Designing a Unified Experience

[Music] Stanford University Welcome to the seminar on people computers and design I'm Terry winegrad computer science in the HCI

group and welcome to glad to welcome you back to a series which has now been going on for about uh 20 years actually and we've had a lot of really interesting speakers and general areas

of human computer interaction um quick administrative stuff if you want to take the course for credit uh you get credit for attending

all the sessions uh and if you are not a scpd student an off-campus student then you need to attend at least eight of those 10 in person uh which you will

verify by signing on a signup sheet this is ronita why don't you stand up because everybody looking for you after class who um every week there will be a signup sheet you check your name sign your name

in on the signup sheet uh it's got the list of people who are on access so if you haven't yet signed up on access just sign your name separately on it uh and at the end of the quarter if you've come to all the sessions you are done nothing

to do if you've done one or two by remote uh then you send a note certifying you've done that um uh and you get your credit we'll send we'll send out details at the end but that is

pretty much all there is and if you're scpd or specifically I've you contacted me about a conflict and we've dealt with it then you just send in a thing at the end saying you've watched them all

online uh there should be a link it works and then it quits working from time to time but if you go to the homepage for the course which is should I write it up uh I won bother HCI human

computer interaction. stanford.edu

computer interaction. stanford.edu seminar among other things you'll find a link to the administrative info which I've just said and a link to um the

online Stanford online uh Services where you can get the talks after they've been given okay so that should be all the other thing which some of you here recognize this is an open seminar so you

do not have to pay Stamford any tuition to come and sit in and listen or to watch it on the web so uh if you're if you're interested if you're not Stanford student and you're interested or you have friends who might be interested uh

welcome to everyone um think anything else administrative we need to cover I think that's it it's simple so uh we have a speaker each week

we've got all but one of them definitely lined up now and you can see the schedule if you go to the web page again hc. stanford.edu seminar there is a

hc. stanford.edu seminar there is a mailing list if you get yourself on that list there's a thing at the bottom of the page telling you how then you'll get the weekly announcement in email it's the same information that's on the web

page but it's nice sometimes to get an email reminder so uh you can sign up for that and

um the schedule is up I think is it Stephen next week I just realized I hadn't checked somebody here no Stephen's not here I believe next week's speaker is

Stephen da who is a postto here in the HCI group who's been doing work on experiments with prototyping if it isn't then he's two weeks from now and I'm forgetting who's next week but I think

he's the one who's next week this week we have Kim Goodwin and um we were just reminiscing I said oh Kim I had you come talk at a class about four or five years

ago and she said uh it was 11 so as you get older things compress right so uh I've known of her and her work for quite a long time she's been at

Cooper design which is uh again a company I've known well for a long time they've done really Innovative things in getting design at early phases user

oriented design at early phases uh into a variety of commercial products um Kim has done hundreds I was thousands or at least hundreds of product projects uh

over a number of years and combines what I think of as a kind of clear conceptual theoretical sense of what's important in design with a just huge wealth of

experience and wisdom that comes from having done it been in the trenches to put in those terms for all these years and usually the people who write textbooks and I won't name names but if you read HCI textbooks they're Ivory

Tower academics they're people who've thought about it read the research literature and so on but she's written a book which really does bring this this kind of wisdom from experience in so

this just came out this year um and I recommend it it's heavy but worth it uh called designing here can you here why don't you run

the wait you get that up designing for the digital age how to create human- centered products and services and the interesting thing about

it is it's not just about the sort of uh computer technical side of it it really deals with the whole process of how do you do design in the real world where you have products and projects and

people and a whole set of issues um so I asked her to come back talk about her wisdom in the field thanks Jerry thanks for inviting me and uh glad

to be back here at Stanford although I do have to preface my talk with a little go Bears um all right so I'm here to talk to you today about uh an area that's been of

interest to me for a number of years now which is designing a unified product experience um because we talk a lot about software as software and Industrial designers talk a lot about

Hardware uh but somehow the two have to come together to form a complete product experience now I don't know about you guys but uh in my lifetime at least

software has gone from being very character-based and and kind of clunky uh and yes this is from war games for those of you who recognize it uh to

being you know sort of graphical but pretty ugly still uh and then eventually to what we know today which is a very rich uh often beautiful sometimes still

ugly uh but visually very rich medium and so um over that period of time you know you would expect that there's been some evolution in the professions that deal with software and of course there

has right first there was uh programming and then eventually this interaction design thing came along and and these two disciplines are pretty different right the uh the programmers figure well

we do the hard stuff right we do the heavy duty technology and there's these kind of lightweight designer people telling us to uh to focus on humans and and bringing this design sensibility and

that was a little alien and to tell you the truth in some companies it still is a little bit alien although I think that that industry has made a lot of progress in this

respect and uh I think that designers uh have learned likewise from programmers uh and from HCI that rigor in design is a valuable thing right so uh interaction

design I think has become a little closer to programming over the years and I think programming has learned from from interaction design over the years and then there came the the even fluffier people all the graphic

designers what do we do with them uh right now that we have to incorporate color and and all this other stuff into our work and I don't know if you know many graphic designers but a lot of graphic designers talk in terms of

emotion and brand and these other terms that um are a little bit alien to some interaction designers and and even more so to many of the programmers that I know uh and so how do you get these

folks working together well I think over the years most uh most design firms at least have found good ways to do this and a lot of In-House teams have found good ways as well

um but these days of course computers are evolving even further right we have software that appears in our our C dashboards uh our telephones for better or worse sometimes worse are becoming

small computers um medical devices lots of these are computerized and if you get these wrong right you could do real damage to people our entertainment is digital right uh and all kinds of

everyday things that we use around the house are just getting so much more computerized and interactive so what is a computer anymore right it's it's not just one of these it's it's almost

anything we touch these days I even found recently that um there are digital showers I don't know about you guys I'm not prepared to deal with a computer interface first thing when I get up in

