Kathy Sierra – Opening Remarks at SXSW 2007 Part 2

[25:27] That’s how we let the user tell us; we have this little button that says “I’m confused”. We still don’t know what to do. So if we can’t just use FAQs, what do we do? Well, we can start by thinking like a human. This is what people were starting to do in the earliest days of interactive CD-ROMs; what can we replicate that’s sort of like the human experience; or at least, how can we minimize some of the downsides, and of course one of the biggest downsides is, it can’t recognize your facial expressions.

So, in real life, if you’re talking to someone, and they look confused, you do something — if you’re being kind, if you actually like that person — what would you do? What are the kinds of things you’d do if someone looks confused when you’re talking to them? Ask them. Ask them a question; because if they just say “I’m confused,” well, you might need to know more context. You might be able to tell the context already. But if you can’t, it doesn’t hurt to ask. You’re going to have to start to drill down and find out, “Alright, what are you confused about? Where did you get lost?”

So we need some kind of dialogue with the user. And again, they’re a research project. Some of you have probably worked on natural-language processing systems, so this is a field of artificial intelligence that’s been going on since the history of the world, and — it’s getting better, but we’re not there yet. However — I was in the field of interactive learning a long time ago, and this was really a interesting area of research, because we were trying to make what was called intelligent tutoring systems at the time; basically artificial intelligence meets CBP, computer-base learning; and we wanted the student to be able to say, “I don’t understand what happening, right here,” and for the system to be able to talk back and forth and have a dialogue — so there’s a lot of research on this. A lot of this is called explanatory dialogues; that’s one of the areas of research. I read this set of books on this academic project on carrying on this explanatory dialogue, and it was fascinating, and at the very end, what they said was — oh, actually we ran out of money, so we couldn’t this whole natural-image processing thing; so instead, we had all these transcripts of things that people had said, and we just said, “Well, maybe that’s captured everything, or most of everything, that people would actually want to say, and let’s just put them in menu items and lists, and let people choose, instead of having the real dialogue; instead of having them type in exactly what they want, and responding to them. And it turned out that they got about 80% — actually, 80% to 90% of the perceived benefit of what they wanted from that, for almost no cost — with just letting the user choose, but choose the right thing, not like what we put into FAQs.

So again, we think of Frequently Asked Questions as, “Well, isn’t that what they are? Frequenty Asked Questions?” But ususally, again, they’re frequently asked questions by the guy holding the tablet PC with the cute little shirt; not by the real person who’s actually struggling.

It turns out that we we can capture, and have a very short dialogue with our users, and go a long way toward helping them feel like the computer is actually perceived; what they were realy thinking. It could be anything — anything that you can use. In fact, in a second, I’ll tell you about — even if you can’t change your software, there’s still something that you can do with this.

So, here’s the big goal, of having WTFs instead of FAQs — and again, in addition to FAQs — is you want to get the user to the right contexts. When I brought up the confused person and you said, “Well, I’d ask a question.” If you can ask a question, you can get people to the right context. A lot of times, with FAQs and online Help, they don’t go backwards. They don’t say to the user, “Ooooooh, I see. You actually went too far. You’re not even in the right place.” They just assume that you landed in the right place and you know where you are, when a lot of times, you don’t even know how you got there, and you went too far. So, one of the goals is to find out if it needs to actually out you in a whole different place, and not have you have to know what that place, or I wouldn’t have chosen this in the first place. Then, give them understandaable questions that don’t assume all this prior terminology knowledge, or even conceptual knowledge. It should be at the level the user is at, at that moment.

Let them choose from a high-level statement. This is how you could start a dialogue. “I’m lost. How’d that happen? I don’t know what it’s called, but I need it.” If you could let the user choose one of these things — and again, this isn’t how you modify all of your help; this would be a special branch, because obviously, you’re not going to present big wizards and dialogue boxes for your advanced users, who are smiling and happy, and just want that little-bit-more-information for something they forgot. This is for the person who’s code blue, who’s more critical.

