Going from College Dropout to Designing Pinterest's Grid

This Soapbox has come and gone!

  • Sahil Lavingia , Founder Gumroad and Former Pinterest Designer

We hope to see you at the next Soapbox. We'll update this page soon with interesting tidbits about the event plus the podcast!

About Sahil Lavingia

Everyone had a great time when Sahil Lavingia, former designer for Pinterest and founder of Gumroad, dropped by ZURB HQ for his soapbox earlier this month.

Sahil was very entertaining, even dropping a few F-bombs while candidly sharing his reasons from dropping out of USC to help get Pinterest off the ground. He took us through the real world inspiration that influenced the unique pinboard He also spoke about why its important to eat your own dog food by actually using your own products.

Feel free to listen to the podcast as you read through the summary of the event below.

Listen to the Sahil's Podcast

iTunes RSS

Early Start

At 14, Sahil built his own products and created some 20 apps, making more than $100,000 by the time he was 15. He got his start by simply opening up photoshop and playing around, teaching himself how to use the tool. That’s when the Web was really beginning to take off. Sahil said he jumped at the chance to construct websites, doing contract work and teaching himself how to code.

That quickly transitioned to actually building my own products because I got bored just building pictures. So I learned to code, and I got bored of just learning how to code to build other people's shit. I wanted to build my own.

But the root of it was a bit of sibling rivalry. Sahil said that he got into computers because his mother asked his brother to fix an problem with their internet connection. "And I got really pissed off," he said. Which is why he got into computers.

Sahil would build products on the weekend, release them and then email blasted bloggers who wrote about apps. All his ideas came about because he wanted to build stuff for himself, such as an app that would allow him to mess around with color palettes while he was riding the bus.

Before it was "I wish this existed" and that was the end of my stream of thought. And now it’s like "I wish this existed."

Sahil said there’s nothing else worth pursing than building something for yourself.

It’s really hard to figure out what people need. People say they need one thing when they actually need another. Or they say they need something and they don't at all.

He said it’s very difficult to gain traction. It’s better to start with one user — yourself.

You’re gonna start with yourself then you’re going to have a pretty good person to ask whenever you want about this product about “is this product working for you.” And then you can worry about getting a second user. If you start with zero, you have nobody.

That makes the feedback loop instantaneous, he said. You figure out first how crappy your product is and you don’t have to what to hear back from users.

Every product I’ve built, I’ve built to solve my own problems.

Dropping Out of College

While a computer science student at USC, Ben Silbermann recruited Sahil to get Pinterest off the ground. Ben found him because of another product he built and got in touch. Sahil then started working on the app while still taking classes. Eventually, however, he packed his bags and dropped out to become one of the founding members of the popular pinboard site. And for the most part, his family was very supportive.

I wasn’t being very stupid about it. It’s very easy to say "fuck college, I’m going to do a startup" or whatever. I wasn’t ... I had a very specific agenda: I wanted to figure out if this career path was for me.

Besides, USC would still be there as a safety net, he added. For him, it was all about knowing whether or not he wanted to make a career out of product design and if he wanted to actually start his own company.

Taking Inspiration from the Physical World

Sahil wore many hats as the first designer hired for Pinterest, from plugging away at the front-end code to building the iPhone app. He was one of the architects responsible for how the site looks to this day. Although, Sahil won’t take 100% credit for it, saying that it was really a team effort.

When it came to the unique user interface, the designers looked to the real world. In the case of Pinterest’s offset grid, the inspiration came from scrapbooking and pinboards.

It’s kinda what people do anyway. I think most of the best products, like Jack Dorsey says all the time, technology should use whatever human things we’re used to.

Sahil said the site has a very physical element to it, like a scrapbook. The offset grid that Pinterest uses has been around for a long time, like in restaurants and college dorms.

It’s just that no one decided to take that, which had already existed and is used all the time. You go to a girls’ floor in a college dorm. What does every person’s door look like? It’s a pinboard.

Pinterest was just the first app to take that and digitize it, making it a core feature, said Sahil. “Everything I’ve worked on has had a very physical element,” said Sahil, referencing Turntable.fm.

Jumping Off the Rocketship

But just as Pinterest was taking off, Sahil jumped ship to build his own product, Gumroad, where you can sell anything you want from music to icons to books. In the process, Sahil gave up ownership in Pinterest.

I totally jumped off the rocketship … but I jumped off with a parachute.

For him, it was more risky not to do what he wanted to do because his ultimate goal is to build a product everyone uses and fundamentally changes the way they interact. Pinterst, however, helped put him in a good place to help him do just that.

It was less risky than how it looked from the outside.

Creating Lemonade Stands

For Sahil, Gumroad is about creating lemonade stands on the internet, not store fronts. That’s because there are a "crazy amount" of roads for people to connect with their potential customers, such as Pinterest, Facebook and Twitter. The world is different than when iTunes first came out, charging artist 30% to get their work out there, he said.

"It’s hard to sell stuff. It’s not that easy," he said. With Gumroad, Sahil hopes to make it easy for artists to share their work to those potential customers and cut out the middleman.

So far, the selling platform has raised $8.1 million, all from heavy-hitting investors. Sahil said that fundraising wasn’t nerve racking or difficult. Everything happened through a casual conversation, where he didn’t solicit for money. He just talked about the product and found investors were willing to fund his product.

It was always about the product, about the user. I don’t really give a shit about raising money unless it makes the product more awesome.

He said if you have to raise money, it’s going to be a pain, especially going after VC funding. He advised that if you’re at that point where you have to raise funds, you might be better off going after a loan. "VC money isn’t the best money when you’re in that situation," he said.

But it’s not about the immediate return on the investment when it comes to fundraising. It’s about the potential of the product 5 years from now, not 5 months from now, said Sahil.

Our conversation with Sahil continued as he took questions from the audience, including how he differentiates Gumroad from other selling platforms, such as eBay and Etsy. We’d like to thank Sahil for chatting with us and to all those who attended the event.

Don't Miss Out on Our Next Soapbox


Ryan: I want to get started here with a little intro. When Sahil Lavingia was a college student at USC, Ben Sullivan recruited him to help get Pinterest off the ground. Soon Sahil was packing his bags, moving up here, and becoming one of the founding members of the popular Pinboard.

As one of the first designers, he was one of the main architects for how Pinboard looks nowadays. But just as Pinterest was skyrocketing into the stratosphere, Sahil left to start his own company, Gumroad, where you can share and sell anything to your followers. Everything from books, to music, to even icons, and that product has been around for 8 months now.

Sahil: Yes, it launched in February.

Ryan: Eight months, and it has raised $8 million. I want to get into all of that, but first, please give a warm, Soapbox welcome to Sahil. [applause] Thank you very much for joining us and making the trek down here to Little Campbell, where it's super hot today.

Sahil: Yes.

Ryan: Yes, great day for shorts. I should have worn some myself. I kind of want to go back in time a little bit and get the secret origins, so to speak. You've been building products since about 13 or 14?

Sahil: Yes.

Ryan: And you sold apps on the App Store on iTunes and made more than $100,000 before you were even 15. Has that tinkering, building spirit always been with you? Tell us a little bit about some of those early efforts, and what you learned about building products.

Sahil: Totally, yes. I actually started designing when I was 13 or 14. All the school computers came with licenses of Photoshop, so I had a few thousands of dollars worth of software at my computer, so I might as well have messed around with it, right? So I did. I just opened up Photoshop and started Googling around how to learn and utilize this piece of software.

I got addicted to making random fonts and layer styles and stuff like that. Nothing really that interesting. The Web was starting to take off, so a lot of people asked if there was anyone around who knew how to make websites. Not knowing how to code or really anything besides making little pretty pictures in Photoshop, I decided that yes, I could do that. I started to do contracting work for a bunch of people. I learned how to do everything else because I had to. That's kind of how I first got started. That quickly transitioned into actually building my own products, because I got bored of just building pictures.

Then I learned to code, and I got bored of just learning how to code to build other peoples' shit, and I wanted to do my own. I just started building my own products. I think at the root of it all was the reason I even got into computers at all. I never really cared for computers for a long time, actually.

There was this one day where there was a problem with our internet connection, and I got home and my mom asked my brother for help. I got really pissed off. I was really annoyed that she thought that he was better. He was actually way better than me at computers or anything tech-
savvy, but it didn't matter. I was mad. That's actually why I got into it in the first place. Which is funny.

Ryan: A little sibling rivalry to start, yes.

Sahil: Yes, I mean even at school, when I started learning how to design, I remember that when I really got into it was because my friend started designing pixel art stuff for some video game. I was like, "I could totally do that". So I did.

Ryan: Sweet. What were some of your own first, early products that you developed yourself? What was it about them that you were either trying to problem-solve or trying to do because you thought, "Hey, that's a nifty idea."

Sahil: Totally. A lot of them were just me stumbling upon something that I thought was broken that shouldn't have been. The first product that I think I ever had an idea for was this thing called Tweader. Twitter plus Reader. What I wanted, I didn't care who built it, I just wanted wall-to-walls for Twitter.

Facebook had these wall-to-walls so you could see these conversations between two friends, and you couldn't do that with Twitter. You can see thoughts between two people, and I thought that was really weird. So I built it. I designed a wall-to-wall interface, but for Twitter. I spent a weekend building that, and then at the end of it, I had this amazing launch strategy, actually.

It was just to spend 8 hours every Sunday. I would try to do a product every week. I would spend 8 hours every Sunday and just email every single person on the planet that could write about it. [inaudible 05:09] all these things. And it led to press every single time.

I think that one got on Nashville Reader, or whatever blogs were there at the time. That was my first product ever. The iPhone stuff came out, so I built this thing that would let you track all these arbitrary data points in your life, like how many push-ups you would do per day, or how many glasses of water you would drink.

I felt like it was weird that for every single stat I wanted to track, I needed a different app. There was a weight-tracking app or a water-drinking app, and I was like, "There should just be one generic one." You just specify the unit depending on what you wanted to track. Glasses of water, pounds, whatever.

I built that data, was in mac world. I built Colorstream because I was always bored on the bus to and from school, and I wanted to like fuck around with color palettes on my phone. I built a little color palette manager called Colorstream that was on Apple's homepage. That was really cool.

Before, I would be like, "I wish this existed". And that was the end of my stream of thoughts. Now it's like, "I wish this existed. Wait a second, I can actually build stuff. I can just build this myself." Hopefully for other people too, if they need it.

Ryan: Right, right. Do you find that that still carries through to this day? You just look and you see, "Gosh, you know, that thing doesn't do what I want it to do. Let me make that do what I want it to do."

Sahil: Yes, totally. I think it's not even worse pursuing anything else. It's really hard to figure out what people need. People say they need one thing when they actually need another. Or they say they need something, when they don't at all. Personally, it's very difficult to gain any sort of traction ever.

You might as well build something where you are going to start with one user. You are going to start with yourself. Then you have a pretty good person to ask, whenever you want, about, "Hey, is this product good? Is it working for you?" Then you can worry about getting the second person.

Whereas, if you start with zero, you have nobody, your feedback loop is like sending an email to someone and telling someone to use thing, and then waiting for the response. Whereas, if you are building something for yourself, your feedback loop is basically instantaneous.

Ryan: Right.

Sahil: I think that's really important. Every single product I have ever built, I built it just to solve my own problem.

Ryan: Right. Any art, you have to do it for yourself first before you put it out there for other people.

Sahil: Yes.

Ryan: Actually, I want to kind of flash forward a little bit to when you were at USC. Go Trojans. You get recruited by Ben Silverman to go and help him with Pinterest. How did he find you? What was the moment that you made the decision that "I'm going to leave USC, move up to the Bay area, and I'm going to start working on this product and drop out of college." Was there any fear or anxiety in terms of making that decision?

Sahil: Yes, there definitely was. He found out about me because of a site called Hack and Use'. I had released a new version of Data, or something like that, and he sent me an email saying, "Hey, Data is really well-designed. I work on this thing called Pinterest. We need a mobile app, and you look like you might be able to do it."

At the time, they were incredibly small. Just 3 guys. They didn't really have the capacity to a mobile designer or developer or whatever you want. They just wanted one dude in college to go do that. And I was that dude. That's how it started. I started working from USC and I flew up there to meet up with them and stuff, but I was working remotely for a long time. Then I started getting hit up.

I continued to build other stuff as well. It wasn't just Ben who wrote me. A bunch of people did. I started talking to all these people. I was up in San Francisco once every two weeks or so during my first semester of college. I kind of figured out that I could just do this full-time.

I went into college, literally had no idea I was ever going to leave. I was 100% set on getting my degree and doing whatever. Coming from Asia, where that's kind of the mindset. Getting all these offers, it was like, "Oh, shit, I could actually just do this now." Why would I wait 4 years to go do this?

My goal was to get a degree and go do this. I could skip 3 and a half years, that sounds pretty good. I can stop paying USC $50,000 a year or whatever. That's how that started. I told Ben, "Hey, I have all these offers. I would love one from you guys, because I think you guys are super early and super onto what I want to go into." That's how that happened.

Ryan: I have a Filipino mother. Was there any resistance in terms of when you told her, "Hey, I'm going to skip the college part. I'm just going to go start working"?

Sahil: A little bit. Definitely from my dad. My mom was actually pretty cool about it. My grandmother was also totally cool with it. She was like, "Oh, yes, so you're leaving college?" I was like,
"Yes?" She's like, "That sounds awesome!" And I'm like, "All right." But I think I wasn't being stupid about it. It's very easy to be like, "Fuck college, I'm going to go to a start-up" or whatever. But I wasn't. I had a very specific agenda.

I wanted to figure out whether this career path was for me, if I was going to get paid a salary. That's how I paid rent. There was enough of that. I wasn't leaving USC. I wasn't just peacing out. I was actually going to go, and if it didn't work out, I would have gone back. That's the safety net. The safety net is USC.

Even if all hell breaks loose now, even today, I could totally go back, and I would have lost a year and a half, 2 years. Which is not that bad for being able to figure out if this is really what I want to do. That cost-
benefit analysis made a lot of sense.

Ryan: Right. In terms of that, it's almost even more beneficial because you're not doing two or three years of a different major and then saying, "Oh, I want to switch at this point".

Sahil: Yes. It's very similar to that. Most people have no idea what they want to do. A friend called me yesterday saying she was switching her major for like the 11th time. I'm like, "It doesn't matter. Whatever major you end up getting, you are not going to use it. You are going to figure out what you want to do and you're going to go do that."

Majors are great. Having a degree, having that piece of paper is still awesome, and I would love to have it if it was easy to get. It's just not for me and not worth four years of my life.

Ryan: Right. I want to talk a little bit in terms of what you did with Pinterest. Because you wore tons of hats. You did front-
end code, you did back-end, you did the mobile app. A lot of us, we use Pinterest, and we are very fascinated by its user-interface, which is unique. Now it's being mimicked by Facebook and even the new MySpace to a degree.

How did that interface evolve on the mobile app? What were some of the challenges there that you guys faced in trying to get it to all work?

Sahil: Yes. The funny thing with MySpace is the first time TechCrunch ever wrote about Pinterest was two years ago when the old MySpace copied Pinterest's layout as well. They're used to it. It's kind of funny. Then they never wrote about us again because they stopped caring. Thank you about the layout and stuff like that. I think how mobile it is unique.

Now it's in a few other places, but we were pretty excited to be the first people that offsite [inaudible 12:39] on mobile. That was fun and challenging to do on such a small screen. I'm not going to take 100% credit for this at all, but it came about because it's kind of what people do anyway.

Most of the best products, like Jeff Dorsey says all the time, technology should use whatever human things we are used to. Whenever we thought about how people used Pinterest, it was like scrapbooking . It was like collecting a bunch of recipes. It was all these things that people did that had a very physical element to it.

How do scrapbooks look? They look like collages. They look like these offset things that you just kind of stick together, and that's kind of what inspired it. There were other digital things. Tumblr themes were probably doing it at the time too. Facebook actually had an offsite [grid] in 2004 that nobody knows about. But that's how it started.

It was funny, someone was like, "Yes, I don't believe you. This is really revolutionary." I'm like, "Wait a second, let's go around." I went around every restaurant on my block, and four restaurants had the Offsite Grid on their wall. They have for 40 years. Just nobody decided to take that which had already existed and was used all the time.

If you go to a girl's floor in a college dorm, what does every person's door look like? It's a pinboard. I think we were just the first to take that and digitize it and make it the core feature of our site. We just felt it was that important.

Ryan: It's like in any kind of products base. We are very much moving from physical products to digital products. And it's like, how do you take that physicalness that we are used to and just moving it into the digital space? Kind of, just to pick up on that thought, are you still doing that now?

Sahil: Yes, I try to. I think everything I've worked on actually has a very physical element to it. Gumroad probably being the least. I worked on Turntable. I built and designed their mobile app. The reason I thought Turntable was really exciting was because of the room.

It was inherently like a very physical thing, right? Conceptually, you understand it. Because if you use bits and bites and no one really understands what these abstract things mean. If you, say hey, you walk into a room with other people and you listen to music. I understand that pretty quickly. I think that works very well.

When people use the product, they use it in a way that makes sense. They're like, "OK, people are in a room. If I like something, people see that." You would see that in Twitter, same thing. They use words that are pretty physical. "Follow" is a very physical word. "Subscribe" is not. Right? I think there is probably a very big reason why they use "follow" versus
"subscribe". Things like that.

Ryan: All right. It's like, we're trained for one over the other, a little bit.

Sahil: Right.

Ryan: Just as Pinterest was skyrocketing into the stratosphere, as I said, you jumped ship. Correct me if I'm wrong, you almost gave up ownership there, which is a really ballsy move on your part. I wanted to know why you left and what was the decision that spurred you to leave?

Sahil: Yes. I totally jumped off of the rocket ship, but I jumped off with a parachute. I think the Valley has a very good support structure. My safety net was still USC, so whatever I did, I was pretty confident that I would always be able to go back to that.

But at the end of the day, for me it was more risky to not to do what I really want to do, which is, I had this idea for this thing that would help people sell stuff really easily. I wanted to go build that.

At the end of the day, my ultimate goal hopefully in life is just to build a product that everyone uses and fundamentally changes how people interact with each other. That's hard to do. That's not easy. Most people will never get there. I might as well optimize for that.

Sure, making whatever millions of dollars in stock like we've had with Pinterest is great, but that's not my end goal. If I did that, I would have, after four years, gone and done my own stuff and hopefully accomplished my real goal. So why not just start now? Pinterest, being on a rocket ship put me in a really good place to raise money and meet people that could work with me on this thing. So I think it was less risky than it might look from the outside.

Ryan: Right, right. In terms of Gumroad. I want to talk about that a little bit, as well. In terms of, where did that idea originate? That was one of the things you wanted to do. That was why you left Pinterest. How is that product different from any other online or digital retail space? Give us the elevator pitch, so to speak.

Sahil: Totally. I think selling stuff online is really difficult. We try not to even be a retail space. We think of ourselves as letting people create lemonade stands on the Internet rather than storefronts. Whereas, a lot of these other services try to create storefronts, because 5 or 10 years ago the Internet was basically these pockets of activity on the web. You would have to visit one of these places.

Using the physical metaphor, you would have to drive to this place and you enter a strip mall of stuff and you could go and buy stuff from Amazon or iTunes or what have you. But now, with things like Twitter, Soundcloud, YouTube, Reddit, Pinterest, Tumblr, Facebook, you would now have these crazy amounts of roads. You can have your own connections with everyone else. You can share stuff to them, etc.

We think there is enough proficiency on the internet that you should be able to create lemonade stands on these metaphorical roads that now exist. Let's say, iTunes. iTunes launched 9 years ago. They took 30%. If you wanted to sell music online, that was the only way to do it.

There was no Facebook, there was no Twitter, there was no Soundcloud or YouTube. Today, all of these things have changed except that 30%
experience. That experience, to the world, is totally different. To me, for example, if Lady Gaga has 30 million followers on Twitter, 40 million followers on Facebook. She has this very personable relationship with each and everyone of her fans. Every fan thinks they are talking directly to Lady Gaga, because often they are, and vice versa.

That's the way that she should sell. She should sell directly to them through the ways that she is used to talking to them. From what we've seen, that increases conversation and fan happiness and revenue, so she can go and make the stuff and do what she really wants to do, which is make music, right?

It's really hard to sell stuff. It's not that easy. Theoretically, it's not impossible, but it's not that easy to do. But that's our goal. Our goal is to take how difficult it is to sell something and make it as easy to share something.

Ryan: It's almost like you are trying to cut the middle man out. Kind of what like Louis C.K. did. He cut out Ticketmaster.

Sahil: It's taking what Louis C.K. did, and applying it so that every other person on the planet should be able to do that without having to spend $200,000 like he did, on his website.

Ryan: Right, right. You've even got some musician now, like Girl Talk and stuff, using this as well. Where is the product now, and now that you guys have been at it for eight months, the problems that you are now solving that you may not have been solving initially when you first got into it?

Sahil: The problems we were solving eight months ago were just making this thing work at all. Doing payments online is not super easy, as we quickly found out. Now we do that pretty well. Scaling that so that it applies to way more use cases, way more customizables, way more flexible and retains the simplicity that existed eight months ago when it was a Bentley with a credit card form.

You upload a file, you tag it with a price, you have a link and you share it. Done. And adding all that functionality that people might want. They want to customize their page or do flexible pricing or a charity model, or give people different things depending on what price they pay and all that.

Potentially sell multiple things in one. Enabling all of that, but it's conceptually the same product. That's what we were really struggling with and trying to figure out.

Ryan: Right, right. In terms of that, you guys have also raised about $8 million in Angel funds. This is your first start-up, this is your first go at it. Was that a nerve-wrecking process? Did you get any advisers or seek out some specific people? I'm sure everyone here would like to know how to make it an easier process.

Sahil: It wasn't nerve-wrecking, nor difficult, nor pain-staking or nor time-intensive, unfortunately. Or, fortunately, for me. It was none of these things because I didn't really want to raise money. It wasn't a thing that I was like, "Okay, I want to spend the next two months trying to raise a $1 million." I was like, "Hey, I have this thing. You might find it really intriguing. I might raise money. If you are interested, let me know."

Every single conversation was casual. I never brought up the fact that I was raising money, ever. It was up to them to bring it up. If you're a psych guy or whatever. Let's say I have this meeting with an investor. Obviously, it might be someone who can have a meeting. If they have to say,
"Hey, are you raising money?" Now they're the ones leaning forward. They are already on 'yes mode'. I never really was very aggressive or reading books or talking to people about how to raise money.

I was like, "Hey, I have this product." It's always about the product, it's always about the user, I don't really give a shit about raising money unless it makes the product more awesome, right? That should be the goal. If you have anyone that you think could be interested, let me know. A lot of times they will be like, "I am interested". You know? Then, that's how that happens.

It's great because you're not raising money until you have raised all your money. Because you never start raising money until you're done raising money. But really, if you have to raise money, it's a pain in the ass. It's going to be really difficult. It's probably a bad reason for me to raise money right now. Raising VC money is probably the worst thing to do when you need to do it.

If you want money, get it from your customers. Get a loan from a bank. Increasing money is not the best money when you're in that situation. But, if you're like, "Hey, I have this cool idea. I want to hire five guys to build this thing and stop working out of my living room" then, it's a pretty good option.

It's all about the potential too. I never mentioned a single number that we had. How many users we had, how many page views we had, how much money we were processing. If anyone would ask, I would just say, "Assume zero".
Because I wanted people to give me money based on myself and on the idea that I had that might exist in five years because of Gumroad.

Not because of our current product or our product in three months. I thought that was really important. It's based on the 1% chance you might be a billion dollar company. Not the 50% chance you might be a $10 million company or what have you.

Ryan: In terms of where are you guys at as far as users and scaling at that point?

Sahil: It's good. I still can't talk about official numbers, but Girl Talk uses us to sell all of their stuff, which is really cool. [inaudible 24:03] used us, Ellie Goulding. Still, the majority of our revenue comes from indie people, like random publishers out of Iowa or Idaho. Musicians from Brooklyn, or whatever.

People who just haven't had an easy way to sell their stuff online, even though they might have a crazy following on Soundcloud or what have you. That's still the majority of our following that we have is celebrities and everything on board. And that is what is exciting, helping celebrities make more money is awesome, but for me it's way more about helping people that might have a 9:00 to 5:00 job.

Even though they might want to create music, they really can't because if they make money off of music and they work at a bank, that sucks. It sucks for them.


Some people like working at banks, that's fine. It sucks if you don't want to. If that's your second choice, then you should change that.

Ryan: Yes, understandable.

Sahil: Yes, it's pretty good.

Ryan: Sweet. In terms of you, yourself are eating your dog food with Gumroad. Because you're writing a book about how you are making Gumroad. And you're using Gumroad. Why do that?

Sahil: It's like I said before. I only want to build something that I can use myself. And using your stuff means that you have this feedback loop. You figure out how crappy your product is. Because normally you don't. If you're building something for someone else, you are going to have to wait for them to tell you what's wrong with your product, and they are not going to do it properly, ever.

If you are using it yourself, the first time I tried to sell something using Gumroad, or the second time after I launched, I was like, "It's really hard and confusing and I even built this thing." Imagine all the other people. It's great that people were using us, but imagine how many more people would use us if we made it more clear, simpler, faster, more elegant, etc.

I think it's really, really important. That's why I started. AI just thought it was cool. If you have cool content, you should be able to monetize it. I thought that building Gumroad was pretty interesting, and writing a book about it would be cool. I decided to do that. I started writing a chapter a month, and I realized that writing a chapter a month is actually pretty difficult.

I've slowed that down. But I still want to do it. Every so often I will randomly decide to write something while I'm doing zero plus and seeing how it does.

Ryan: Awesome. It's almost as if you are your best user test in this case. Kind of just figuring out, "This sucks. I can't do what I want it to do."

Sahil: Yes. It's really important. It's funny. A lot of the time, we will be at an interview in engineering, and they will be like, "So, what are you guys working on the product?" My default is, "Our product is entirely shitty. Every part of it sucks. We need to make it better." That's how I think about our product and it's how I think about every product I've ever worked on, because I need to make it better.

It also turns off a lot of engineers. "Wait, the product sucks?" No, no,no. It's good, but my mindset is that it sucks because it could always be better. If you think it's good, you have less incentive to make it great. Whereas, if it sucks...

Ryan: You're just hitting the middle of the road. You're not setting your bar higher and higher in terms of that. Gotcha. Very good.

Sahil: And self-deprecating. Which is my setback.

Ryan: Right. I think that is everyone's setback. And it's our favorite here. Thank you for answering my questions. But I really want to open up to the audience. There's a lot of you here and I want to make sure everyone gets a chance to at least ask a question. We'll start off.

Audience Member:: Just a question about how you position Gum road, sharing and selling stuff. Ebay, Etsy, other marketplaces for you selling things. How would you help us differentiate or understand what you do is different, better, cooler or whatever?

Sahil: Yes. The question was, how would I differentiate from current marketplaces like eBay, Etsy, etc. I say, we are not even a marketplace. For us, our ideal goal is to turn your Twitter feed, Soundspace, or writing forum into a marketplace. There are ways on the Internet people already connect with each other, and I think marketplaces exist when there is not enough efficiency there. People need a separate place to just sell and transact.

Ideally, you have enough efficiency and you are just missing the transactional piece, and maybe the security and contacts piece, which we also provide that we can do a good job. Lady Gaga might not need to use iTunes if all of her followers connect with her on all these social platforms. That has yet to be proven, but that's the vision.

Why do you even need to sell your book on Amazon, if everyone that is every going to read your book or cares about it is following you on Twitter, or following someone who follows you on Twitter, then Twitter does a pretty good job of distributing that stuff for you.

Audience Member:: You would be more or a commerce-enabled sharing something like that.

Sahil: Yes.

Audience Member:: OK.

Ryan: Yes?

Audience Member: What recommendations do you have in terms of all the different kinds of virtual distributing? What have you seen in terms of trends for currently, what are the most sold virtual goods? What's your vision for the future? What kind of virtual goods do you think people will start creating and selling?

Sahil: Yes, totally. What virtual goods are pretty popular and which will be? Right now, on Gumroad specifically, books and eBooks, etc, are very popular. A lot of how-to guides. Music is really popular. The one that is really random to me, and thinking about it now, makes total sense, is comics. Comics just blew up on Gumroad.

You ask most people who make comics, they all know what Gumroad is because either they have used it or one of their friends has. It makes sense because comics have been serialized before TV shows were serialized. People understand that kind of consumption.

People have these rabid follower bases. If you read comics, you read comics. It's the thing you do. It's not like, "Oh, I watch TV shows" or whatever. It's pretty core.

Ryan: Yes, it's a pretty core element.

Sahil: Point proven. In the future, I was talking with a friend and he was really excited about Gumroad because he thinks that the way people are going to sell physical things are going to change. He is a big believer in 3D printing, and he thinks that in 50 years, maybe, every town is going to have their own 3D printer. Instead of buying a cup on Amazon and then they ship it to you.

The two biggest costs of selling physical things are inventory and shipping. Having the thing there and then shipping it. With 3D printing, what if you bought a digital receipt for this cup, and they sent that to you. Which is free to send, right? It's like sending an email.

Then you manufacture the thing right next to you. Remove inventory and you remove shipping, or at least make it extremely cheap. So that's pretty exciting. I have no idea about that space or anything like that, but it seems really cool. If we could help that happen, that would be amazing.

Ryan: There is one question.

Audience Member: I have a question about your design [inaudible 31:05] Pinterest. Where you guys doing a lot of this new start-up, doing
[inaudible 31:09 - 31:16]

Sahil: Yes. A Lean Start-up, iteration, etc., on Pinterest. A lot of it was done before I got there. We were actually never big believers in Lean Start-up methodology or whatever. I think some of it is great. I think most of it is obvious. You should iterate and talk to users, yes. Things like that make a lot of sense. It would be stupid if you didn't do that.

But at the end of the day, we wanted to build this product because we had a very good sense of what people wanted. Ideally, we were users of the product ourselves, so we knew what we wanted in our own product. We actually didn't do a lot of user testing and things like that.

We did do stuff where people would write us a board email and say, "Hey, how do I delete a pin?" We said, "Oh shit, we should probably make that more obvious". That's a different type. That's reactive rather than proactive. That's the best, because if you are proactive, people always have these ideas of what they want but they never actually tell you.

Like the faster horse kind of thing? If you asked people what they wanted, they would have said, "a faster horse", not a car. It's kind of like that. When you launch something, they're like, "Hey, it's broken". It's good for that. It's good for fixing stuff. But it's not good for inventing new stuff. There was not a lot of iteration. We thought a lot about how we get on a specific screen. Why do people go to this screen? What information should we expose?

Why do people care about this? Things like that. One test that I used personally was, "Will people be pissed off if we kill a certain feature?" A lot of times you get attached to something because it looks beautiful. But really, no one actually cares, and if you killed it, no one notice tomorrow.

If there's a feature where you removed it, you would get a million emails the next day, and be like, "What the fuck?" Those are the features you should probably stick around. But people don't really think of it like that.

Ryan: Yes?

Audience Member: I see [you] using Gumroad for [inaudible 33:09]
products, are you guys partnered with anyone [inaudible 33:11]?

Sahil: A lot of people have started using Gumroad for physical products. We have this one feature where we let you collect shipping information on top of email and things like that. People use that and then handle fulfillment themselves, and delivery and shipping. But we don't do a good job at all. We are focusing entirely on digital just because there are so many problems yet to be solved with that. Once we solve most of them, we can be like "OK, how can we take this and apply it to a design t-shirt, for example?" Ninety-nine percent of the code base look and feel, etc, is going to be the same. that's how we think about it. We have started talking about some fulfillment products, etc, but we haven't gotten very far.

Ryan: There is a question way in the back.

Audience Member: Do you have a strong social media presence
[inaudible 34:00]?

Sahil: Yes, right now you do. Do you need to have a social media presence to use Gumroad? Yes, we have definitely got a lot of feedback from people being like, "Hey, I don't have a following. How do I sell shit using Gumroad?" I think most people will try to figure out an answer, I just say "There's no way". Gumroad is not the fit for you.

Seriously. I mean it's not. It's like if you are musician, you are not going to sell your stuff on iTunes. Apple is not going to be like, "you can maybe spend a weekend making a song with your friend and sell that". It's not a fit. Don't try to force it.

I think that loses your product because you built it for all these use cases that don't work. You might have a marketplace that's kind of shitty, but if it's not you don't really try, if it's 90% there, it's 0% there. It's pretty binary. So we don't work with that at all. We think that the web is going to get better on servicing those people.

Inevitably, in my idealistic vision of the world, if you create cool stuff that people care about, you will get found in some way. You might have four followers on Twitter, and then someone cool sees it and re-tweets it, and now you have 40,000. Or something like that. But we'll see.

Ryan: There was another question. Yes?

Audience Member: When you came to Pinterest, how much of that
[inaudible 35:18] design was already there and how much did it change
[inaudible 35:22]?

Sahil: The overall concept of Pinterest hasn't changed since I got there. It used to be square. And then that didn't really work out, and then it was offset. We messed around constantly with changing the look of it. Once you figure out, basically the only innovation in the look is things don't have to be the same. They don't all have to be squares. They don't all have to add up to the same thing. That was it.

Audience Member: Did you come up with that decision to turn a different way [inaudible 35:58]?

Sahil: Yes. It's like reality, right? Like I said before, if you open up on the wall of a restaurant that you go to or a dorm room door, or when scrapbooking, you don't make everything perfectly square.

Once we figured that out, it was pretty clear what we built after that. Now it's everywhere because of things like the masonry plug-in and things like that. It's a lot easier to do that today than it was three years ago. Masonry? It's this thing that makes it really easy to do that offset grid look. It's what we call it internally.

Ryan: I saw your hand go up.

Audience Member: I was wondering, have you ever had challenges identifying digital rights with people who sell? If they sell a song that Lady Gaga wrote, they don't have a right to sell that, right? Two questions. One, do you have challenges around that, and how do you manage about it?

Sahil: Yes. We actually don't have as many challenges as you might think. Because you need a following to sell something, who would you sell it to? You're not really going to sell it to anybody. Anyone you talk to is going to know that you're not Lady Gaga, right? It's your Twitter account, it's your Facebook or whatever. It's very socially unacceptable to sell stuff that you don't own. It's totally socially acceptable for you to give me a Lady Gaga song for free, but the minute you charge for it, it's blasphemy.

In that context, it never happens. Literally, we have sold a tens of thousands, hundreds of thousands of things. Zero times has that happened, which is great. But people still care. Our biggest challenge is people caring, rather than people actually doing it. Especially big corporations.

They are like, "How do you handle DRM or how do you handle someone selling Lady Gaga's thing?" We're like, "We don't." They freak out. No one is going to actually do that. Now Sony uses us. Universal uses us. All these people got over it. iTunes is doing a good job. All the stuff you download on iTunes is DR-free now.

They paved the way for that pretty well. I think it's a problem. Our goal is to make stuff easy to buy, not hard to steal. That's the future of the Internet. If you don't like that, stop using the Internet.


Audience Member: [inaudible 38:02] hear about it. If I have an account on Gumroad and I get on Lady Gaga's Facebook page and she posts a song, can I download the song right there on Facebook and not have to go through Gumroad?

Sahil: Almost. Unfortunately, Facebook has restrictions on that. So you click through, and this very simple landing page, which looks almost like a Tumblr [inaudible 38:44] or something, and then you can transact with it in 30 seconds. It still requires a click-
through out. We are trying to work on making it in-stream 100%, but that will require Facebook's blessing, which is hard to get.

Ryan: It's like you have to go to the mountain and ask for it, right?

Sahil: Totally.

Ryan: Anyone else with a question? Yes?

Audience Member: Regarding the [inaudible 39:10] masonry Pinterest and earlier, you guys were talking about how things on Facebook but also MySpace had copied that.

Sahil: Yes.

Audience Member: Since it's really just like a tool, anybody can use
[inaudible 39:23] layout. Would you really consider that being a copy of Pinterest then or [inaudible 39:34]? Like in the future, if somebody copies how Gumroad has the social e-commerce system and you go to buy it
[inaudible 39:49] pay for it, would you think this is copying or kind of this inspiring [inaudible 39:54]?

Sahil: I don't think it's copying at all, no. I use that because everyone else says it. I was never pissed off that people copied any of the things that I built. I was more flattered than anything else. Sometimes it's annoying if people fundamentally copy everything, you know? Like, clones are very different.

Pinspire, for example. Or Sex.com.


Don't go there. But if you do go there, it's really funny. That's different. But yes, if you copied an idea, no. I think it's great.

Ryan: Anyone else with a question?

Audience Member: Gumroad [inaudible 40:45]?

Sahil: I actually bought the domain a year and a half before I built Gumroad. It was just a domain I had at GoDaddy. I wanted to launch the thing really badly and I didn't want to wait to find a name, so I used it. It fit pretty well.

Audience Member: How did you come up with it?

Sahil: My mom did actually. I wanted a vague name so I could use it for some random thing, and it worked out.


But I wanted a name that was generic enough and implied fun with gum, and some sort of connection to movement. Road. We were looking at things like Gumpath. That was taken, so we went with Gumroad and then it turned into this.

Ryan: We have time for about one more question. Someone with some really good, pressing question in the corner over there?

Audience Member: What kind of technologies are you using right now?
[inaudible 41:39]?

Sahil: Yes. Our stuff is built on Ruby Rails. My SQL is primary data store. We use things like [Reddis, Mongo, and MCash]. We use all that standard stuff nowadays. It's all on AWS. Used S3 for storing files. What else? What other [inaudible 42:00] do we use? GetHub
[inaudible 42:05]. We jQuery for JS. Nothing too fancy. It scales well. If things don't break over time, that means you over-optimized at the beginning. I'm sure things will break. Things have broken.

We'll hopefully get to them. It's a little different with us, because a lot of times we will get a mass of traffic, like when Liz Cleaver tweeted about us, we got this absurd spike. It was not on TV, it's kind of this, right?
But it was a tweet from this big celebrity. You can't really plan for that. Or, you don't think about how to plan for that until it happens. Now we've gotten better at that. But there are problems like that we try to solve.

Agent X, for example, is an amazing web server for dealing with a shitload of requests. Ruby, as an app, is not. So what you do is tell AgentX to deal with as many as it can, and then any ones that need to be dynamic for example, [inaudible 43:10] uses Ruby at the center. So you use page caching for that, which we do. There are things like that. You just learn them as you go and hope you don't go down too much.

Ryan: OK. The last question, right over there.

Audience Member: Yes, a question about user [inaudible 43:26]
company. How are you getting around your [inaudible 43:30] out from the very beginning? Have you done any marketing or have you sold [inaudible 43:34]?

Sahil: Yes. We have done a bunch of press around funding, things like that, automatically. We have that. That was our initial stuff. Then we do a lot of manual. Reach out to label agencies, talent guys, managers, agents, etc directly. So that's how we got ways for
[inaudible 43:59].

But in terms of actual scale, I think that needs to be word of mouth. Some sort of thing that's just inherent in your product. Otherwise it's really hard to grow. Pinterest, we didn't have share buttons on our site until we were hitting ridiculous numbers. Part of it was laziness.

Two is because adding Share bottons to your site is not going to make it explode. Your site has something really interesting about it that causes it to do really well, and then Sharebot might amplify that to some extent. It's not going to make your site succeed or fail.

I have a friend who works on this thing called Supply, which is a blog network, and he refuses to put tweet buttons on it because he thinks good blog posts should spread automatically. If a blog post is not good enough for you to manually copy, paste, and tweet about, it's not that good. Or doesn't deserve to be tweeted, and just because it's on the way now, you're not going to do it. That's his belief.

So yes. It's just built into our product. If you want to sell something, you need a following. You tweet it out to your 20,000 followers, now 20,000 people might know what it is. Hopefully some proportion of them will be like, "Oh, this is really cool! I could use this to sell something". Or,
"I'm going to buy this and then tweet about the fact that I bought this".

That's why I spent a lot of time on our flows just making sure that people understand what we do pretty quickly because I think that's pretty critical in our success. But that works. We have yet to do any sort of paid marketing or something like that. We have millions of venues.

Ryan: Sweet, very good. Well, we're just about at time. Thank you to Sahil for coming down and joining us at ZURBSoapbox.

Just to let you guys know, we have Joe Gebbia from Airbnb next Friday.

So, come on back, same time, same bat channel, so to speak. ZURB channel. All right, have a good one. Take care, everyone.

Learn from Past Influences