Woman: From the campus at the Wharton School in San Francisco, this is Startup School Radio. Here is Y Combinator partner, Aaron Harris.
Aaron: I'm a partner at Y Combinator, and this is Startup School Radio live from Wharton's campus in San Francisco on Sirius XM's business radio powered by the Wharton School. At Y Combinator, we fund early stage companies and work with them to make them into billion dollar businesses. You've heard of some of our companies like Airbnb and Dropbox. And others, you'll hear about soon. Every year, Y Combinator does a conference called startup school where we bring on early stage founders, late stage founders, CEO's and really amazing people that we love, and have them tell their stories, talk about everything they've learned, the screw ups and the successes.
Here on Startup School Radio, we'll be bringing those founders to you on a weekly basis, broadcasting every Wednesday at one p.m. eastern. 10: a.m. Pacific just trying to help you think about starting a company or learning how to do it better.
Today, I'm honored to have Optimizely's Pete Koomen. You've probably heard of Optimizely if you have a website. And if you don't, you really should know about them. They help you make your website better, make your users happier, and help you do more business through the magic of A/B testing. We'll talk about what that is and how Pete decided to do that, and what they've done since they started in just a second. Later on today, we'll hear as well from CEO and Founder of Lawn Love, Jeremy Yamaguchi, about how he's making the process of finding a great lawn care provider, pain free.
If you're thinking of starting a company, you're an entrepreneur in the early stage running a start up, or just generally have some questions for us, we'd love to hear from you. Our number is 844-Wharton. That's 844-942-7866. Be sure to follow us on Twitter @bizradio111 or you can follow me at @harris. So without further ado, let's turn it over to Pete. Pete, welcome to Startup School Radio.
Pete: Thanks, Aaron.
Aaron: So you've been at this startup game for a little while.
Pete: Yeah, almost six years to the day, actually.
Aaron: Oh, six years to the day. Is six years ago, is that when you left Google?
Pete: So six years ago is when I left Google. And my co founder and I started working in late 2008, we were both moonlighting. So that would make it more than six years now. And we had two startups which never really made it off the ground during that first year, and then started Optimizely actually midway through our Y Combinator batch in 2010.
Aaron: You had two failed startups before you got to Optimizely?
Pete: Yup. We definitely learned the hard way there.
Aaron: So lot of people look at a company like Optimizely and founders like you and Dan, and say, "Wow, these guys have gotten a lot of press. Their company has all these people, nice big office. There's no way they've ever failed at anything." So what happened with the first two and how did you keep going through it? What were the first two? Let's start there.
Pete: Yeah, it's a good question. And they actually are pretty unrelated to what we ended up working on with Optimizely. So we started working together, as I mentioned, really late 2008, both moonlighting at the time. And we started working on something that was totally unrelated. It was a startup around building an educational platform for young kids. So we wanted to make it easy for young kids to learn math on the internet. And we figure we like math, we like computers. Probably all kids are the same way, and so we build a, you can tell.
Aaron: By the way making the job, we like math, all kids like math, pretty big jump.
Pete: Yeah, it sounds a little silly but it's actually I don't think that uncommon among pretty early founders. So we actually had a third co-founder as well. We started this really in earnest in early 2009. It was called CarrotSticks. It was an amazing experience. We worked on it for probably about nine months. It evolved a lot during that time. It started out just as a way to practice math on the computer, which it turns out didn't really excite kids too much. We learned a lot. We ended up changing it into a game where kids could compete to solve the biggest number of math problems in 30 seconds, and they ended up actually liking that.
But we realized by the end of it that, first of all, building an education startup is really, really hard. That space is littered with the carcasses of ambitious companies. And we also, we weren't parents, we weren't teachers, we weren't kids, we weren't administrators, we weren't close enough to the problem to really solve it in the way that we thought would lead to a big, big business.
Aaron: So I want to pick up on the first thing you said. The second part is super interesting with this idea that you weren't the target audience, you really didn't really know enough about it. But the first thing you say is educational technology is really hard. A lot of ambitious startups, and I think you said lying on the floor or litter the way or something like that. Isn't that true of all startups? Isn't that all really hard?
Pete: It is all really hard. And that's something we learn just over and over and over again. It's really hard. And that alone, the fact that education is the difficult market to build in, would never been enough, I think, to dissuade us from working on it, but it was that combined with the fact that as the more we learned about our customers, the further apart we realized that we were from them and their experiences. And we thought maybe there is a great business here. Maybe there is a way to have a big impact. We're probably not the best people to have that impact, and so we decided to move on.
Aaron: So that's an interesting remark to say, "Okay, this is going to be hard, but this is probably not the place to put that effort, right? We are not really going to be able to understand it." So how do you get from education technology over to Optimizely and say, "This is the thing where we can have... Well first, let's explain what Optimizely is. Because, I think, in San Francisco, you say the term A/B Testing. And I was like, "Oh, yeah. Of course. That makes sense." But what are we actually talking about here?
Pete: Right. So A/B testing is not a new concept. It's something that's actually been around for a long time, and I will explain what it is. So A/B testing is a way to make your website better by testing your audience, by testing your visitors. And it doesn't only apply to websites. It can apply to mobile applications or just about anything else that a business presents to its customers. But the idea is that you've got a website. You have lots of people landing on it. You want them to do something typically. So if you're a political candidate, you probably want donations or volunteers. Or if you have an e-commerce website, you want people to buy stuff or you want people to sign up for an account.