the morning but um uh they have them so all of these uh Hardware products are kind of interesting because you never hear somebody say yeah the interaction

design the software in that phone is great but the hardware is terrible or vice versa you don't hear that uh if one part of the product doesn't work very well people condemn the whole product it's kind of like a group project in

school right somebody screws up the whole team fails uh and it's interesting because in industry it's very common for companies

to divide hardware and software teams uh it's very common to be handed a physical device prototype and told okay here now make the software work one time I was

doing a uh a consumer glucose meter and somebody thought that it would be really simple if it only had one button because what looks simpler than one

button well think about it okay glucose meters do two things at at their simplest right they take a reading of what your glucose is today and they show you the history of what your glucose has

been so at a minimum you've got to be able to select and move back and forth that's really hard to do with one button so they had to redesign the the hardware once we started getting them to think

about the software uh and that would have been a really avoidable problem uh if people had integrated the two so what we have is a visual design discipline

that's largely rooted in two Dimensions uh industrial design that's largely rooted in three dimensions uh and then we have the fourth dimension of time and behavior which is the strength of interaction design and the other thing

about interaction design is it does overlap two-dimensional and three-dimensional form a little bit because the controls we interact with are are in those spaces now one of of the things that I

find really interesting about bringing the more traditional design disciplines of of visual and uh industrial design into a product development process is they introduce a new tension that I

don't think used to exist and that tension is really between structure and function and emotion uh and that's something that I don't think interaction design is all that good at uh as an

interaction designer for a long time I don't think that's a strength of my discipline uh and so I really appreciate it when you know I'm pushing for structure and form and somebody says

yeah but it's just kind of uninteresting right maybe it's very efficient but it lacks something um one of our our teams at

Cooper was doing a uh a museum kiosk at one point and had the interaction very clean very efficient and you know what the response was yeah that works really well but it's

kind of boring and so they intr produced a little bit of inefficiency into the behavior because it just made it more entertaining and the end result was

actually a design that people spent probably 20 minutes with at a sitting which is pretty remarkable for a museum kiosk so it's not always about efficiency when you think about the

average uh Enterprise app it is sure but consumer devices in particular there's got to be something else there because let's face it we don't make product

purchases just based on rationality Okay so so I live in a neighborhood where there's not a whole lot of parking so I have a really good rationalization for why I bought a mini and I can tell you that it gets

good gas mileage and it has plenty of cargo space and it's all very virtuous right but that's not why I bought it really if I'm honest with myself it I I bought it because I wanted it because it was cute uh and it was fun to drive and

that's really why I bought the car and I'm a very rational person okay so emotion is never really missing from the equation and so to get this unified experience we've got to have a unified

process that incorporates the best of all these disciplines uh and also incorporates the best of people outside of the design field one of the things

that I've learned over a number of years doing design Consulting is that design is only half of my job and that's not an exaggeration literally at least 50% of

the job is building consensus among a huge range of stakeholders and getting everybody moving the same direction uh doing upfront research to make sure we're designing the right

thing making sure everybody's got good process and decision-making tools one of my jobs is to enable the business decision makers to make good decisions right and make sure they have all the

the right information about you know what's going to be well received what's not and I can collaborate with Engineers to figure out well what's going to cost how much to build uh and we can take all those factors into

account and so those of you familiar with user Center design processes will find this pretty familiar um this is the the basic process that uh

uh I find is really successful in getting good design on a pretty quick timeline and building consensus along the way I think all three of those are essential in the real world uh first you

have to do a little project planning which I'm not going to spend time on today because it's kind of a boring topic uh there's a bit of upfront Research into uh what's going on with the business how are we going to make money doing this or or save money or

whatever the objective is uh and into who are the users and what is it they really need uh using that as a a starting point to figure out what should the product even be what is it

we're going to build uh before we start trying to build it then we turn that research data into some models that we can use within the design team and across the broader product team uh in

order to communicate about what we've learned in the research so we all get a shared uh language a shared set of tools to use when we're evaluating should we go this way or should we go that way uh

and then from there we figure out what are the requirements for this thing what does it have to be and do what does it have to accomplish and is there any of that that can wait for version two

right uh and then we actually get to the point where there's a solid design concept we call that framework definition uh in other words what's that initial sketch look like uh how many

screens does it have what are the hardware and software input mechanisms how do we display information how do we interact act with the thing at a high level but a sketchy level because that

starts to let us assess cost and feasibility and and lots of other things that are very important uh and it starts to give us something concrete to look at this is one of the the biggest challenges in industry is people spend a

long time on requirements and they thrash a lot because they don't sketch something early enough right they go straight from a list of here's what has to be in the product which might be a

page or it might be 200 pages of text uh and then they go straight to building it and usually if they spent a bit more time sketching very low Fidelity just doing really quick turnaround on things

they would get to a better answer and be able to build it a lot more efficiently and then there's a lot of work to be done in detailed design down to what every what color every pixel is and and uh what finishes on the hardware and so

forth uh and ultimately design is mostly done but stuff always comes up in engineering right no matter how closely you collaborate with Engineers as a designer there's something that shows up that was harder build than you thought

it was going to be and so something has to get tweaked a little bit along the way now um any of you who've taken classes with Terry I'm sure uh have heard about the value of iteration probably some of you uh in other classes

have as well um a lot of these steps are iterative okay I don't want you get to get the impression because I've displayed it linear uh in a linear fashion that it's a linear process it's

mostly linear but there's some recursive bits to it too so let's dig into each of these things so research um you know when I say research a lot of

times clients will say oh research we're not an academic institution we can't afford the time to do this research could be as simple as spending an afternoon watching some people use their

iPods and figuring out what's working and what's not working right it could be that cheap and easy or research could be a month of travel around the world

looking at uh how doctors uh in different places perform surgeries okay so um it could be a real range of things and it kind of depends on um how much certainty you need right if you're

betting a whole company on a product well you want to do a lot of research if it's a quick and dirty kind of uh rev to an existing product maybe a little bit will

do the uh the main tool that I find is useful in uh upfront research is based primarily on ethnography now that might sound a little bit odd because you probably

think of ethnography as the study of a culture right you're if you're going and um going to some foreign country and and sitting down with people whose culture is really different from your own think

about it though users might be pretty different from you they might have uh attitudes and beliefs and experiences and and training that's really different from yours um I think about this all the

time anytime I do a product involving uh kids because I just don't think like a kid anymore and their reality is very different from mine even though I was like them once it's a really hard term

remember that um likewise there may be a lot of uh assumptions that don't come to light based on uh pretty common methods that people use now what most companies do is

they go out and they say customer what do you want from our product well have you ever seen an episode of The Simpsons where a car company asks Homer to uh help them design a car the Homer mobile okay

that's what happens um Ford has an example called the Edo back in its history okay we don't go and say what do you want CU people don't actually know um Steve Jobs

says the worst thing you can do is listen to your customers and in a certain way he's right uh there's there's a way in which I think that's uh rhetoric but

um I don't even ask customers how they behave right if I say what do you do every day what what are you going to say I'm going to pick on Terry for a sec Terry what did you do this morning

before you came to work I had a meeting you had a before I came to work before you came to work I got up I showered I brush my teeth I discussed the day's plans with my wife

okay so you got up showered brush your teeth discuss the day's plans with your wife were you in the nude when you did that uh partly okay see he didn't

mention getting dressed so thank you Terry for indulging me um people always leave out the getting dressed part somehow um and

that's because when we do do something every day our habits are really ingrained we're unconscious of them and so we can't just ask people what they do we have to go watch them and we need to

go into the context where they would use the product and if we're inventing a product that doesn't exist yet we look at analogous behaviors right we look at what they're doing with paper and pencil and and face Toof face conversation

today and we figure out what's really going on there what are the fundamental behaviors we have to account for so just to give you an

example here's um here's a photo of uh a surgeon replacing a knee joint putting in an artificial joint so uh user

research is not always quite that um graphic but uh uh so this is one where if you've got an industrial designer and a visual designer and an interaction designer in the room they're going to

look for slightly different things okay so the interaction designer is going to say hm there's an electronic system there but the surgeon just ask somebody to bring in a piece of paper what is on

that piece of paper and why is it not on the computer right that's what's going through the interaction designer's mind because that's a huge issue okay the industrial designer meanwhile is looking at huh what's the angle of that display

is that really positioned right is that the right size and shape um is that ergonomically going to make sense uh and the visual designer is going oh man that thing is covered in plastic are you kidding I have to make the the controls

big and really high contrast then if they're going to show up okay so everybody's taking something a little different from the research so ideally they're all going to get a chance to observe people in

context okay the other thing that particularly Visual and Industrial designers are doing is they're looking for emotion and brand context right what are the kinds of things that people

relate to so if I'm designing a product for example uh for somebody who reads dwell shops at Crate and Barrel and uh has furniture from Ikea versus somebody

who shops at restoration Hardware reads Better Homes and G Gardens and and you know goes to William William Sonoma and so forth there's a very different aesthetic implied there right there's a

sort of different lifestyle that product is going to look different for those people right you might not know how but you can kind of see that that would be true um because something that fits into to this person's life aesthetically is

maybe not going to fit in so well there okay so we're going out and we're Gathering a bunch of data maybe for a day maybe for several weeks but we're doing this with a pretty small team cuz

having 20 people go out and do user research is a really expensive uh and B it's hard to spot patterns because patterns are really what matters we don't just want to say oh we saw this

one user doing this and designed the product based on that we want to say wow of the 20 users we looked at 10 of them worked this way and another seven thought in this totally

different manner how are we going to accommodate both of those big patterns in the design of this of this product okay so those are the kinds of things we're looking for and it's a lot easier to spot those if you've got a small team

doing the research so generally uh I find it's useful to send two or three people out to all of the product in uh uh all of the research versus having 20 people each do one or two interviews

each okay but you have to turn that data into something that the entire product team the product managers and the the salespeople and whoever else is involved in decision- making you've got to give

them some information that will help you make your case um I once observed a designer say to a room full of Engineers um here's my design isn't it

great and he had done all of these uh sort of semi-transparent layering things lots of animation and and this was a while back so it was really going to

kill system performance and and the engineer said well why did you do that why is that good and he kind of shrugged and said well cuz it looks cool and

that's not really going to be a very compelling reason right he couldn't convince everybody that his design was right because he didn't have any basis for an argument research doesn't just

help you know what to build and how people think it also helps you build credibility because you can say you know Mr CEO you're absolutely right I wouldn't do it that way either but all

these users we saw look look how they all behaved and that way it's not you saying I'm more credible than you are cuz somebody in the room is always going to outrank you

that's just a fact of of how we all work um but you have better data than they do and you can make a better case so you've got to develop some shared understanding here so interaction designers are

looking at all the data and thinking how do people work right what's their workflow today where are there holes in it where can I make it more streamlined where do they run into problems what information do they need as they make

decisions and so forth and they're not spending tons of time documenting this stuff because ideally when we design the product it's going to be better than the the current

workflow um to some extent interaction designers are also looking at people's mental models of the data Maybe doing some taxonomies of how people think

about things um maybe thinking about how they they conceive of the workflow right uh I don't know if you guys have ever seen the a video there's sort of a uh

famously bad usability test of um windzip anybody ever seen that video no well um they've improved on it since then but back in the day right the thing

you had to do with windzip is you'd open the program and you'd basically tell it to create a container and then you'd put your files in that container and then it

would compress right and every time I would go to use windzip it was kind of a moment because it just wasn't how I thought about it right cuz the way that we all think about compressing a file is

hey here's my file crunch it for me right uh and so it was a little bit backwards uh and so you watch this usability test and people

kind of get to this point and go what do I do because their mental model was different they just conceived of the process in a very different fashion right imagine if we all had to find our files based on where all their little

segments were stored across the dis as opposed to just going to one spot and grabbing them right probably the biggest uh amount of time that interaction designers spend at

this stage is in uh modeling the human behavior in the form of personas okay so a Persona is a uh a distillation of a behavior pattern and for any product generally you're going to find that

there are two or more personas people who behave in really different ways now these might be totally different demographics right you might have adults and kids making different different

personas um they might be related to market segments right small business large business people have different degrees of specialization in their tasks so maybe those are different Behavior

patterns or it might be around their mental model uh that knee surgery for example uh the design team found that um surgeons would approach the surgery in

two completely different ways you'd think orthopedic surgeons would all be approaching a surgery in a pretty similar fashion but there were two very distinct workflows just based on sort of a philosophy of of how they thought the

surgery was going to go better and so that led to two different personas okay so what we're doing is we're looking through our data and we're we're comparing behaviors so we're looking at

uh for example this is a data set from some uh financial advisers okay so for example some financial advisor spend a little time educate educating their

clients others spend much more um you know do they use pre-approved tools how often did they transact how big are their transactions we're looking at a lot of very objective

behaviors and and then we're looking for correlations and we're seeing huh people who did this almost always did that why is that right and H these people were

almost all the same in all these ways and so that distills into a behavior pattern so personas are distilled and derived from the data they're not just fictional people we make up because we

think it's fun it's about expressing that behavior pattern okay and so the the benefit to then turning that into a persona

is that uh it starts to take on a life of its own a little bit now you could just say to everybody on your team here are the behavior patterns we saw you could present them in sort of a dry

scientific fashion and everybody would look at them and nod their heads and probably forget about them but if instead you present those Behavior patterns as archetypal people as

characters you give them a photo and a name they start to become a little bit more real and it's interesting because I do find that people start to use the personas names when they refer to

certain kinds of behaviors and they start to empathize with the personas a little bit and we all know they're not real people we know they're fictitious but people start to get

excited about you know making aanda happy uh and they start to say well you know Pete and Marine just need really different things so you know how can we make both of them happy in this instance

and so it starts to get people excited and it really does a lot for building consensus and it does a lot for helping you prior prioritize features and functions and so on as well uh because you can say well here's our presumed

list of features but you know what our most important personas don't care about these five let's postpone them let's do them later because there's always some

trade-offs to make in the process and so um back when I started doing personas at Cooper 12 years ago they were pretty much behaviorally focused right it was

it was really about um what do they do all day right what's their basic process how do they think what are goals uh and so they were there was a little emotional context but they were

pretty functional in nature now as we've really Incorporated uh visual design and Industrial design and so forth um Persona started to get a little bit more

of that flavor in them as well and so uh it can be really useful to start incorporating a collage of what this persona's world is like you know what kinds of brands do they have an affinity

for and so forth uh starts to give you a little bit more complete picture and uh uh when it comes to Industrial design sometimes it's useful to incorporate some models of what what

would this persona's physical space look like right take a look at this uh this little sketch for example this was actually thrown together using

sketchup's little Free Library of uh um uh digital objects and so forth and of course Herman Miller being the smart folks that they are put a bunch of their Furniture in there but okay let's

imagine we're designing an entertainment system for this guy right what does his environment tell us well we can see he's into pretty clean modern look uh he's

into uh he's fairly knowledgeable about about design Classics and so forth he's got some some well-known designers in his furniture uh his his area there is

pretty sparse right um lots of cables and clutter and so forth are going to drive him crazy so can we solve those issues right it starts to help us understand what context will our physical device if we have one uh exist

in and so we're not trying to spend a lot of time on these tools generally we're spending just a few days putting together these models that encapsulate

our research in order to say to everybody here's what we learned here's what's compelling and believe it or not people really relate to these tools I've I've seen you know rooms full of

Engineers go oh what is it you're going to make me sit through these fluffy you know creative writing exercises Persona things you know there's a lot of

skepticism but once they see how we use the personas they start to get into it and they kind of adopt the personas and I've had development teams like print Persona t-shirts and play Persona

Jeopardy and do all kinds of crazy things that um actually are good for team building believe it or not I that's kind of a a soft topic but um uh it is actually pretty

important so one of the things that I think an emphasis on Visual and Industrial design brings to the the process is that an interaction designer is mostly going to focus on end goals

right what are the things the Persona wants to accomplish using our product and so they're going to be kind of functional whereas a visual designer and and to some extent an industrial

designer is also going to say but how do they want to feel using the product right what's their experience goal so think about um some of the interactions you have in your day-to-day life right um how do you want to feel when you go

to the bank right when you're entrusting your uh your life's earnings to somebody how do you want to feel about that experience

anybody safe right you want to trust this institution and so you're probably not going to have cartoon characters and purple and orange on that homepage are you right that doesn't make you feel

safe there's a reason most banks use lots of navy blue right there's a reason that they use very conservative type faces and so forth because they want you to feel like we've been here forever

we're going to be here forever we're not going to do crazy risky things with your money even though we know better at this point that they do um but their visual design is geared to make you feel a

certain way right hopefully their visual design is telling you the truth about who they are but uh um same thing with physical products right um when you when

you look at devices in a hospital for example what color are they what colors do you

see white beige beige teal maybe some teal or some blue right colors like green and blue make people

think clean they make people think of light occasionally red for danger right um but you're not generally going to see Hospital products that are brown because

that kind of makes you think dirt right um you're definitely going to see a certain feeling that people are trying to evoke right how about if you go to a spa what colors are you going to see

there you're going to see Browns and Golds and other kinds of warm things right you're going to see surface textures that are are softer and more soothing and so

forth okay so how we want to feel about a product is definitely incorporated into its visual language into the colors and materials and and other ways that we

portray it okay so we distill all of our data and we say hey guys here's what we learned so what does it mean what does this mean for the kind of product or or

experience that we need to build and so here's where we do some requirements now this terminology can be a little confusing CU sometimes when the designers are called in a product

manager has already written something called a PRD or mrd or some other acronym that means requirements document okay and so sometimes that's a one-page

thing that says if we build roughly this much uh this product at this price we'll make roughly this much money that's my favorite kind of requirements document um and then there's the 200 Page thing

that says you know item 13.1 12.5.2 a uh the list box will contain these things well how do you know the list box contains these things how do

you even know there's a list box have you drawn anything yet no okay and so sometimes requirements documents are right on sometimes they're arrived at by good process other

times they're unfortunate pieces of of documentation that you kind of want to set aside for a little bit uh and so uh it really depends on the process and the skills of the people putting those

together ideally our design research is informing those requirements okay because we learn a whole lot out there in the world that uh that maybe people haven't learned by looking at the

competitors or or asking the customers what they want okay and so we're at least going to um tweak an existing requirements document if it's there or

ideally the requirements document is going to start at this point um and so the most important tool I find in developing requirements is storytelling

okay uh some of us were talking at lunch earlier about how do you get from that blank whiteboard to something that you can look at and talk about something that's concrete and that's a tough

transition for a lot of people um it's hard to go from understanding the problem to solving it and so one of the most human tools we can draw on one of the most creative

tools we can draw on is storytelling think about your childhood for a minute okay reflect on how good you were at making up crazy stuff when you were little right that cardboard box was a

school bus one minute and a spaceship the next right you were very creative you were generative and that's an important part of the process it's not all of it but we've got to get to

something concrete and so if we say all right now that we know our characters right our personas and what motivates them let's put them in a situation and

imagine how that ideally unfolds right let's give them a magic black box and see what that experience ideally would look like from their point of view okay so here's the part of the process

where sometimes people get uncomfortable because they say how do you know you're making stuff up well okay yes you are making stuff up and that's what you do

when you create on the other hand how do you know what kind of birthday present your mom would like how do you know um that your girlfriend's going to be

mad at you when uh when you didn't meet her on time for a date right how do you know these things because you know what those people care about and you know

what they like uh and you know how they think and behave and that's exactly what we're trying to do with our users is use that human understanding and trust that part of ourselves uh in order to figure out well here's what we think they'd

like that doesn't mean we won't do some usability testing and stuff later on but we're using it to generate and so generally we're going to start out with some small number maybe half a dozen maybe a dozen scenarios in other words

we're going to say let's put our personas in this kind of common situation and imagine how that could unfold what's that ideal experience look like and from that we're going to generate a whole bunch of different

proposed requirements that we can all then talk about and figure out can we accomplish those things okay now from that emotional side though

storytelling doesn't quite cut it okay so one of the hardest things any designer will tell you is uh the client who says I don't know what I want it to

look like but I'll I'll know it when I see it right um visual design and and Aesthetics are incredibly subjective right everybody's got an opinion uh they

they can't tell you why they like it or don't like it um but whether a CEO likes it often isn't a determiner of whether users are going to like it and so one of

the important points to build consensus about about is what do we want our product to say what personality do we want it to convey okay so what what do

you want to convey if you're Kmart what do you want people to associate with your brand anybody not cheap value right what

do you want to convey if you are sax Fifth Avenue Luxy luxury exclusivity right uh and what do you want to say if you're

Apple other than pay me lots of money for my products right sty stylish simple right one of the the things that you'll see in in every Apple product is they're

aiming for uh let's assume our users are busy and and don't have time to learn a whole bunch of stuff okay let's just make it work and so getting some consensus on what you're trying to say

with a product is really important CU if you don't all know what you're saying well you're all going to look at that design and and take your best guess and

so if you started out your project by talking to all the stakeholders all the uh you know the marketing people and the executives and and so forth and saying what do you think our our product is

about right what what is it you want people to take away well they're going to give you a whole bunch of words and then users are also going to give you a bunch of words about how they want to feel about that experience maybe not in

direct response to that question but they'll come up over the course of spending spending an hour or so with people and so what's really useful is to write down all those words on a

whiteboard and then start to group them and distill them into clusters and start to substitute words that are a little bit more visual right uh so for example

if you want if you want to say intelligent the word intelligent doesn't necessarily evoke anything visual does it right if instead you say brilliant what does that make you think

right clean bright maybe a little sparkly right uh that starts to make you think in a little bit more visual terms and so we we start to translate what we heard into

a visual a set of visual attributes um called experience attributes typically we're going to get to a small number of these that are are most important uh and and generally they're

going to uh create some some sense of tension so for example um you might want a business tool to to feel both expert and approachable right it can't be

intimidating but you've got to trust that it's it's very rich in uh in technical expertise so how do you balance those two words right uh and so these things are going to lead to some

visual expression now when we start doing this most people in the room are kind of squirming like I don't know you're getting into fluffy territory here I'm uncomfortable uh and so people

can't visualize this often so what you can start to do is use imagery to start to evoke these things okay and you have to focus on on the emotion

now if you're going to use a word like secure for a piece of software right some people might think all right well

images of um bars and guards and and weapons those are all about security well what you want to focus on actually is what does secure feel like it feels

warm and enclosed um it feels solid uh it feels a little bit heavy and you're going to start to use some visual language that that evokes that stuff right uh so for for example that

combination lock up there at the top it has really thick lines it implies a certain strength it's not dainty uh and so these start to help people understand what you mean when you say ah okay

here's that feeling of security that we want to evoke make sense and sometimes we have to say all right but we don't want to take this too far okay you know we we might say we

want our product to feel youthful but but we don't want it to feel babyish right we don't want it to feel infantile we just want to be young but but we're not going to push that too far okay and through discussion of these kinds of

things with a big room of stakeholders we can generally get consensus on what's most important what's least important so that by the time we start to develop the

visual and Industrial look uh we actually have a basis for that discussion and we can say okay remember we wanted to say this here's why it's that color here's why it's that shape so

that it's not anymore about what you like what I like it's about what we're trying to say and we have a basis for that okay and so we've done a whole bunch of analysis we've got some consensus on

what the product is going to be so then we've got to start visualizing the thing okay and here's where most I think design processes break down which is the industrial designers go over here and

the interaction designers go over there and nobody even calls the visual designers yet uh and they do some separate things right there's some wireframes going on maybe there's some hardware and then eventually you try to

make those two work together well that's that's the problem right if instead you spend the first few days of a design uh phase with all three of those disciplines in the room together

all sketching and looking at possibilities you get much better coordination so one of the first things you have to figure out with a physical product is what what platform are we

talking about and what architecture so what do I mean by by those words um I I use the word platform to refer to uh a

set of components that is is a specific kind of input and display method right and the architecture uh as an industrial designer would use the term is about how

those those components are arranged so let's take a look in an example imagine we've got a little mobile device okay we could use multiple different input

methods right we could have something that's basically a touchcreen we could have something that has soft keys around the perimeter right or we could use some

sort of little 4-in one controller right we could do any of those things and so we're going to have to figure out what's ideal from a design perspective and how does that intersect with what's feasible

from a cost perspective because building Hardware uh has a fixed cost for every unit that you build uh and then let's say that we decide all right we're going to go with this 4in one controller

because cost is really a critical Factor well then we have to figure out the architecture right what's the orientation of that controller ver is the screen is that a horizontal layout

or or a vertical layout uh how does that work okay and so we're going to be sketching a whole bunch at the Whiteboard right and and using those scenarios that we started to generate to

imagine how do we step through an interaction using um soft buttons how would we step through that interaction using uh a touchcreen right and we start to see at the Whiteboard oh man the soft

buttons are really clumsy or you know there's not a huge difference between the soft buttons in the touchcreen and so when we start to talk to other stakeholders we can say well here's the design benefit here's the cost

difference uh in conjunction with our our engineering friends and uh and that way we can all make an informed decision about what technologies we're going to use and we're looking at things like

well what kind of data do we have right is it mostly lists if so well a vertical screen is going to pack a whole lot more data in there than a horizontal one right um if we're mostly watching movies

well maybe it needs to be horizontal or maybe it needs to switch back and forth with a gyroscope who knows uh and so we're just kind of working through possibilities using the personas and the

scenarios as our our filter so after a few days then uh the interaction designers can dig a Little Deeper uh on their part of the problem and the industrial designers can dig a Little

Deeper on their part of the problem and so the interaction designers are saying okay based on our requirements and what we've decided about our technology options here are the pieces we have to

fit together and then using the scenarios you can start to break out screens okay so imagine that you're telling a story about um let's use a

simple example like online shopping okay so imagine you're walking you're poking around your your virtual store you look at this product uh doesn't meet your needs you you look at a list of your

search results or category or whatever again and you look at another product oh maybe this one's interesting and you kind of bounce back and forth between overview and detail a lot okay

and so you kind of want all that information readily accessible with you know no more than a click or so on the other hand by the time you've made your decision and you're ready to say here's

my address and here's my Visa number and so forth you don't need all that product list information anymore therefore it doesn't belong on the screen at the same time right so that's a very simplistic

case but what we're doing with the scenarios is we're saying what kinds of tools and information are people using at the same time and that starts to tell us what to draw how many screens do we need well we need a screen that contains

these things those things and these other things and then the flow in the scenario starts to tell us well in theory uh the stuff you use first should be at the top left stuff you use last

should be at the bottom right and that's not always going to be the right answer but it's going to give us a starting place okay and so if you let the scenario Drive what you're laying out you get to a sketch very very quickly

and then you can start to throw more scenarios at it and evolve it uh and before you know it you have a product design that's pretty solid often within less than a week you're starting to see

things really solidify and you've been able to to uh look at a lot of different possibilities okay the industrial designers meanwhile

are starting to uh you know explore foam models uh do some more sketches um check out the size and weight and feel of things uh working really closely with mechanical engineers to figure out how

much space do I need to fit all the components inside here so that they're not designing something this thick when they need that much space

um but at this point the behavior and the basic form of the product is expressed in a pretty sketchy manner now um not all designers do this some

designers will go straight to detail what I find though is when the behavior of a product is still kind of rough and you haven't had a lot of time to refine

it if you present it in detail people will assume you've thought it out in detail so you kind of want the Fidelity of the expression of the idea to match the Fidelity of the idea if you

see what I mean it's kind of like uh in a lab class right you're not going to express things to the nanometer if you measure them with a ruler um because it implies a false Precision we're kind of doing the same thing with pixels right

you don't want to show pixels when you've only had a week to work through something and and it might still shift okay uh and so generally you're going to see pencil sketches or things like that

interaction design um for layout of physical controls I tend to do it in something really crude like this I think was drawn in PowerPoint which is a horrible drawing tool and that's good

because it forces you not to worry about the fussy details and and what we're kind of saying here is all the things having to do with sound live on this side of the telephone and then there's a

screen with with a list and then there's some voicemail controls and things over here so you kind of get a sense of that architecture without worrying about how it looks okay because the minute that you show how it

looks people will focus on that and they won't be able to step back and figure out if the structure makes sense okay but we do have to start working on that visual design language as well and

so we do the equivalent of swatches right so using that set of experience attributes those words that we all agreed on we start to do some visual language studies so for example for that that little telephone sketch I just

showed you imagine the that we came up with three experience attributes well one is approachable so it's got to be kind of bright and cheerful and friendly uh another's exceptional right so we're

going to show lots of Polish and maybe a little bit harder glossier look uh and trustworthy you know well maybe that's kind of soft unassuming uh and that kind of thing and so our color palette and

our shapes are going to be driven by that at the same time that visual designers are doing that the industrial designers are exploring what that would look like in Hardware form so for that

word exceptional for instance where we have lots of glossy highlights and stuff in the in the pixels well we might have some chrome trim right we might have something that's just very severely