So, one thing you could do is just ask a couple questions; maybe even just one, and then narrow down the user choices, and then just present a big list. It could be that simple. It doesn’t have to be anything that’s really elaborate and sophisticated. So it is like way better context-sensitive help. A lot of times, the problem with context-sensitive help — especially if you look at things like Tooltips — it’s like, if I’m in code blue situation, I don’t care what the damn thing is called; and I don’t even really care what it’s used for. I care how it relates to the thing I actually think I’m trying to do, right now. So, most context-sensitive help is too tree-focused when it needs to be forest-focused. This can give you a way to have way better context-sensitive help, and if you don’t know the context because your software — that’s another question. Your software probably should know where the user is; but even if it doesn’t, it doesn’t matter, because you can still emulate context-sensitive help without having it actually know where they were, just by asking, just a simple question.

So, give the user a way to express himself.

This is not the guy. That’s just a gratuitous puppy photo. Those of you who have been to my other talks know why I put the gratuitous puppy photo — okay; something just happened to your brain.

Anyway, so, just for fun, I thought, “Well, why stop with WTF?” What other emotions might you be feeling when you’re interacting with a piece of software; or when your users are interacting with a piece of software?” I know you assume they’re all smiling, but — maybe instead of words, you could just have the face, and they just click on the face. Yes, that’s my face; and because all button have tooltips, — (laughter)

You know, this guy’s on a deadline. This one; I don’t know what he’s thinking, and I’m not sure I want to know; and I’m not sure I want to know how to respond. This is the most common one. Now, the problem with that one is he’s no really saying, “You suck!”, right? He’s feeling, “I suck.” That’s the big problem. He might be saying, “I hate you guys for what you’ve done.” He’s hating you because you’re making him feel like an idiot. So I took the trouble of modifying the Office Assistant in Microsoft Excel. What it does now, is it says what would you like to do, but this would be more appropriate: you just click on the button, and then hopefully, it would know what to respond.

So, even if you can’t modify your application, you can still do it with your documentation; and a lot of times, it’s just a matter of reorganizing it. This is our book series, and a lot of people have a lot of questions about what are the things that actually make these books work so well. Burt and I, when we designed this series, because we came from interactive software and artificial intelligence, we were horrified by the thought of trying to put a user experience in a flat, 2D, un-interactive book; and especially because we’ve come from the field of learning. It was, like, how are we going to compensate for all of the problems? And we came to the same conclusion: what is it a student can do in a classroom that they can’t do with the book? Well, they can look confused. And they can raise their hand and ask a question.

And a teacher, or a tutor, or a mentor, you know, a sidekick, whoever it would be, would do something with that information. They wouldn’t — well, the good teacher, anyway; some of them just keep on going — but if the person looks confused, the teacher will often ask a question; find out more about why they’re confused; and then they might just try it a different way. They give you another chance. Most of the online help and FAQs — it’s like you’ve got one shot. It’s like the computer saying, “I’m only going to say this once. One way, and if you don’t get it, that’s your problem.” And that the kind of thing we’re trying to solve.

So, this is just a book, but we put in context-sensitive — they’re similar to frequently asked questions; they might look like it on the surface, but they’re captured from real user questions when users are really confused. This is what you want people to say. We get e-mails like this all the time. They say, “Oh, this part was driving me crazy and I had this weird-ass question, and then — there was the question, right there. I didn’t have to go to the back of the book. I didn’t have to go to another section. It was right there, just in time, and at the moment, and the context, at which it was driving me nuts.” So you’ll know when it’s a little creepy, because they’re like, “Oooh, how did it know?” Just the littlest bit of customization can do that.

So, sometimes, it’s just a more practical grouping. A lot of you know; my big passion are my horses, and I’m in this horse home-study program; and one of the things they do, is they know that — when you’re working with your horse, you can’t be watching the DVD at the same time, or holding some big book; so they give you these little pocket guides. But they’re all for a specific context. When I’m out there, I’m just not, in general, working with the horse; I’m doing a specific thing. They put all the troubleshooting, the pitfalls, the tips — everything in that context, is in one place. If my horse turns around and, say, charges me, I’m not going to run back to the house and look it up. I need to know right then, why is he charging me — at this moment, when I was doing this particular thing. There might be fifteen different reasons, but this troubleshooting that they’ve listed in their book is relative to the thing I was just trying. They tell you how to work your way out of it right then. Again, this is crucial, because I might get pissed off, and never use your software again; but with a horse, I could die, so it’s kind of useful that it’s context-sensitive.

