all the bad design. Right. Why do we have to suffer that? But let’s talk about a few examples of some specific user interface technologies. So first of all, personal computer graphical user interface is there that last, last, last several stage of being old technology. But we really know how to do it. We have a lot of insight into how to make a good graphical user interface and how I make a good PC application. So they’ve been through that. There was a lot of bad applications in the old days, but a very confusing systems, but not we know how to do it for the Web. We are almost there. I don’t think we’re quite there yet, but we’ve that stage with the best practice guidelines have definitely been defined. We have thousands, thousands of Web design guidelines to tell us how to make a good website. But first, you need everybody to follow suit. And I felt accepted that the Web is an established, you know, design. So we still get some some video sites to come up from time to time. But maybe we do have the best practices known as a pretty solid take over into that established technology, agile development. On the other hand, we’re not there yet. We are at this stage where we know that if we just do it all the way, it was kind of condition for programmers. We will have a better user experience as a result of it. I mean, the original idea was great for programing things and for many things, but that for designing things and so that we know doesn’t work and we have found out why and how the bad things are kind of well documented. And we have some insight already now into how to do better. In fact, we had a one day seminar today about how to do better user experience that we do know a lot of good advice and how to do it right. But that said, I don’t think we have a to set me up yet. So I don’t think we can claim to say it’s a small problem, but we don’t do what we have. We know a lot about how to make it make it work, it seems like, who are designing for mobile devices.
We have been through that phase of having that design, finding out why those designs are bad, finding out why they don’t work. And we are now we have a number of good apps, a good mobile design to look at and we’re kind of getting into when we saw about getting away, did you say we know how to make mobile work?
And I don’t think we’re quite there yet. And one really kind of striking example of that is entire area of responsive design, which quite often means, you know, bad design because you can’t really have exactly the same on the big screen, small screen. So we have to figure out how to make the design kind of work on all different screen sizes. And that I mean, we can display all the different screen sizes. That’s a technology issue that that is no comfort to that, but could make a good user experience, a different screen size. It is a much harder question. And and so we sort of stage we’ve had to figure out what are the things that trip us up when doing responsive design. So that’s a little bit of an earlier stage. And so I will illustrate how maybe change my example with the of that whole painting and this guy who really love to be paint cowboy, you know, so here’s another painting by the same guy. And what’s the difference between these two paintings? Well, one thing you might notice in this painting is very long shadows. And so that might lead you to think that it was painted like sort of sometime in the evening, maybe like getting closer to sunset. That’s a because it seems reasonable, but that’s not really the big. Between these two paintings, you will never guess, because I’m only showing you photographs of these, and when you only see photographs, you don’t get what’s really the truth striking difference between these two paintings. You’re going to see them for real. What is shoes and what is small and size matters. And they’re really different experience wise to seem to be painting. But you see the small painting. So the Glass-Steagall was artificial intelligence. So that’s at that early stage. Now it’s being very hyped up right now and people are really thinking that that’s going to be our salvation for user experience. I think it could be like this painting of the Soviets. So one possible future is that it’s like a beautiful you enjoy your vacation and this woman is doing all it could in Brooklyn law in a city like it will compay. So Vesuvius has those two aspects to it. Right. And I think it is the same. It could definitely have some good things to it, but it could just as well create a lot of monster trouble for us. So what is the artificial intelligence interface? What is going to be the user experience on that? I think that sort two different ways of looking at it. So one is really, as I’ve been saying, it’s like the same as all the things I’ve just been talking about. Every other technology is like, OK, it’s people are very thrilled about it, but it’s not going to be as good and it’s going to be a lot of trouble early on. And I do believe that. I mean, I do believe that in that age, it’s like everything else because we’ve seen it so many times. It’s got history will repeat itself one more time also. But that’s another way of looking at it as well, which is this actually in certain ways, from qualitative ways different than the other user interface technologies that we have had. So I’d like to discuss discuss those differences. But let’s remember this kind of first point that people were like driven by technology associates and are kind of overhype things like this. And it’s our job to be the reality check and say, well, remember, the people like, OK, so let’s talk about I’ve got to start from the lowest level of interaction design, which is a basic level of the kind of words that we are communicating to the computer. And so the big difference between a guy and every previous user interface technology is that most of these cases with the recognition based user interfaces, by that I mean what by contrast, you were saying that this graphical user interface, if I click an icon, I think that icon is no debate about it. Well, I mean, I think that that’s the icon. I click on the story, but in a high, if I speak a word or if I had write something, if I make a gesture to one of these things, I mean, there is a debate about what I did with where did I say what role did I write? Which of the different command that I did? I mean, well, I mean, it would be, you know, this is going to interpret or recognize what I do, but it’s going to do it wrong some of the time was affecting an icon, I think that icon. And there’s no no discouragement. And so the fact that it’s going to be this recognition that we have to have a way of recovering from this recognition. So that adds a layer of complexity to the interaction design, which is right there at his advantage. Now, this will always be with us, but it’s certainly getting better and better and just getting better in a very astonishingly rapid pace. So by way of example, going to refer to Google. Now, you’ve got five times better in the last four years. I mean, this is five hundred or so four years ago that this recognized twenty five percent to one quarter of all words were recognized. We try to say something to it. And just a few weeks ago, they announced they were down to five percent misrecognition rate so that they cut the error rate by five by a factor of five. So in four years. So that’s hugely impressive. And I think it is easy to predict that that improvement will continue. That is exactly how fast, but it will therefore continue. So they’ll get it down to one percent wrong and then some years later, maybe even down to a tenth of a percent. So one out of every thousand words wrong. And then I think we can say, OK, it’s still not perfect and it’s pretty good to perfect. And beyond that, you know, computers can be better than reality. So it’s not just a matter of the computer being as good as a person at recognizing speech, but it can also do things like accommodate accents, which can be difficult for people know if you’re used to hearing the words pronounced a certain way and somebody pronounce it in a different way, then the difficult problem for a person. But the computer can just be taught to recognize all of these different accents and one day to teach them how to recognize the things that so so definitely there the possibilities for improvements here and things like a bit of a reality. So I think it’s not. All that bad, but let’s look at the next level, which is a sense, I suppose I can combine my words into an actual command structure to the computer and the question of what can I say? When can I say? And here we have certainly already we could see some issues in the of many of his current systems have very awkward ways. The phrase sanctions that I work on, its kind of rules as opposed to human rules. So you’d have to say things like, OK, hey, Siri with annoys me, you know, in a normal way of saying, hey, wife, if you want to talk to my.
So let’s not do that. So that’s really very, very important.
And even if we get over that, those type of annoyances, there’s also the deeper problem of reformulation, which is that if I state my problem in a certain way to the computer, it’s not always the right way. And sometimes I want to get alternatives, alternative variants of what I said. And that’s something affected. Quite the most important graphical user interface is because a computer can show me not just you want to ask for this, but here are some, you know, related items that you might be interested in as well.
And this is really well suited for scanning over a two dimensional I think, look at a big screen can really do a lot of great work there in and you know, it’s hugely improved conversion rates and sales, by the way. So so that’s good.
Whereas in speech, linear user interface, it becomes very awkward to do to present these kind of charges. So is one of one thing that that’s an issue here, and it’s that people have to try back to how the computer needs to work, which is contrary to our basic ideology. I mean, if there’s any one ideology that we have in users and the and said the technology should adapt to the people and the people should be like they are and be left look left enough to live their lives. So that’s very, very unfortunate.
There’s also right now a much kind of initiatives to put a lot of these speech recognition systems into cars and put apps into cars and do, you know, manage your calendar and things like that while you’re driving until you smack right into something, because this is there’s a huge risk of driver distraction for those type of interfaces. And I think it could they could easily end up killing large numbers of people.
And you know, that that’s one of the biggest causes of traffic accidents these days, is people being out there on their mobile phones. And when I say people, I don’t mean like you holding your phone handset. Handsfree is just as dangerous because it’s not a problem of whether your hand is on the phone or the steering because you have two hands. Right. But what the problem is, is the cognitive lower cognitive distracted. I talk to somebody who’s not the car. My mind is like halfway out of the car and your reaction time goes up and your accident rate goes up. And it’s very it’s about as dangerous to have a phone call as it is to be drunk when you’re driving. And if that they to talk to a person. Just imagine how dangerous it is to try to manage your calendar. I mean, the car and all of that is going to be really, really nasty. So that’s those kind of issues with that. And, you know, this is generating a lot of talk right now about conversational interfaces.
And it would be dangerous, of course, but even for other systems, I don’t think it necessarily is going to be that good. So one of the reasons that so much talk about that is this weekend. So we weekend is the most successful system in China these days as China is the world’s number one, you know, online market.
Of course, everybody else is very excited about what works in China. And they want to learn from the China, from the Chinese and everybody. So that makes all sense. But they’re learning the wrong lesson from Wiecek because we can check for historical reasons, because a long time ago, that’s what they do. But today, that’s not why people in China use WeChat. So we did some research in China with Chinese users on how they use we tend to think about WeChat and so forth. So this is just one example of a mental model that we had the users cross in one person’s mental capacity.
The IS colleague, there’s a peripheral part of this mental model and was most central, not just for this person but but for a lot of users was much more central to these things, like bills to scan QR codes, like a version of bar codes, I guess some kind of integration with the physical world with a really good thing that the wheelchair has achieved, but also integration abilities that they have in which they very simple payment across different physical world and online world and person to person payment.
All these things are very smooth and nicely integrated in wheelchair. So integration of useful facilities, integration with the physical world, those are reasons people use WeChat and the way they use it is quite often by using made use of graphical user interfaces, very rarely actually by using chat.
So it’s a misconception with this idea that we’ve had is that so conversational user interface, they open up one level to the task level, which is like I’m trying to get something done with the computer. There’s a big actually big difference here between A.I. and traditional systems because of different computer systems. They do a zip code and this is both good and bad. I mean, it’s good that the computer doesn’t have control, but it’s bad because what I tell us to do is not always what I want to do. And that’s the problem for that. People don’t really understand the computer facility. And so forth. So we know that people will ask the wrong thing quite often, but that said, it’s still kind of a comforting feeling. The computer does and is told in countries with a high, the computer does decreases or just thinks would be good.
So the computer, it will get agency in these eyes, the river systems, and that wasted time. You have aspect of user experience that we have otherwise been beaten so much into. And I mean, some of those are several examples of this that we always see now as like these examples of if you say somebody was on Amazon.com, they were searching for things like like pregnancy books and baby books and stuff like that. And, you know, like a year later, they still get pregnancy books and recommend. And I think they have to fix that one by now. But that was for a while and they the situation. So the thing is that the easiest means understand the bigger context of news, not just for the specific simple things, like if I buy one book, I will buy more similar books, but maybe not for for certain types of problems, but certain types of issues. So let me show you the painting here. This one looks like it’s a ship in the middle of the city very often doesn’t make sense. No, that’s not actually what it is. I believe we recognize each element in the painting. We can recognize a ship and we a recognize buildings. So our first thought is, man, a ship in the middle of downtown. No, what happened was that the good old days they used in these stocks where they would build ships at around those cops were warehouses. And so that’s what’s happening. So you just see it’s a ship that’s kind of being built. This is it. It’s almost finished, but in the process of being built and it’s of warehouses around it.
So understanding and context is crucial for interpreting this picture correctly, not just recognizing each element that will need to bear once Australia first order will be will be wrong, finally will aid the more kind of society. A really broad based question here. So I have a list of what today is the most famous design products.
And what characterizes this list is they’re all from huge I mean, truly huge companies. And that’s because it has turned out that the best way to do artificial intelligence is by machine learning from inventors datasets. I mean things that are so huge and immense. They are mainly collected by various privacy violations. But anyway, so that these big companies have them, but other people don’t have them. And so I think that is one of the really biggest challenges and problems with A.I. is if we’re going to be restricted to to be driven by these like a handful of huge corporations. Because if we think back to some of the previous revolutions and user interface technology, they’ve really had the opposite effect that the PC revolution makes, that now thousands of companies could create software and could make application and sell applications. In the past to the mainframe era, it was only a few companies like Big Iron type of that would make applications and the software that they did was terrible as a result. But the PC software, thousands of companies make software that they were able to do that by this new technology. And sure enough, maybe the PC applications had been designed there, but also maybe had good design and had a lot of interesting innovations in user interfaces and didn’t new things and empowered a lot of like average normal users to be able to use computers. The PC revolution was much a huge, huge advance and usability and similarly for the Web revolution, but even more so because now it’s not just thousands of companies that could create Keesey applications, but millions of companies could create websites. And basically any company in the world could just put a website up. And if they wanted to write, they could create a new design ideas, new features, new services, new products. They could do a lot of different things. And again, 100 per cent, so many bad websites, particularly the dot com bubble. There’s so many bad websites, but also so many good websites. And again, when you have millions being attempted, even as many as there’s going to be a lot left over there to go to. And so that was the beauty of those two revolutions. The USA of technology was that broadening the base and broadening of the number of things that were done and services offered and that just sort of generally speaking, diversity said that it would cater to do whatever you wanted. It wasn’t it wasn’t a that this is what you do know. That is really a big brother society. We be very easy to get into this ISIS that if that turns out to be the way the only way it’s done so. Concluding, I like to remind you, stage Wednesday and Wednesday is voted to stay or vote in the state and Odin was the king of the gods in the resolution of my last Thursday. And some serious things, just as tomorrow is Thursday, which is Thursday and the thunder of the day. But events go back to open. So Odin vs Odin say so. It was a hail of bullets, but he had many names.
He was also called the old father. Odin was also called the old one time because he only had one. Now, why did the king of the gods only have one time? Because he had sacrificed his other eye to gain wisdom. And this story, by the way, tells you that the Vikings are not against bloodthirsty modern television shows like to make them out. I mean, look at television. The Vikings are just like the killers and they remember the pillage and it just really nasty guys. But they actually respected Wisdom’s Alexander the King of the Gods had sacrifices to get one guy to gain wisdom. So wisdom is something that we in our field gain by upsetting the users.
That’s I think is a really core or thing. And user experience is you go back to the people, go back to humans or serve the people of Syria. They do that have use different Assai and see which works, which don’t work. So and that’s the way we gain our principles, both the very specific design principles and our raw raw interaction principles, as well as both of those to which we can be nice and cool is, you know, our wisdom.
UX wisdom comes from observing the users. And I think it’s kind of very happy that we can gain always by observing users, by looking at you guys. You do that and you don’t have to sacrifice your eye and you’ll be able to do so.