modern because it doesn't look like every other phone out there right and there are a lot of little details that will start to imply quality now you can see that this is still a sketch right we still got some kind of crude lines

sketched in there it's not finished and we'll show a series of these to start getting people to say all right we want to go this direction versus that direction and then finally we've got a

concept that's pretty solid we've got a direction for the design language uh and then we can do all of the little detail which is where um you spend most of your time as a designer actually is in the

detail and so we're going to look at things like the physical control behaviors right how many detents do we need in a hardware knob you know is it a jog dial or does it just stay in one

position that kind of thing um you know how does that knob feel and so forth uh at the same time the interaction designers are chunking out

the the design into pieces now um I'm sure you guys are are familiar with the sort of waterfall versus Agile development method thing um actually some blend of the two I think is really

pretty useful right uh once we've got a concept laid out we can chunk the interaction design into pieces so that it can start getting built as we go um

usually something like a three or four week chunk works really well um but what we want to do is touch all the detailed design twice so that we're getting down to the level of

controls uh and language on the screen and things like that once and we're documenting that or in some cases maybe building a a workable prototype of it we're getting all the engineers and

subject matter experts to look at it maybe we're doing a round of usability testing on it and then we're doing a second pass at it to I say finalize but essentially finalize it even though

something might shift a little during implementation okay so ideally we're chunking it and iterating so um we're doing chunk number one we've got people looking at it designers are working on

chunk number two and then they tweak chunk number one after some feedback comes back this tends to be very fast and very efficient and this is where the visual design and the interaction design merge

okay so the basic visual language the visual designer is turning into a visual system uh and they're doing things like establishing a layout grid right so that you know the layout's not halfhazard

you're always doing things in certain columns and certain Heights um if you look at any uh well-designed software from a visual point of view you'll see

that they use some fixed units and some fixed proportions right you'll see that things are a certain number of pixels high or wide and they're always aligned in certain ways okay and so what works really well but is a real pain from a

project management point of view is to have the visual designer and the interaction designer working in the same file so the visual designer sets up the layers generally in fireworks or

something like that uh and and all of the components the interaction designer does a rough pass at laying out the screen based on what's been done at the Whiteboard uh and then the visual designer takes another pass at it

polishing it and tweaking the visual system okay so um I tend to find you'll see a lot of interaction designers out there who say they only work in wireframes and they live in illustrator

I don't think this is a great idea because how do you know all that stuff's going to fit on the screen until you've put it in pixels at a realistic resolution it's really easy to fool yourself in a wireframe so I tend to

think that pixels are a really important medium um or you could do this in you know your favorite prototyping tool things like icons can wait until pretty late in the process because they're not

structural right um You can fill those in toward the end of the uh of the process and generally you know either you want to build a pixel perfect spec

uh I mean a Pixel Perfect prototype or you want to do uh a pretty detailed spec you can get away with really loose specs with small skilled teams when you have a great big distributed team with varying

skill levels um I find that if you don't do a pretty detailed spec you get things going all over the place uh and clients who are Outsourcing find this even more important

um but you do want to use ability test ideally somewhere in here um ideally before you do your last pass on the spec uh having people you know walk through the the product uh some sort of

prototype with some realistic tasks makes a big difference and then at the same time you've got the industrial designer collaborating really closely with the mechanical engineer and they're sharing a file as well they're sharing a

cad file uh and they're they're laying out exactly how all the parts are going to be assembled um where all the part breaks and and specifying whose touchcreen you're going to use and so forth uh for that to get ready for

manufacturing and so they're pass pass ing the file back and forth and the uh the industrial designer is mostly looking for does it match the design intent right let's tweak this curve a little bit let's change this surface a

little bit uh and the industrial designer is working closely with the visual designer to specify the finishes and the materials and so forth that are

used in production so all of this actually works very very well the hard thing is it takes really intense collaboration pretty much every day so

you don't get to go hide in your cubicle for weeks at a time uh in fact if the various members of the design team aren't looking at each other's work at least once a day if not twice a day

they're going to diverge and it's going to get inefficient similarly if the designers aren't meeting with the engineers at least once a week and preferably more often um they're going to start doing things that are too hard

to build for the timeline and so um constant collaboration is really really critical uh you know you you've definitely got to get efficient about it and uh you've got to get skilled at it as well people think collaboration just

happens automatically when you put people in a room and it really doesn't um it's a skill just like writing code or just like uh you know drawing a picture is you've got to learn how to

listen to Somebody's idea right help them kind of give birth to that idea Midwife it a little bit draw them out then reflect back do I understand your idea

correctly and only then start to critique the idea one of the things that I see happens a lot is is people uh you know assume they know what the other other person's saying and then they start to critique it and they don't

actually know what that person meant and that just kind of kills the collaboration right there so you know as you do group projects in class think about how effectively you're collaborating and uh make that part of

your your goal as well because it's it's an absolutely critical skill um so assuming all this works well you get the Holy Grail right you get the uh the great integrated product experience where people think oh

yeah they they couldn't think of the hardware and the software behaving differently because they go so well together uh and that's that's the basic idea so we do have a little time for questions if anybody needs to run away

that's where you can reach me um [Applause] questions how would you go about avoiding people coming up with specific

details kind of working out backwards the process that you mean how would I avoid somebody saying well I want to do this kind of control and then kind of try to

especially for like the kind of user cases right you're coming up with a profile what user is going to do how do you avoid people thinking well I would want it to be this way so I'm going to

come up with a user that does it this um believe me I've seen clients try and do that where um I actually had a client call once and and basically say well we have the product spec we kind of want to

retrofit some personas to it um and my answer was no I'm not going to help you do that um the uh um I really find that research is key uh