Most importantly, we don’t word things in FAQs and online Help. Something happens to people, and they stop being human when they write. Some of you have heard this story — I worked at Sun Microsystems, and so I like to make fun of them, even though I love them, and one of the things they did with the documentation is, for a while, their editorial department for the group that I was in, which was Sun Education; they didn’t want us to use contractions, because it would be difficult to localize, and some other total bullshit. So we said, “Wow! Okay. Who in Hollywood doesn’t use contractions?” Like in a movie — well, it’s on the screen, anyway — in a movie, when the director wants to communicate that someone is not human, even if they look human, you can always tell, because they don’t use contractions. That’s how you know; and that’s how we were communicating to our users, like we were not human. So, talk like a person.

Now, there have been some really useful studies about how tiny changes in wording, to use more conversational language in anything that you use to communicate with your users — anything at all — tiny changes in language to personalize it. Just using the word “you” makes a huge difference. Probably one of the biggest studies — so, between 20% to 46% more solutions to transfer problems, and what that means is, the person was able to transfer what they learned, to their real job, whatever they were trying to do with it; and they made only the subtlest changes. This wasn’t like we went from formal, arcane, academic language to slang. It was just slightly changing things; and the biggest change was using the word, “you.” It made a huge difference. We have think to about that.

One of the interesting things is, “Well, why do we get such profound benefits from slightly changing the language to make it more conversational?” One of the theories is that, when your brain is reading something that sounds conversational — it sounds like it’s talking to me — your brain doesn’t always know the difference between a real conversation — I should say, part of your brain — and a conversation that it’s only reading. Your brain, when it reads conversational language, it starts going, “Crap. I’m in a conversation. I gotta hold up my end. I better pay attention.” Just that little change, and all those little changes add up.

So, you want to talk like a real person — talk like Virginia, not Spock. Talk like Tantag. And, you get a choice. (laughter) Two different versions; but they’re both festive. So think about who you might talk — Tom? No. So again, don’t talk like a non-human. (laughter)

[39:34] And engineers are the worst of this. Part of my job was to extract forcibly knowledge from the heads of some of the engineers at Sun, and it was challenging; and if you saw them in action with their friends — I’m sure this has all happened to you, and you probably all do this — they would just suddenly grab a whiteboard marker or a pen and napkin and start writing, and suddenly, everything makes sense, and they come alive, and they explain it; they haven’t sucked all the life out of it. But then, when they go to write about it, they go into “writer mode.” Most of all, me included — we aren’t writers, and we just shouldn’t even try to be. So we should just be talking. Your editor’s going to clean up some of the — you don’t actually have to type “um” and “uh”, whatever else weird stuff that you would say — but just talk like a person. If you just took a transcript of what most of you would say to a user if they really asked the question, and you just slapped that up there, as messy as that might be, it would be more effective than what most online help and FAQs are written like today, because they’re not written for real people.

This is one thing I keep harping on, on my blog and everywhere I get a chance, which is, “Why do we treat the people who have already paid us, so much worse, than the people wh haven’t paid us?” So we put all this effort into making — the manuals are readable. The manuals assume I don’t know anything, and tell me wonderful things, and get me all excited, and make me feel comfortable, and try to help me understand that, yes, you’re going to be with me to get me through the “I suck” phase; and then we juts drop them down the mindshaft, once they’ve paid their — in this case — almost $2000 for that T200 camera. I know.

It’s about real people. So, none of this sounds sexy; if you’re talking about online documentation, most people never want to hear that the key to passionate users is actually just helping them learn. They don’t want to hear that. They want to assume it’s some big marketing thing. Now, the good news is, this is a hell of a lot cheaper, because again, even if you don’t change your product — if it’s a marketing question, you can outspend the big guys — although that’s starting to not work any more — or you can just outteach your users, so that your users get up that curve, into something they can feel passionate about.

