SpoolCast: Crafting a WebApp and Design from Scratch, An Interview with Sean Kane Jared Spool: Well, hello everyone. I am talking today with Sean Kane, and he's going to be speaking at the Web Apps Summit in March. And Sean has the we've caught him at a transition point in his life. Hey, Sean, how are you doing? Sean Kane: I'm doing pretty good. How are you doing, Jared? Jared: Yes. You've recently changed jobs, have you not? Sean: Yeah, I was the director of User Interface Engineering at Netflix until recently. And I had this opportunity to go to a really small, very early stage startup called "GetListed," and came in as a founder. I'm the VP of user experience there, so what I'm trying to do right now is build out the technology team and the design team, user interface engineering team. And build this web product from basically nothing.So it's a huge change for me in the fact that I had the entire company of Netflix to help support my efforts. And now it's me and a couple of other guys trying to build things from scratch right now. So yeah, it's a big change. Jared: So now, when you were at Netflix you were there for quite a while, right? Sean: Yeah, I was there for five years, almost exactly five years. Jared: So in that period it grew tremendously. How many users did you guys have at the time that you got there? Sean: I think there was about 500,000 members, and by the time I left there was somewhere around seven million. I don't recall exactly what the number was they reported. But yeah, it was up around seven million at the time that I left. Yeah, it grew a lot.I got to see the company move from its old little building that was falling apart to their new glamorous headquarters that I think you came and visited up there. Jared: I did. Sean: It was a lot of change over those five years, a lot of it for the better. I learned a ton about the way to build really fabulous websites. Jared: Yeah, 14 times the number of users from the time you started from the time you left. That's quite a legacy to leave there. Sean: Yeah, yeah. Jared: Now the thing about Netflix, right, is the entire business runs off this website. So if the website ain't working, you guys are dead in the water. Sean: To some extent. You've got to keep in mind there's an offline component, which is the shipping piece of it. Jared: Mmhmm. Sean: With Netflix you build your queue of movies that you want to see, and then in a separate system the movies get shipped off to you through the postal mail. So one of the nice things about the site was if the site (for whatever reason) when down or if there was problem with it or we were doing an update the shipping and stuff still worked. And so people still got their movies, and could still return their movies. They just couldn't pick them.The website is the only interface to deal with Netflix, unlike Blockbuster where you can go into the store and talk to people. You have the option of calling on the phone to talk to people in Netflix or use the website, and that's it. Jared: So to that end, you had a lot of good management support for that website. Your management really understood the value of the website and why it had to be usable, and all those things. Sean: Right, absolutely. As far as the management teams I've worked with in the past, they're as good as it gets. Reed Hastings is just a super smart guy. He is big into data mining and letting data prove everything. So with that, not only were we driven to create an extremely compelling user experience across the board, everything from picking that movie on the website to getting that shipped within a day to your house.And creating that awesome experience where brand new people show up to the site, and they think, "Oh this might work, this might not." But then when it shows up to their doorstep the next day, it's always this amazing thing.So not only trying to create that experience, but also trying to make sure that we really deeply understood the why of how everything was happening. So with any interaction that was going on the website, we tried to monitor and measure that, and make sure that experience was indeed a better one than any of the alternative ones we had worked with in the past. Jared: Now last week we talked to Bill Scott, who is desperately trying to fill your shoes over there at Netflix. He talked a little bit about the culture of measuring everything and looking at how the design changes perform on the site. While you were there, that was a big part of the whole user experience team's work. Right? Sean: Yeah. If you could take feature idea that everybody at a web company will oftentimes come up with, there are a few different ways visually actually from a design perspective to implement a given feature. It might even have different subfeatures within it. If we ever got into a situation where there were different opinions about what the best way to do something was, we kind of took the tack of: why don't we just build each of those experiences out?And so in parallel we would build out four, five, six different implementations of the same experience to try and find out which one was going to achieve the business goals of what we were trying to accomplish. And also balancing that with the usability goals or the qualitative aspects of that. So we had a great research lab there. We'd bring people in, and we'd watch them use the site. We used that to inform the subtleties of how we might do stuff.But really at the end of the day it was a lot about the quantitative data. How were people interacting with things? What was the end result of what they were doing? Were they adding more movies to their queue? Were they getting their first choice more often? All the certain things like that, we wanted to make sure we deeply understood it.Oftentimes, what we found was that it was surprising. There'd be one person in the room that had an idea. Everybody else would disagree with it, and thought, "There's no way that one would win." Maybe at other companies that would become a very political battle, where people are trying to angle to get their idea through. But since we built them all, and at the end of the day we learned that the one guy that had the idea that nobody else agreed with was the winner. And the data shows it; there's no room to argue.You just say, "You're right. I was wrong," and you learn from it, and you move on. You do something else. So, I can't tell you how many times that I thought my ideas were the winners and they absolutely weren't.[laughs] So it's been a huge learning experience in the way people interact with websites, and the way you build things to meet the goals of your end users. Jared: So now you've moved over to GetListed, and this is a brand new startup. And you guys are still kind of in the cone of silence about what you're building. But we can talk a little bit about what it's like to be in a startup after being in this organization where you had all these resources.At Netflix, you guys became twice the market cap of Blockbuster. So people understood why this was important. You had this beautiful set of usability labs to do the stuff. You had a lot of system data to work with. I mean, you guys had millions of users, so you're collecting data all the time.Now you're in the startup mode. What's different? Has it been a shock, to some extent? Sean: To some extent. I mean, the part that I'm really enjoying is the fact that instead of focusing on one aspect of just the user interface engineering, I have to wear everybody's hat. So, I'm doing everything now.I've always loved to do that. That's one thing forcing me to learn a lot about stuff that I haven't given much thought to in some time.Everything from writing out product specs to talking to vendors and getting people in the door to review technologies. There's other stuff I'm good at which is hiring and whatnot. Doing a lot of different things and compare and contrast. Netflix was absolutely wonderful.Tomorrow is when my first two people start. At the moment I'm sitting in the Sunnyvale office that we have. I'm the only one in at the moment. It's been that way for a few weeks now. Jared: I have this image of you sitting in the conference room putting out a design idea then running to the other side of the table and saying, "That'll never fly!" Then saying "No, it will!" and getting into a big argument with yourself and then storming out of the room and being happy that you won your way while being really upset that you were being so belligerent. Sean: [laughs] I don't think I've taken it that far. It's really different. It takes a lot more personal time. I have to manage my time a lot more carefully because we're still really trying to crank that big wheel, get it rolling. Once it gets moving it will start moving very fast. It's just a lot of manual focus.I described it to a colleague of mine, it's like driving a big semitruck without power steering at the moment. You can keep it on the road but it takes a lot of attention. I'm looking forward tomorrow to getting the new hires up on board and starting cranking and really moving things forward very quickly.It's just the biggest thing right now is how do you take all these ideas that we have? We've got, I would guestimate, two years worth of features that we want to build. Two years in a normal environment of maybe 15 people. We have a ton of stuff we want to do. It's really a matter of prioritization. How do you focus that down into the most essential things to get it out the door? Once it's out the door, how do you make sure you iterate on that very, very quickly so that you're constantly evolving the product.When we launch what we're doing, it's not going to be polished. It's going to have some quirky elements to it. It's going to have some features that aren't fully realized yet. Just bare core essentials of what we want to do. It's very important after we launch that, to iterate on that and constantly show that the site's improving and keep our members that are going to be coming in engaged and keep them interested in what we're up to.I'm sure you'd agree that when you find a gem of a new little site and you constantly see it improving, you feel like you're are part of that experience. You feel more like it's yours. Listen to feedback. Listen to what people want to do. Start doing some research on how people are using the site. I think it will grow very quickly. At the launch time it's just the bare bones of what we want to get out. Jared: Right. Do you have a vision that goes out several years at this point in terms of what this thing is going to be like, but you've segmented it down to the small piece of that that's going to be the initial launch? Is that how you're doing it? Sean: Yeah, Yeah. The thing is, we want to prove that we're on the right track as quickly as possible. This gets into the way Netflix thinks about things too. You try to do things as fast as possible and as cheaply as possible, because you will fail and you'll probably fail a lot. The more you can reduce the time horizon on those failures and the cost of those failures the more you can attempt to get things right.Whereas if you spend six months, eight months developing something and it doesn't work, you're out all that time. You're out all that money. And you don't have much to show for it at the end of the day. But if you could do things very quickly within that couple of weeks and try things over and over again, you'll get it right.Your article, you brought it up again every time I talk to you, it seems to be extremely relevant. Your "The Quite Death of the Major Site Relaunches," it was titled. Jared: Yeah. Sean: There's a lot to be said about evolving an experience over time. If you do a bunch of rapid iterations on things, that's exactly what happens. Nobody's caught offguard by the fact that the entire site's different and they have to relearn it all.You saved money on that development and you've saved money by focusing on these little incremental things. You learn how those are affecting the business. So the experience evolves over time verses just doing something that's going to change it all.There's a couple of things that happen there. You don't have the ability to look back at the old experience and compare them one to one, unless you decide to run these two experiences simultaneously. You're looking at historical data. A lot can change over a week as far as the way user's behavior. There can be seasonal things affecting that. Summer shows up, so people don't get on the web as much. Or Winter shows up and more people are on there.If you're doing big huge site relaunches and big huge development efforts, it's really hard to compare those things unless you're running them simultaneously.From my perspective, it's OK to get something out that's not 100%, because 100% percent is going to cost a lot of money and take a lot of time. But if you get something out there that does what it needs to do and it does it pretty well, within a few quick iterations on it, then it starts to do what it does really well. Then you start enhancing it and making it a much more robust product down the road. It allows you to get to that same end without costing as much money and you learn a lot about it along the way. Jared: I know it's still early, you're just getting your first people in there, but is there anything you had in your Netflix life that you didn't think you were going to be able to move over to your new GetListed life, but in fact you've been able to jerryrig something that's actually working? Did that question make any sense? Sean: I'm trying to think of an example. Yeah, it makes sense. I don't know if there's anything just yet. The one thing I was counting on that actually was a surprise that seemed to get a little bit easier, I think it's the startup climate.I consider myself being somebody that's fairly good at hiring and recruiting, and one of the things I noticed was that in trying to find a couple of people people really want to work at startups right now. The environment in the Silicon Valley is great for working in startups right now.I was actually pleasantly surprised to see that enthusiasm from the people who wanted to come onboard and try things. As far as something I wasn't expecting, I can't say there is something yet, but I'll definitely let you know if something comes up. Jared: How about something you had in your life at Netflix that you're missing now? Something you wish you could have brought with you. Sean: Just the people and the level of experience of everybody that's there. It's something I miss a lot. If you don't know something or you want to have some more clarification or you need some help with something, there's plenty of people to ask and to work with. Everybody that's there is extremely competent and extremely bright, and so, even just hallway conversations about random stuff always was very interesting.I have to build that all now, and so that's one thing that, the sooner I can get that going, I think, the better it will be. But it's like I said. I'm sitting here by myself, so that's the one piece, is just missing the people and the people that I've worked with in the past. But I get a little bit of that through talking to Bill Scott. He keeps me in the loop on all the gossip over there, so that's good.[laughter] Jared: Is being in a smaller organization giving you any sort of flexibility that you didn't have before? Sean: It does, because if I want to do something granted, I could always do it before, but there were priorities, as far as the business priorities, and so it was a little bit harder to get some personal, passionate stuff out, unless you kind of did it as skunkwork projects.But right now, I know what needs to be done. It's up to me to set those priorities and figure out how everything's going to happen. So there's a lot more flexibility, from that regard, that I get to pick everything. But at the same time, I can't as quickly go and sanity check that with as many people.On the GetListed team, right now we have the other two founders and myself, we have somebody in our sales organization and the two new folks that I hired, and we have actually a really great technical advisory board. So once a month we all get together and meet, and so that's where I've got to present everything. I talk to those guys occasionally as well, but it's just not that daily, inthehallway thing. Jared: I've got to formulate this question right. In terms of sort of visualizing, you're doing two things at once. You're building this application at the same time you're recruiting staff. OK. I think I've got the question. This is where we're going to be editing.[laughter] Jared: As you're choosing the people to be on your team, are you thinking about the functionality that you're building and trying to match people to the specific functionality? Or are you trying to build a more general team that is going to be capable of all sorts of things that you haven't even imagined yet, and so you're not really looking at the functions that you're building versus the people that you're trying to hire? Sean: That's one thing I learned at Netflix, which was: always try to hire great people, superbright people, because really bright people have this ability to just do new things. And it's not really that big of a challenge. They enjoy learning new things.And so, we have the technology. We want to build our stuff. We want to build it in Java. We want to try using the Tapestry framework to build it. And we have some other things that we may want to do with Adobe AIR and some other technology that's out there that's kind of interesting.So we have these kind of musthaves, as far as the technology components are concerned. But if we decide to change direction in the future to something else, or try new technologies, the folks that I think I'm bringing in and the folks that I want to bring in down the road, if they're bright folks, they can change on that dime with us. Because one thing is certain: the fact that we are a startup and we will have to change. So we need to be able to respond to that and do it very quickly.Yeah, so I would say I want to build both a specific team and a general team. [laughs] But we have those specific needs that we need to meet right now, and we just need to be able to adapt to changes in the future that we don't know about. Jared: How do you find people who meet that criteria of just being really, really bright? I mean, do you just put out an ad that says, "We're looking for bright people. If you're not bright, don't apply"? Sean: [laughs] To some extent. I think when you put a job ad out, there's a lot of luck in play in that. Some of it's on instinct. There's a lot of it on, when somebody actually starts proving themselves in their ability to deliver, you can get at some of that though references and through the interview and just getting a general sense about somebody. But at the end of the day, it's all about how somebody actually performs.The other thing is I try to really make sure that the people that I know, that I've interacted with in the past, that I've worked with before, just try to make sure that I keep in good contact with them, because your network is your best resource. So it's always great to keep in contact with really great people. You may not work with them within the first couple years that you've met them, but you never know when things are going to change, and so I think that's extremely important.I've had some success stories with people that I've talked to for many, many months, and sometimes a year or so. And finally, when the situation was right because that warm tie was there I was able to get a great person in the door. So yeah, a lot of it's just nursing your network. Jared: So, as you're building out this team, are you thinking of it in terms of specific roles? Or are you going for generalists at this point, and you just want people who are just really bright and can basically do everything? Or are you thinking, "I'm going to have an information architect. I'm going to have a programmer. I'm going to have a user interface developer"? Sean: Yeah. The way I'm looking at that right now is, at the moment, we have an alpha prototype kind of built right now, if you will. And so, that was done by a couple external consultants, so that had the designer working on it. So we have designs. We have our logo. We have some technology, just something to help us prove that model. And that was kind of built as, like I said, a prototype, to make sure that it's going to work like we think it will.And then this secondary platform, what we're calling our beta product, that's the platform that's going to be built to scale. So we have some luxury in that I get to leverage the designs that have been put into this alpha prototype. And so, at the moment, I'm not looking to fill designer roles. I'm stepping in, myself, to the information architecture piece, and a friend in engineering on that.So yeah, there's specific roles that I'm looking for somebody that really knows the technology, the middle tier and the backend side of things really well, and then somebody that knows kind of the middle tier towards the frontend pretty well. That being said, there's always room for a really great generalist that can get everything done and help everybody else in their various areas.And then, down the road not too much longer now, once we kind of get closer to the finish line on shipping the first release we'll start looking more seriously at hiring designers and product managers and start growing things out from that regard, because that's when we're going to want to start looking at the various components of the site and starting to think a lot more deeply about how we can implement the various features that are on the home page, for instance, versus take another section of the site and that's got its own set of features that need to be worried about.And that's not something that I can continue to do forever, so once we get across that finish line, we can start looking at those elements and start focusing on that.So, yeah, to answer your question, there definitely are the specialists. There's a little bit of room for generalists. But then, the plan is to continue to bring in specialized people over time. Jared: OK. So, next week or in the next few weeks, I guess, if I understood what you were saying your team basically triples in size. Sean: Yeah. [laughs] Jared: So, with these two new people coming on board, what are you planning as sort of your first steps to get them as productive as possible? Sean: The first step is to figure out I have lots of documentation that I've written about what we want to build. And we have a visual design for it. But the final design of both the software systems and the information architecture, I wouldn't say that that's finalized.So I'm anticipating to spend a few days just talking, designing on whiteboards, writing stuff down. And then it's about just getting in and building the very core bits of functionality, not worrying too much about how it looks, just making sure that it works and it follows the architecture that we designed out in those days previously. And then, from there, it's just a matter of continuing to put the polish on it until we feel it's at that ship point.So, like I said earlier, we want to make sure that we don't put too much polish on it because that will just delay the ship for little gain. And if we can ship it and then improve upon it, then we'll get the best of both worlds there. Jared: To get these folks up to speed, you said you've got stuff you've written. Are these just Word docs, or do you have prototypes? Sean: We have a functional alpha prototype, so, in the event that I don't have something fully specced out otherwise, we can always fall back and look at that. Otherwise, yeah, I have written a lot of Word document PRDs. I've written flowcharts. There's even some stuff we're trying to get patents on.Yeah, so there's a lot of documentation, and there's a lot of thought put into all the features that we want to build. And we just need to make sure that we then specify. What we need to figure out, still, is what exactly does the database design look like? What business objects are we going to build within the Java layer? What modules are we going to build on the UI layer, and how are those going to behave, and what are the rules associated with those?So those are some of the stuff we still have to figure out. But, with a few people in a room, that's stuff we could probably crank out fairly quickly, I'd imagine. Jared: Right, right. As you bring these new folks on, what are you thinking your biggest risk is, in terms of getting to where you want to go, from a productivity standpoint? Sean: I think it's the project management component of it. There's a lot that needs to be done in a short amount of time, and so how do you make sure that you're spending your time in the right areas, doing the right things that are going to result in, basically, the most bang for the buck with everything you do, and making sure that those things happen?And especially, there's some pieces where we're dealing with thirdparty technologies. Some other pieces, we're dealing with third parties, and making sure that all those ducks get in a row in a very efficient way. I think that's one of the biggest risks, is just making sure that the project runs on time and the time is being spent effectively. Jared: Right. OK. That makes perfect sense to me. Are you doing anything special to sort of keep that in check and monitor that? Or do you just..? Sean: No. Right now, I've gotten some experience with our alpha prototype and the guys that are working on that. I haven't done the project management thing to a real large degree until recently, so I've learned a lot about that. [laughs] I think it's just a lot of communication. It's harder to take something as aggressive as this and what we're trying to do and do it in a vacuum.That being said, one of the other reasons you want to talk about hiring great people is that, if you have people working for you that you can trust and that you know will do a good job, you can start to rely on them and you don't have to explicitly lay out every single detail.Also I think having people that are part of the company is going to help there a lot. Some of the things I find, when doing these similar types of things with third parties, is that their interests are completely different than the interests of the company. And so it's a lot more necessary to spell out and dot every I and cross every T, all the little, subtle details to make sure that that specification document they hand off is something that they can execute on.But if somebody is working within the company, they want the company to succeed. They're on the same page, as far as the "why" of why everything's happening. They're a lot more likely to dot the Is that aren't dotted and cross those Ts and just kind of do some of that stuff autonomously, which is a big, huge help in these types of situations. Jared: So, at the Web App Summit in March, you're going to be talking a bit about what these experiences you're having are going to be, so can you say a little bit more about what you're hoping to talk about there? Sean: Yeah. I'm hoping to have some specific examples of things that I was able to bring over from Netflix and apply, and talk about the successes of those, or hopefully not the failures of those, and talk about how I tried to adapt what I learned there into applying it at Netflix.In light of the fact that we can't do a lot of the testing yet, because we haven't launched it and we don't have the traffic to do that testing, how can we kind of achieve those end goals in some of the stuff that I'm hoping to do between now and then, and in efforts to get the site up and running?Yeah, so a lot of comparison and contrasting between the two. I'll give some examples of stuff, some of the methods that were very effective at Netflix and how I'm trying to replicate those with GetListed.One of the questions from of the talks that I did at the last Web App Summit and other talks that I've given that I've never really had a good answer for, and now I think I'm getting much better at being able to answer that question was always: "Sure, at Netflix, you guys have all this traffic, so it's easy for you to do the quantitative testing and get statistically significant results back. I'm a startup. I don't have very much traffic. How do I deal with that?"I didn't really know how to answer that question before. I think I'm figuring out a way to answer that question. I hope to have that question answered [laughs] by the time I give the next talk at the Web App Summit. Jared: Well, that sounds good. We'll check in with you, hopefully, before then, and see how things are coming along. And we will certainly be anxious to hear what this big, mysterious, secret thing is that you're working on... Sean: [laughs] Jared: ...as the details come out and we get to learn more about it. It's exciting to get a glimpse as to what it's like from the startup world. And I think it'll be interesting to see where you are in March and beyond. But thanks for taking the time to talk with us today. Sean: Absolutely. Jared: And we'll see you at the summit in San Diego in March. Sean: OK. Great. Thank you very much, Jared. Jared: Thank you. Sean: See you there.