I think that if you can get people to not jump the gun and and lay out a bunch of requirements before they've gone and and looked at how actual humans behave

um that is a huge help and even though the design team is primarily driving the research it can help to get product managers or other folks to tag along for a visit or two just so that they can see

reality and you know photos and video clips from research are very powerful um and I I think they help quite a bit because people can see oh you're not just telling me this is so you're showing me six different photos that

demonstrate that people use our devices like this um data is your friend you know I think that it's it's really difficult to win those arguments otherwise

because the people most often wanting to do that retrofitting or the product manager or the CEO or somebody else who's making decisions about the product it's usually them and the only way you're going to win arguments with those

folks is with data in in my experience so um not a perfect answer uh you're still going to see it happening on occasion but um always going back to the

data is a huge help other questions comments yeah one of the things that uh bber plank taught me in the years that we taught together was that you know we teach them we often

teach the normal user centered design pipeline to start out with observation and then do storyboarding and prototypes and then build the final thing and and Bill told me once he says you know look sometimes you've got a neat technical

Innovation and the question is how can we get this out there so hey we just invented tree Maps what are tree Maps good for and so I think there are sometimes where you've got technology and the question is what personas would

want this technology sure absolutely I think that is that is definitely the case I mean if you look at how Innovation happens in industry in general I think it comes from two

directions right um I mean you can look at uh uh people have long lists of sources of innovation but I think it's fundamentally two one is understanding human beings and and how they behave and

what they need and the other is oo here's a neat technology what can we do with it and so yes I have on occasion had a project where somebody says hey what can we do with this pile of

technology and and we still go out and do a little research in sub personas um and it's very imaginative right and and we're not quite as sure that we're going to hit the market that way right we have

less certainty that we're going to come up with a good solution um but still by saying who are our users and what do they need um you can still come up with good uses now on occasion though I've

used the same approach to kill technology right um had a client once who uh who said you know we have this great tool for turning VHS tapes into digital video you plug our dongle into

your computer you buy our software Tada we said okay let's go do some research and and uh we looked at how people were using their photos and their video and and it was one of those things

where there wasn't uh a lot of uh there weren't a lot of digital video products at the time it still a pretty new category but what we found was people

would would swear to you that that wedding video was one of their most important possessions and they would throw it out the window in a house fire right after the kid and the dog to make sure they saved it but in the past 20

years they never watched it you know uh and and we found that like people converting slides and things like that well they didn't do that themselves they sent it out to a service uh and the

services existed and they were really cheap and so we went back to our client and we said you know based on the behavior that we see we're not convinced this is a great investment for you to

build this product but we found another opportunity which is people really wanted the ability to share their lives with other people around them and so if

they could share digital video based on recent events that would be a big deal and so you know our client didn't build the first product and they they started working on the second so you know I

think sometimes um research tells you that you've got some bad ideas too and that's okay um you know design and and this kind of user research is certainly not a complete solution it doesn't guarantee

success I do think it increases the chances of success quite a bit what else yeah could you comment maybe from from

aspects of design that that hit on personas the relationships among people as well as the overall process on differences you might see between tral

desktop um in my experience it's not so much a difference in the the personas as it is in um their willingness to compromise I think the biggest

distinction that I see between the mobile context and and a desktop or sort of tradition traditional web context is when people are sitting at a desk they kind of want a really good tool when

they're out in the world they're willing to put up with crummier tools in order to get that piece of data right it's kind of like the Swiss army knife right when you're at home you want a real pair

of Scissors by gosh when you're out in the field and you you need to snip off a button or something that annoying little pair of scissors on your Swiss army knife is just fine but you would never use it as a substitute for real scissors

right your expectations are lower in that context I don't know if I answered your question though context seems more complicated better things to do like

driving oh yeah well yeah checking on the cup the cup of tea right and and you know personas are are a piece of the puzzle HCI principles are absolutely another piece of it right I mean we know

how much attention people can afford when they're driving and it's not much um and and so we've got to consider those things as well one of the things that any good designer should bring to

the the equation is not just that user research but absolutely a headful of design principles and patterns too um but yeah context is absolutely critical and and the personas are part of putting

that context together um there are a lot of people who will try to debate what is good or bad design right and and they'll say here's a design principle that's absolute there are not a lot of design principles that are absolute because

what's good design in one context might be really awful designed in another right Terry you have a question about training yeah take two polls here one is you say to one student you become a really good programmer say to the next

student you become a really good visual designer and so on you say now get together on a team Kim will run the team and you'll all do great job the other to really be in this you should know some

know some programming where on that Spectrum yeah um interesting question and and it's a multi-layered answer I think uh because this is an ongoing debate in the design

Community as well which is how specialized are you um I think that if you want the absolute best programming and you want worldclass visual design and you want worldclass

Behavior then I think people who are specialized are going to do the best job at those things and so um I think you'll find a lot of consulting firms have very

specialized roles that said if you're a designer uh or even a designer/developer in a startup you may be doing all those things yourself uh and so I think there's always going to be a role for

generalists but I think there's also a role for Specialists um I would say that at a minimum though you want to be somewhat literate in these other fields right I am not a programmer nor would I

ever want to be one but I have enough knowledge to kind of have an idea of what sorts of things are hard and what sorts of things are easy right and and I know enough that I can help translate between the programmers and the

marketing people which is often a designer's job um but you know I want to leave the the programming to the people who are really good at it um I think any good interaction designer

needs to be at least visually literate you know you need to know uh at least the basics of layout and visual hierarchy and things like that um likewise you need to know some physical

ergonomics um I think that Engineers should be at least somewhat in literate in those things as well um but I think it's working vocabulary right it's it's like uh the cardiologist and the

radiologist can talk to each other but they don't have to master each other's Specialties um one of the the trends that I've seen in in some of the academic programs is um in design and I

don't know how much this is happening in in computer science but in design a lot of design schools have students attend all of the same classes for the first year regardless of their design

discipline um which I think is an interesting Foundation right because there are certain fundamentals of design that you need whether you're doing visual design or industrial design or architecture um I think there's some

room for that kind of overlap in a computer science program as well I'm not as familiar with the the Cs programs I'm sure you are more so than I am um other

questions anybody okay thank you all right thanks everybody [Applause] for more please visit us at stanford.edu