So everybody's got some goal or some set of goals for their website. And when you run an A/B test, you just design multiple versions to that website, and then every visitor gets one of those versions randomly. And you measure which of those versions gets more people to sign up or more people to buy or to donate. So is it's this idea of doing a random controlled study in real time on your own website visitors or your own app users, etc.
Aaron: It sounds pretty similar what you might do in advertising on paper. What's been done all along is like, "Hey, let's try this ad copy in the newspaper and in ad in the paper, and let's try this ad copy and see what drives more sales."
Pete: Exactly, And actually, this goes back to the days of catalog sales when companies would produce multiple versions of their catalog and mail them out to different parts of the country and then measure which version of that catalog got more people to call in and buy things.
Aaron: Oh, I had no idea that catalog makers used to do that.
Pete: Yup, it's an offline practice. It goes way back. It's pretty interesting.
Aaron: How much were they actually able to measure and isolate what variables were making a difference in that scenario?
Pete: I wish I could answer that question. I don't know the answer to it.
Aaron: Because as you think about this, if you're a test, you have to isolate the thing that you changed that actually made the difference. And if it's a multi-hundred page catalog, you might change a lot of things and you have no idea if the customer actually saw that. So I guess on the web, it's a different story, right?
Pete: So on the web, we are a lot more able to measure things that are important to us. And as I mentioned before, every business or organization has their own goal that's important to them. It might be sign ups. It might be purchases. It might be revenue. It might time on site or number of articles read or sign ups in an app. There are whole bunch of different things that people can measure. I think back in the day, there are probably cruder ways of doing it. They probably ran simpler experiments. Nowadays, with modern A/B testing software, you can run large multivariate experiments where you're testing many things at once and trying to understand the interactions between them.
Aaron: That's interesting. So you can actually test multiple things at once. But if you have a large enough sample, then you can start seeing the interrelation between those things, and actually, which ones are more important.
Pete: That's exactly right. Those are called multivariate tests.
Aaron: OK, that's pretty cool. So when you guys had this idea, I think you mentioned that it was possible to do A/B testing before you were there. People were doing different types of it. But were there other companies doing A/B testing online before Optimizely showed up?
Pete: There were absolutely were other companies doing it online. They typically tended to be much larger companies like Google or Amazon or Netflix are famous for this kind of testing. What we did is we made it possible for anybody to run an A/B test. We made it much, much easier to do this. And I'll tell you where the idea came from. So we were really struggling during that first year. We started this company CarrotSticks, and that didn't really work. And we started at a second company right after that, and that failed much faster.
Aaron: What was the second one?
Pete: I won't even go into it, to be honest. We were at least a little closer to the problem. We realized with CarrotSticks that the easy part was building the technology. We didn't realize that at first. But the easy part was building the technology, the hard part was getting customers. And so our second company, we just tried to build a way to help other businesses get customers. And without going into what it is, we realized after about two months that that wasn't gonna work either. And we moved on again.
And what we settled on was something that came out of my co-founder Dan's experience on the Barrack Obama presidential campaign in 2008. And so let's rewind a little bit. We were both working at Google. We'd gotten to know each other very well there. We both had a real strong interest in entrepreneurship, and we were actually ready to jump ship in 2007, late 2007.
Aaron: How long had you been at Google at that point?
Pete: I'd been there about a year and a half; Dan, just about a year. And it felt a little early, but we were ready to go. We were excited. And then Barrack Obama came and visited the campus. And I remember going to see his talk and thinking, "Wow, that was neat." And Dan saw the same talk and it was just absolutely flawed. Quit his job three week later, moved to Chicago, became the director of Analytics on the Obama campaign. And over the next nine months, everyone knows what happens there. But over the next nine months, Dan and his team ran a whole bunch of A/B tests for the campaign. And they had a huge impact. They raised tens of millions of dollars and additional donation?
Aaron: Had anyone done that before online for presidential campaign?
Pete: I think this is a pretty new way of looking at it. When he jokes, they lumped his team into the new media team, which was just their way of saying everything that we don't really understand get's lumped into there. And so anyway, the takeaway really was that this is a very powerful technique, but it was also really tough to do. It took a lot of work for them to run these experiments. And it's funny, after the presidential campaign, I think Dan moved to Washington and quickly decided that he preferred the faster, passive doing things. And so [inaudible 00:12:11] and came back here, and that's when I left.
Aaron: How long did he last in Washington?
Pete: I think it was just during the transition time.
Aaron: Okay.
Pete: Yeah, before the president took office. And came back here, and that's when we started working together. And it wasn't until just about a year later after two failures that Dan, I think, started thinking about this whole A/B testing thing. Again, this is really powerful, but it was really hard to do. I wonder if there is an easier way to do it.
Aaron: That's crazy though. He had this really powerful experience with how powerful this was, and then completely forgot about it basically. It was like, "Oh, let's go do math problems for kids."
Pete: Yeah. I'm sure there are plenty of sayings about good ideas, and how sometimes, it's the most obvious one right in front of your nose that ended up being the best. But I guess you could call it maybe a mix of just burnout and excitement for our original direction. But it wasn't until we'd realized that, "Hey, we really weren't as great as we thought we were," and B, that the stuff was really tough that went searching around for a problem that we could really prove to ourselves that someone would pay for. So that was really big.
With CarrotSticks, it took us about six months to earn our first dollar. With our second idea, it took a month and a half. And with Optimizely, we're able to earn revenue on day one before we'd written any other code. And that illustrates how we changed our thinking to focus on making sure we found something people wanted before we built it.
Aaron: If you're just joining us, Pete Koomen from Optimizely is telling us how they came up with the idea of Optimizely out of all the ideas they had. This is Startup School Radio. If you want to ask us some questions, please give us a ring. You can reach us at 844-Wharton. That's 844-942-7866. So Pete, the last thing that you just said was we needed to make something people actually wanted, which is the model of Y Combinator. It's on all over our shirts. It just makes something people want. And I think it's a really tough idea for a lot of people to actually internalize, because they can find things that maybe they want. "I want a puppy." But that might not be a startup.
I think you gave me the answer. The way that you actually figure out something that people want, that's actually a business. It's that they are willing to pay for.
Pete: That was our proxy. Because we need up building a business that sold things to people for actual money, I think, unlike a lot of tech startups, we had an easy proxy there for demand. It was just trying to get someone to pay for it.
Aaron: How long did it take for you to get someone to pay for it?
Pete: It was a single day, actually, with Optimizely.
Aaron: One day?
Pete: Dan called a couple of agencies that he'd worked with on the Obama campaign and just described, "This is what we are working on. Will you pay us $1,000 a month for access to early versions of it?" And sure enough, two of them signed up. And we earned more revenue that one day than we had the entire preceding year with our two startups.
Aaron: Wow! Did you have the software yet?
Pete: We had nothing.
Aaron: You had nothing?
Pete: We had nothing.
Aaron: So you sold a promise of future software?
Pete: We sold a promise of future software.
Aaron: Would that have been possible, do you think, if Dan didn't already have personal relationships with these agencies?
Pete: I'm not sure. So on one hand, I'm sure the relations helped, and they certainly were open avenues that we knew exactly who to call for a customer. And that's just finding a customer can be difficult. But at the same time, I think that our growth after we launched signaled that we'd found a need that was real and burning for a lot of customers or a lot of companies out there. And we spread for about two years after we launched. We spread purely on word of mouth alone. We didn't spend anything on marketing. I think that's a testament to how big this problem that we stumbled on was.
Aaron: I want to go from that and take a question from Terry in Kansas who's looking for advice on coming up with startup idea when that idea is already out there, which is what you went through. It's the idea of A/B testing existed but you did something with it. So Terry, you're on the air. Let us know what you're thinking about.
Terry: Yes. So Pete mentioned it earlier when you say you would let that just dissuade him. But I've been working with a couple of bodies for a while now on a financial service idea. And then all of the sudden, we started seeing multiple companies do somewhere. And it's made me think, "Well, we don't know how to program. We don't have these resources these larger companies have, why even bother?" And how do you overcome that? And then if you don't have the resources, how do you get them?
Pete: That's a great question. This whole question of, "Am I barking at the wrong tree?" is something that I think every startup founder faces at least once, probably multiple times. And I don't know that there is any formula for deciding that what you're working on is not actually the right thing. Because probably, the most essential ingredient in any founder that I've seen is perseverance and the ability to ignore that the notion that what they are doing is crazy or stupid. And so it's this real, it's a weird contradiction. On one hand, yeah, what you're doing is ridiculous, and that you need to pay attention enough to reality to know that. On the other, hand you have to believe it's possible. Otherwise, you never would have gotten anywhere in the first place.
So I don't know the answer to your fist question. I think that probably, I would suggest that this ought to comes from your customers. If you really believe that you're trying to solve a problem that people have and are willing to pay for, and you've talked to people and you understand that that is a burning need of theirs, then I would say that that's fertile ground for something to grow out of. I think our first two attempts were really building things were there's no burning need for. As Aaron said, it was our puppy. It was something we wanted or we liked but it didn't really match what people were willing to pay for or really wanted. And so if you can go out there and you can find people that are willing to pay for what you're doing, then I'd say you're on the right track somewhere.
As for finding resources, that's another tough one. Probably the second hardest thing after just finding a need is hiring. And what we ended up doing in the beginning was with our early employees, we weren't able to pay them much money. In fact, we paid them a lot less than they were making elsewhere. But we offered them a pretty meaningful stake in the business. And you and your co-founders really need to be on board with that if you're going to go for it. But it's a great way to build a small team that's all pointed in the same direction because you are all incented by the same thing, which is making this a success.
Aaron: Yeah. One of the things that we tell companies a lot is not to really think about your competition. Because at the end of the day competitor aren't what kill startups. It's more either the apathy of the market, the fact that no one really cares or knows about what you're doing or you simply run out of time. There constraints that happen when you try to start a company. You have X amount of dollars in the bank that can last you X number of months without having a salary, or people are willing to give you a certain amount of time for the equity that Pete is talking about before that they need a salary.
And if you can't make an impact, if you can't prove that people really want what you're doing in that time, it becomes really, really hard and increasingly difficult. And you have to look at those signals that at some point reach a decision that says, "You know what? The signals are pointed strongly enough in direction that I have conviction that this idea is not the right thing to pursue."
So Terry, thanks for calling in with the question, really great question. If anyone else wants to call in and chat, we'll try to get you on. Our number is 844-942-7866. So Pete, you just said something that I think was one of the coolest parts of the Optimizely story, which is how you made your first hire. So we'll skip a little bit. You get your first couple of thousands in sales. You start building the software where you guys do Y Combinator, then you need to hire. Because all of the sudden, you know you have this incredibly powerful thing on your hands, but you don't have enough time to do everything between the two of you. So you're selling to pretty large companies. So there is a model here on how you're supposed to hire people to sell to companies, which is hire people with experience in a rolodex. Did you do that?
Pete: We tried that. It didn't really work out. So first of all, I guess I should say our first real sales hires were Dan and myself. Neither of us has any formal sale training, but we really believe in what we were doing, and that I think was enough in a lot of case. We were able to close deals, significant deals with customers. And having that connection, forcing ourselves to sell the product early on, was I think one of the more important things we did. It really forced us to get as close as we could to the customer need.
But we realized that just like anything else we were doing, we eventually would have to find somebody who is better at it than we were, which in most cases is not too hard to do. And it's hard to convince them to come on board. It's not hard to find someone who's actually better than we were on most of the stuff.
Aaron: I think you're selling yourself a little short on that one.
Pete: In any case, we found somebody who was working for a big competitor of ours, and they'd had a lot of sales experience. And we hired him. We were very excited about it. And I think we made several mistakes there. This was someone who is remote, and so we weren't really connected with him the way we should have been. This customer to founder or engineer connection is so important earlier on. We didn't have that. We broke that by doing this.
This person was someone, I think, who was used to working with a lot of resources. They had a team around them. They had the power of a large brand. They weren't used to being in an environment where no one would really take their calls, which is what you're in when you're starting a company. No one knows who you are. And so that didn't work. And for a while, we weren't really sure that anybody else could sell Optimizely. We were the founders. We were also engineers. We knew the product in and out. We could do it, but is it possible to scale that model? We didn't know.
Aaron: It's got to be a pretty scary moment?
Pete: It is a scary moment. We tried again. Eventually, we got a referral from one of our investors. And this one was pretty unorthodox. This was the cross buddy who played with one of our investor's associates. And he was running a struggling mattress company in Seattle. His name is Matt. He's amazing. He came down and he said, "I want to get in to tech. I'm willing to do anything to prove this to you guys." And so we were skeptical. And we said, "All right. Well, we are just going to start with the 12-week contract. We'll see what you can do." And I remember listening to his first customer calls and just wincing through them.
Aaron: Was he terrible or was he just not what you wanted?
Pete: I think maybe a combination of both, to be honest. But he had so much hustle. And he learned so fast, that eventually, he was doing much better than we ever had. And he really built the way to sell Optimizely. He hired someone else who was a retired ballet dancer, who was not the grossing sales person at Optimizely, I think, in our history. Matt, that first sales person who worked well, now runs our entire European operation. We really found folks that a lot of run way, a lot of potential.
Aaron: Which is exactly what you look for in founders themselves. You don't necessarily need people who have years of industry experience. It's more the ability to rapidly come up a curve on learning and just the ability to go through brick walls again and again and again, which is what you and Dan had done through crashing your first two startups before Optimizely.
Pete: Essentially, yeah. It's that hustle, that ability to... You identify something wrong, you just find a way to get around it. And that's what Matt did.
Aaron: Yeah. You're almost A/B testing. You're hiring.
Pete: One could say that.
Aaron: Good. All right. So this idea, just to come back to this A/B testing thing, I can see how this is useful for large corporations with website that have millions and millions of people coming to their sites, and they have the knowledge or in-house team that's super knowledgeable about statistics and can figure what is meaningful. Is that the focus of Optimizely? Is it useful for other people? If I have a square space page and want to make sure people buy more of my beef jerky or something like that, can I use Optimizely?
Pete: So you can use Optimizely. We have plans that range from free that anybody can use up to plans that support much further.
Aaron: You give away your product for free?
Pete: We do give away our product for free. And Optimizely is useful whether or not you have a smaller website. There's probably a minimum traffic level. If you are a startup founder, you have just a trickle of customers, I wouldn't recommend investing in A/B testing. I think you're probably wasting your time at that stage. It's probably a better idea just to get in front of your users and understand face to face what's going on there. But once you pass a certain level, I think A/B testing is a very useful way of trying your hypothesis and making decisions. It's also useful for folks who have apps.
There's so much focus on trying to get people onto a website or trying to get them into an app. There is about a zillion ways to advertise out there to bring traffic in. But where I think most companies have trouble is converting those visitors into customers or converting those users into repeat app users. Engagement is very difficult on mobile applications, getting folks to sign up. Same deal on websites. And that's where Optimizely focuses.
Aaron: And as the world moves increasingly to mobile, we all use, I think, our mobile phones more than we use our laptops or our computers anymore these days. That is probably increasing to the forefront of what people really have to be thinking about in terms of optimization.
Pete: Absolutely. I think so. The app store is an extremely competitive place on both iOS and Android. This is one way for app developers to really understand empirically what works for their users and what doesn't?
Aaron: I wish we had enough time to dig in to what you've seen in terms of changes in traffic and differences in how people should A/B test mobile versus web. Not only are you testing with inner mobile app, what should change, but your mobile against your web. It's a whole... I don't know. The possibilities are mind-boggling and really confusing.
Anyway, Pete, I just want to say thanks for joining us at the top of the show. I hope you can stick around for the rest. We have Jeremy Yamaguchi joining us in just a few minutes. We are going to go to break shortly here. I want to thank everyone for tuning in so far today. You're listening to Startup School Radio on Sirius XM's Business Radio powered by the Wharton School. We are channel 110. And you can follow us @bizradio111 on Twitter, or follow me @harris. This is Startup School Radio. Sorry, it's actually @bizradio111, not 110. I misspoke. I'm Aaron Harris, and this is Startup School Radio. We'll be back with you in just a few minutes.
Woman: You're listening to Startup School Radio powered by the Wharton School. Here again is Aaron Harris.
Aaron: Welcome back to Startup School Radio on Business Radio powered by the Wharton School, Sirius XM channel 111. I'm your host, Aaron Harris. I'm a partner at Y Combinator. I've been speaking for the last half hour with Optimizely co-founder and CTO, Pete Koomen. Pete, thanks for sticking around for the send half hour.
Pete: Thanks, Aaron.
Aaron: And now, we are joined by my next guest, Jeremy Yamaguchi, founder and CEO of Lawn Love. Jeremy, you and Lawn Love were part of Y Combinator summer 2014 batch. So you've been out for about six months now?
Jeremy: Yup, that's correct.
Aaron: And you've been described in the past as "Homejoy for our lawn care." I think the press used that once or twice. So thanks for joining me today. And why don't you tell us first a little bit about what Lawn Love does, and then we are going to go back in history. And while you're listening to Jeremy, if you have questions or thoughts about starting a company or you're an entrepreneur in the early stages of running a startup and have a question for any of us, we'd like to hear from you. Our number is 844-Wharton. That's 844-942-7866. So Jeremy, what the heck is Lawn Love?
Jeremy: Yeah. So Lawn Love is an on demand lawn care marketplace. We like to think of ourselves as Homejoy for everything outside the walls. We often or also commonly applied the Uber-for-X moniker, although that's not as relevant as something like Homejoy handy. We're not nearly as location-centric or on demand or something like Uber. But yeah, we're marketplace appears, qualified, pre-screened long perils with consumers.
Aaron: That's a super easy way to explain it. It's funny. You said, sometimes people use this for that example of how to define it, which can be useful. It packs a lot of information into a small phrase, but it can also be really misleading a lot of the time or just not really get the new ones of what makes a startup great.
Jeremy: Yeah. I recently heard of something that was, I believe it was Uber for real estate. And that's really hard to understand.
Aaron: You push a button and a house comes to you.
Jeremy: Yeah, something of that nature. It teleports you to a house.
Aaron: That would be worth a lot.
Pete: I'd pay for that.
Aaron: Yeah, I'd pay a lot of money for that, just teleportation. So how did you decide start Lawn Love? Is this is the first thing you've ever done as professional or were doing it since you were three years old?
Jeremy: No. Actually, I don't have a super deep body of experience in the lawn care space. I'm not sure I'm not. Plenty of it belongs to the kid. The lawn of genesis story isn't one of these super romantic ones where I felt some profound pain and I needed to go out and solve it. It actually came out as a by-product of another startup that I built and sold in a very similar vertical.
Aaron: Okay.
Jeremy: It was called Golden Shine. It was literally Homejoy just three years before Homejoy. [inaudible 00:30:26] boots strapped. March 2009, it was an on-demand household service marketplace that did have a more romantic story where I felt this acute pain of needing to hire the housekeeper to come clean up my terribly messy home. And not being able to see an easy solution there, I decided to go and build one. And that's how I came about with this industry vertical and Homejoy. I'll say Lawn Love is the natural by-product of that.
Aaron: So what year did you start Golden Shine was it?
Jeremy: That was '09.
Aaron: '09. So the internet was a pretty different place. I guess that's Pete right around when you were starting Optimizely?
Pete: Yeah, it was the Wild West back then.
Aaron: It was crazy. Mobile phones were there but not really the dominant force they are today.
Jeremy: Sure. Uber-for-X wasn't a thing.
Aaron: Uber wasn't a thing. It just didn't even make any sense. So how did you build traffic to Golden Shine first? Where did you get customers?
Jeremy: So we actually have a fair amount of experience in SEM and SEO.
Aaron: Okay, learn SEM and SEO. Can you explain them a little bit?
Jeremy: Sure. So SEO being Search Engine Optimization and SEM being the paid variant of that.
Aaron: Got it. So people go to Google. They search for housekeeper or/and now, in Lawn Love's case, gardener.
Jeremy: Yeah. So lawn care gardening, things of that nature. So our background, to take it a step back further, is prior to Golden Shine, I started and run a web dev consultancy with a lot of home and marketing things like that. So we had a fair amount of experience in this space and we're able to rank really well for key terms specific to what services we are providing. So that's how we built a lot of the initial demand gen. And then there was also the supply side of the marketplace where we would get in front of consumers, acquire pros and knock yourselves with maids and carpet cleaners and window washers to sign up for the platform and render these actual services. And that was also predominantly through [inaudible 00:32:16] actually in the early days.
Aaron: Yeah, because you have this challenge where as Pete built a piece of software and then went out and sold it, which is just super easy, you had to go and get. You had to build software, then you had to find service providers, then you had to find customers.
Jeremy: Sure. Yeah. So this business has a unique set of challenges in that we're not selling uniform widgets that come off in assembly line and are largely the same and easy to control. We're working with human capital. And humans by virtue of being human beings can have all sorts of unique characteristics and can be challenging. So being able to build a business that works well with people of all walks from all walks and of all shapes and sizes is really critical to building a scalable company in this space.
Aaron: So how do you use software to actually start systematizing that or feeding people into the repeatable pattern that you need?
Jeremy: So you build pattern matching tools where you take a look at what your best possible service provider looks like, and you try to model around that so that the other people that you bring on board and new hires have those same characteristic and they're likely to succeed. So that's one useful way you can do it. On the customer's side, of course, we take customers regardless of what they look like whether they're great or not. Thankfully, they almost uniformly are always great. But yeah. There are a lot of things you can do to try to bring software to enable these otherwise largely human efforts.
Aaron: Right. So do you have an ideal model of what the ideal lawn care provider looks like?
Jeremy: Sure. Yeah. It's someone who, in our case, because we tend to work really well with individual solo providers or people who have one or two individuals working under them, small crews. Our value prop isn't as compelling for large scale 20 to 50 person teams. So we tend to look for people; smaller groups who have deep, longer experience. They're genuinely skilled at the actual work and can render the service well. And then the next step is just to look at how well they communicate. Is it easy to communicate ideas and needs to them? Are they responsive?
One of the challenges here is that sometimes, there's a bit of a language barrier. And we just have baseline criteria that you need to be able to speak decent enough English that customers can communicate their needs and that we can as well.
Aaron: Right. How long did it take you to figure out what that ideal person looks like. Was it just right off the back you knew or did you have to A/B test the kind of person you needed?
Jeremy: I would say we A/B test it indeed. I would say around about 87% of what that should look like within you first week of working with people, and then you refine and iterate from there. It's more of a soft thing. It's not a rigorous A/B or multivariate processes on this one.
Aaron: Pete, would you say that that is fairly accurate even on the web? Can you get most of the benefit of optimizing a site or a flow for a customer quickly, and then the last 20% takes a lot longer?
Pete: It's sometimes like that. We try to push our customers to focus on the low-hanging fruit. Because they're usually, when you've got a new site or a new page or design and it's totally unoptimized, there's usually a lot of easy things that you can tweak to understand first before setting up. We talked about multivariate testing earlier. That's really complicated. Usually, it's best to start with simple experiments. Try the resources that use your gut or intuition about what might work.
Aaron: What I love about this interplay here is what we're seeing is that the ideas around what you do on the web and what you do on real life and how those things meet up are really similar. All these things play out against both places. And I think when I put it in the framework in my head, it's all really just about this idea of iteration towards the thing that works for your business whether that's what your customer wants or what your customer needs or what makes your business make sense.
I'm curious, for you guys, it's scaling. Once we finally found this product/market fit, scaling has been just a constant challenge. And for you, it must have been doubly so because you are trying to build two sides of a marketplace at once. And I imagine you probably started out doing it pretty manually. How did you keep up and how did you decide where to invest?
Jeremy: So scaling is and will continue to be perpetually a challenge. Early on, the marketplace just tend to have strong chicken and egg problems where when you launch in your market, it's very difficult to... You have to set up a large body of service providers first in order to be able to turn on the demand side. The way that we built this company does circumvent that to a certain extent where by virtue of being a full stock marketplace, we don't actually present the user or our customers with a long list of service providers that they then select from. We think that it's a much better experience if they just push a button and get a service provider who is well-qualified, is going to do great work, and is reasonably affordable.
Aaron: So this is interesting. Because when I hear the term marketplace, I normally think of, "Oh well, I'll have all these choices in the market and I can choose based on my perception of quality and price." So why don't you do that?
Jeremy: Well, our core premise or assumption, I should say, is that we think the choice is actually quite painful. So users don't want choice. They want baseline expectations to be met. So in the same way that Uber operates, you don't want to be able to pick your driver on the basis of their hair color or what car they're driving. You just want to know that their hair is reasonably clean and doesn't stink and that their car is officially new, and that's all you need. And then after that, the challenge isn't like the value prop is how fast can I get to your car and how affordably can you travel?
Aaron: So did you have that? Was that idea fully formed when you started? Did you start somewhere else?
Jeremy: No. Actually, that was fully formed and probably a by-product of the prior company that I built.
Aaron: Okay. So let's talk about that, because it's fully formed for Lawn Love, sure. But that's a process to learn how to get to that point, right? So where did you start on this idea of the match-up and why did you move away from it over time?
Jeremy: So we actually started there and have stayed there for these reasons. But I guess just to back into it, it was probably just a matter of looking at the existing experiences. So even though Uber didn't exist at the time and we didn't have these other comparative services, it was really a look at what is it like when this works really well. And customers like to be able to just call someone who is reliable. At that point, in the household services space, the leading player was Molly maids and Merry maids. And they don't have this whole marketplace matching up system either. They allow you to just call a phone number, order service, someone shows up.
Part of the value prop is consistency and reasonable affordability. And that was, I think, the initial leading indicator for us, and we modeled around that. And then it turns out that that's actually a really good way to build a marketplace because it allows you to scale across geographies without having these deep chicken and egg problems that are super hard to overcome.
Aaron: I know that we got this totally wrong in my startup initially. We were also building a services marketplace around education technology to our earlier point about the carcasses of brave ideas. But we had this idea that we actually wanted to provide all that choice. We wanted to say, "Hey, you can select the tutor with the 4.0 GPA from the best school and they'll charge you for a $400 or you can choose the person with the 2.0 GPA and no experience and your subject matter for 15." And we thought that was a really good idea, and we went up to the market with it. And while some customers were really happy, what we found was that that repeatability thing that Jeremy, that you were talking about, just wasn't there. Every single discussion was a discussion, right?
And it was a constant haggling, and users were confused. And so we through this process of iterating on that idea. And it took us a long time to figure out, "No. The right way to do this is to actually just make the match," to say, "Look, we have this database of really good people, but it's hard for you to chose because you don't actually understand. Most users hitting the site don't actually understand the difference. What is an incremental point of GPA or a better degree worth in dollar terms. Nobody knows that, so you have to set that pricing yourself.
Jeremy: Yeah. I think the amount of choice that a user wants is directly proportionate to the intimacy or a personal nature of the service. So if you're having someone teach your child, it's not surprising that people are deeply concerned with exactly who that person is and much more concerned than they would who is outside caring for my grass. In that scenario, so as long as the service is great.
Aaron: Well, some people have a healthy relationship with their grass.
Jeremy: Usually, they do it themselves, in that case. But yeah. So I think you had a particularly strong challenge there, because of the one personal nature of the marketplace you were building.
Pete: Yeah. I wish we hadn't gone after something so personal. The challenges were so massive for us in getting there. Lawn care, it's not as personal. I never quite thought about it in that framework, but it's like a depth of personal thing.
Jeremy: Yeah. Homogenous supplier is also a key concern when you're looking at a market and seeing whether or not you want to get in to it. And that's certainly true for taxis, and we think it's mostly true for gardeners as well.
Aaron: So when you guys were getting started, you mentioned that this wasn't one of those romantic stories where you had this inspiring passion for the field, but you found a problem that people were willing to pay for, and you build a service around it. I'm curious. How did you first learn this stuff about your potential customers? You just go door to door?
Jeremy: I wouldn't say that we don't have passion for the field, we do. But the passion doesn't lie in, "Oh my God, I really, really love grass." It is more, "I really love making things that dramatically improve people's lives." So in this case, let me just paint you a short picture of what it's like if you had a lawn. I'm assuming you don't live in the suburbs.
Aaron: I live in San Francisco, so no.
Jeremy: Yeah. So uniquely, not my type of audience. But if you did have a lawn and you were going to outsource the care thereof, your process would probably be going to Yelp or Craigslist and being presented with a series of service providers. You'd give them a call and you'd probably have to wait a number of hours because the phones were in the tracks and they are actually out mowing lawns. When you do hear back, they'd set up an appointment days or weeks in the future to come out and physically set foot on your property just to give you a quote.
It didn't take us long to build this. We built some site-navigating tools that allow us to measure up the script on that regard. We can give you a quote in under two minutes, and it's 99% of the time, it's totally accurate. That is already a dramatic improvement on just getting an estimate for your yard. And then the improvements continue from there. There's ongoing optimizations around quality control. One of the characters of this space is that quality tends to degrade overtime. So we have just basic ratings, reviews, performance optimization stuff to your base computation to allows us to ensure the quality remains high over time and/or continues to improve.
Aaron: It's so interesting to hear. It sounds like something so simple on the top level, right? "Hey, we got a gardener show up." But when you start digging in to what makes the business work and what makes it work better than other people who have tried some other things, it's these details that you're getting at.
If you're just joining us, I'm Aaron Harris, and you're listening to Startup School Radio. My guests are Lawn Love's Jeremy Yamaguchi and Pete Koomen of Optimizely. So I want to turn back to both of you and start with Jeremy and talk about something that I think people might find surprising, which is you had your web dev consultancy, then you had Golden Shine which you sold. You knew what you were doing. You knew how to build a business, and you still came and did Y Combinator. You went through the program. Why?
Jeremy: I like the use of the word 'why' in this setting. But I think it's clear. YC is the leader in the space by a magnitude, many orders or magnitude.
Aaron: But what do that actually...? Because we do different things for different people. And I'm curious about your experience, what you thought you were going to get going in.
Jeremy: So what I thought I would get is largely what I did get. And that is, number one, advice from people who have deep startup experience. And YC is uniquely well-positioned here in that there is a lot of people who have started companies and have anecdotal evidence of maybe one or two iterations to the corporate building process. Y Combinator partners have seen literally hundreds of iterations of the same process. They've seen companies do things that typically succeed or universally fail, and so their pattern matching skills are very, very high. Are your pattern matching skills high, Aaron?
And by virtue of that, you guys are very well equipped to provide good startup advice. So I expected to get good advice, I did get good advice. That was the expectation number one. And then also, probably, it's a by-product of the good advice and the strong alumni networking on the advantages that that represents. The fund raising advantage is pretty significant too. And I did expect to get that and did get that as well.
Aaron: What was the thing you found most surprising in terms of a thing that you thought like, "This is how I should run this business because it's how I've done in the past." And I don't know, a partner said, "No, no. That's no. Try this way." Or, "Try iterating in this direction," and it led you to a better place.
Jeremy: So I think one of the single biggest values I got was this uniform and unilateral and just obsession with growth. And whenever you start a company, I think most people don't, their ambition is not to not grow. They very much expect and hope that they will grow. But YC has this obsession around it that helps you cut through a lot of other noise that's inherent to starting a startup. Startups, I think it's a PG who said that they don't generally die; they drown. So you have way more things to focus on than you have to time and available energies to focus on. And this obsession around growth enables you to cut through a lot of the noise and focus on the things that really matter which turns out to be growing your company.
Aaron: Yeah. It's funny. When you say growth, it means a lot of things. I think it means different things to different people. And I've heard a criticism which is like, "Oh you focus too obsessively on growth. You should focus on this piece of the business or on product/market fit before growth." And what I always say to them is growth isn't one dimensional. Growth is the by-product of getting the other things right. It's the top line metric that only works if you have a product that people want, if you have a way to communicate that product to them, if you consistently make your customers happy so that they don't go around trashing you to other potential customers.
Jeremy: Yeah. Product/market fit is a necessary condition for real sustainable growth anyway, right? So you're going to get that if you're focused on the outcome.
Aaron: Yeah, as long as you're iterating underneath. That's the willingness to change and challenge basic assumptions. I think it's one of those things that a lot of people don't do. Either, I think, probably because it's uncomfortable. I think part of it might be a creativity thing like, "Oh, I can't figure out what the next thing to do is." But I don't think that's actually the main issue. I think that if you're running a business, you actually understand it pretty damn well and you have really great ideas about it. But I think that challenging your own assumptions is hard. It's hard, A, because it's uncomfortable. But maybe more importantly, it's really hard to step out of your day-to-day and decide to do something. Today, I'm going to do something different and see what happens.
Jeremy: Sure.
Pete: I think being constantly self critical is one of the most difficult parts of being a founder. And to your early question, so joining YC, we actually, when we left our respective positions, both Dan and I came from Google. And we really thought we knew what we were doing when we started the whole startup thing. And we thought about making it a successful startup in the world. Great, how could we not know what we were doing?
Aaron: How many people worked at Google while you were working there?
Pete: Thousands, thousands. In any case, we started and we decided we weren't even going to try to get in to Y Combinator because we didn't need that. Because it was, at that time, I think we thought of it as a thing for new college grads who really didn't know what they were doing, which was not us, because we did know what we were doing.
Aaron: Because you guys were so smart, you had all these successful startups in your past.
Jeremy: So smart.
Pete: And so it took two startup failures through that year to reduce us to this level of desperation where we really truly understood how little we understood, I think, at that point. And as a founder, I think constantly being reminded of that and being willing and open to admitting at all times is difficult. It's hard.
Jeremy: It's what drives the iterative self-improvement process. Is that right?
Pete: Yeah.
Jeremy: Like this sense of, "I'm not where I need to be," and this, I guess, ambition for future you that is dramatically bigger than the current you. It is a painful part of being a founder but it also core to what makes you improve and get better over time.
Aaron: Yeah, absolutely. You can't grow a business without growing yourself a lot.
Jeremy: Yup.
Aaron: We're getting pretty profound here. It's weird to think about the fact that when you found a company, it is more than just having a job. And I think people who haven't done it think of it as like, "Oh, you just have a job. It's just you have a little more freedom or whatever." You're operating without a net, right? That's one big thing. So every time you make a mistake, it really hurts. Especially once you have employees. You make a mistake and you think your mistake led to revenue dropping, and then you start thinking about the fact that you might have to lay people "Oh my God."
And we were joking in the office. I think Jessica Livingston, one of my partners, was saying that it just gets harder as you grow, and your journey, you are in your earlier part of your growth. And Pete, how many employees do you guys have now?
Pete: So we are about 350 now.
Aaron: How many offices?
Pete: We have three offices. One in New York. One here in San Francisco; it's our HQ. And one in Amsterdam.
Aaron: That's wild.
Pete: It is wild.
Aaron: Has it gotten easier or harder?
Pete: It's only gotten harder. I remember sitting in Y Combinator and watching these amazing speakers come through, and assuming that once you really got to that point, there was more coasting and you had other people who do all the hard work. But it's been anything but that, a joke that we now have 350 bosses in our offices and then thousands of bosses outside the company who use or products. And yeah, it's tough.
Aaron: Do you hire differently at persons 300 plus than you hired at the beginning?
Pete: Yeah, I think so. The profile that you're looking for is probably different, but also by virtue of the fact that when you're two or three people, you really can only hire people that you know or who know people you know. You're just basically looking at anybody else in your immediate network and going after that. And we now have resources that help us actually look much further in the field and you can spend a lot more time defining large executive roles and things like that.
But when you're starting, you really want folks who are... Like for our first sales hire, we really needed somebody who is able to write the playbook. They were able to come in, there was nothing that, and write it. And later on as you build the sales team, you want people who are able to just do amazing jobs following the playbook, crushing.
Aaron: Yeah. So just for my final question, I'm just going to turn it back over to you, Jeremy. You're next biggest challenge, what happens next for you?
Jeremy: So I would say supply side scaling is hard especially while maintaining the quality that we are looking to hit. So it's pretty easy to grow quickly, but it's much less easy to go quickly while doing great work and making sure that you're only working with the best service provider in your market. That's a challenge, and that's something that I think we are up to.
Aaron: Having seen where you started almost a year ago, I think when we first met, where you've gotten now, I have no doubt that you're going to get there.
Thank you so much to both of my guests, Pete Koomen and Jeremy Yamaguchi. It was awesome having you both. For more about Pete and Optimizely, you can check out Pete's book, A/B Testing: The Most Powerful Way to Turn Clicks into Customers. And visit them online in optimizely.com. You can find more about Lawn Love at their website lawnlove.com. You should definitely get them to take care of your lawn.
This is has been Startup School Radio. We air live every Wednesday at 1:00 p.m. Eastern, at 10:00 a.m. pacific right here on Business Radio channel 111. If you have questions about something you heard on today's show, email us at Business Radio at Sirius XM. You can also follow me at @harris on Twitter. Thank you for joining us today. And special thank you to my senior producer, Lisa Mantineo, and associate producer and engineer, Dion Simpkins. This has been Startup School Radio. Be sure to tune in next week for Startup School Radio. Thank you.