Now, again, where there’s passion, there’s always a user kicking ass; or at least, a user really on his way, and knowing that he’s going to get there, and knowing that you’re going to be there to support him. Now, why? Why is passion so connected to being really good at something? Because being really good at something makes that thing a higher-resolution experience. For example, some of you may be going to the music festival. There may be some of you who are going to get laid or for beer. Some of you may actually understand something about the kind of music, and you may have some deep appreciation for some aspects of the music. You’ll hear different notes; you’ll hear more notes; you’ll hear things the rest of us don’t hear. I’m not a music expert, but I have a little bit of experience with mixing boards, so it kinda sucks, because I’ll go to a concert, and I’ll be like, “Oh, if I could just get my hands on those faders”– so it’s a little bit of a higher-resolution experience for me.

Okay, look at this text. There are a bunch of designers in here — or you claim to be. I know you probably know what both of those fonts are, and I think there’s even a web site dedicated to the eradication of the one on top; (laughter) which is what? [inaudible response] Excellent! Now — don’t say it yet — the one on the bottom I actually like, and then someone wrote to me and said, “Oh my God, that’s so 1968!”

So, wine — and a whole lot of you have seen me do this slide before. It’s my favorite one, just because I think it’s just bullshit. This is what people claim about wine. If you’ve seen the movie “Sideways,” — I don’t even know how to pronounce these things. The 1986 Chateau — thank you; see, these people — “has sumptuous aromas of cedar wood,” something in French, “wood smoke, and dried herbs with a subdued bouquet of minerals and celestial black currants; and, of course, the tannins suggest subtle nuances.” People actually say this, and believe that they perceive that. That’s a higher resolution; there’s a greater bit depth of wine for them. On the other hand, Robert Scoble says this: (laughter). It’s just a one-bit experience for him. Red, white, all the choices you have to make.

This picture — some people have a higher-resolution experience when they look at that — and I’m talking about resolution — so, Molly is a designer; she might say good lines, good negative space. She’s perceiving things about that picture. On the other hand, Eric might say, big JPEG. And Bill Gates might say, another customer. So, it depends on the resolution of experience that you have.

One more time, until you can all do this on the whiteboard. My last gratuitous coffee picture.

The way I like to end — same way I always end — is to talk about the importance of what we’re doing. We think that we’re writing silly little Web apps, or maybe really important world-changing Web apps, or games, or some kind of software; whatever we’re doing, but so often, we get bogged down sitting in the back room in our little cubicles or wherever; or maybe our cool little designer Airstream trailers, and we’re working, and we don’t think about the impact of our work. I know a lot of you are really savvy to the negative impacts of our work, but we don’t think about how much we’re changing people. Now there are people building websites that are going to change the world, but if you think about the experiences that you give a user; if you can make a user just have a slightly better experience, at something he’s working on; or something he’s having fun on; if you can make him a little bit better at a game; just have five extra minutes of free time because you sped up his work; and most importantly, if you gave him an opportunity to be in the “flow” state — I know most of you are familiar with that “flow” state — and the psychologists tell us that people report that when they’re in the state of flow, those are some of the happiest moments of their life; because all the troubles of the world, everything they’re worried about — it vanishes.

It drops away; because as all programmers know, when you’re in flow, you’re just, “Oh, man, I am one compile away.” Which is not necessarily true, but when you feel that way, nothing else in the world matters. That’s the kind of experience that we’re giving to people all the time; so, we don’t have to change the world one giant scale at a time. We can just change it one user, one little five more minutes of joy at a time, and I feel grateful that I get to do that every day, and so do you. I think we should all give ourselves a hand for helping people.

And thank you all for being here. This is great; I’m done.

Thank you. You guys are awesome.


2 Responses to “Kathy Sierra – Opening Remarks at SXSW 2007 Part 2”

  1. 1 kristytheron99 June 26, 2015 at 10:21 am

    Amazing! You provided sufficient information about the topic. I am sure people will praise you to do this commendable work.

    academic transcription services

  2. 2 Nirv May 11, 2017 at 3:29 am

    Hi, on occasion I get a 400 site message when I view this site. Just a heads up, cheers

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: