Rapidly Scaling Twilio Into a Worldwide Juggernaut
Jeff Lawson, Twilio, CEO and Co-Founder
Jeff Lawson, CEO and co-founder of Twilio, was a great deal of fun when he dropped by ZURB HQ to get on his soapbox. More than that, he shared the challenges Twilio faced into breaking down the telecom monoliths that have existed for decades.
Jeff wasn't just insightful. He was funny, telling us a story about Dave McClure and his mysterious Jedi PR tricks. He also told us why one of the first things he built with Twilio was an application that would Rickroll unsuspecting victims on their phones.
One of the things, however, that still resonates with us was how culture plays an important part in scaling a company.
Feel free to listen to the podcast as you read through the summary of the event below.
Breaking Down the Telecom Monoliths
When it comes to the telecom industry, it’s nothing new, said Jeff. It's over a 100 years old and hasn't changed much over the years.
Network of copper and fiber under our streets. Then you could plug a phone into it. And that was what people considered telecommunications.
Telecommunications, traditionally, weren't very user friendly for software developers, said Jeff. It was big software and big hardware that you couldn’t modify to suit your needs.
These monoliths aren’t built for the software people.
Jeff saw an opportunity in this old, stodgy way of thinking and doing business. In his other entrepreneurial ventures, Jeff and his cohorts wondered what if there were applications that could reach out and touch someone. Apps that could talk to customers, vendors and business associates in new and interesting ways. After all, as Jeff put it:
Communication is really fundamental to how we interact, how we do business.
Jeff said Twilio basically breaks down these monoliths into smaller building blocks, putting the power of telecommunications into the hands of developers. It allows them to take those building blocks and arrange them into different applications to solve business problems for their companies.
You've Just Been Rickrolled
In 2007, no one was safe from being Rickrolled, thanks to Jeff who built an application that would play the prank on unsuspecting victims through their phones. He built it as a way to demonstrate what he was building to his friends at the bar. For it to work, you had to email an address which would then trigger a call. It was just supposed to be a harmless prank to explain Twilio. As Jeff joked:
We were never ever supposed to be talking about that ever in the life of the company, because … not at all the use case that we actually want people to associate with Twilio.
After an evening of drinking and sushi, however, the Rickroll gained a life of its own. Just two days before Twilio's official launch, Jeff was out with investor Dave McClure, whom he had previously shown the Rickroll app. Feeling a little buzzed, Dave asked Jeff, "What's the email address?"
Dave then pulled out his phone and activated a Rickroll on someone. Afterwards, he said to Jeff, "So I just Rickrolled Mike Arrington."
That is to say, Micheal Arrington, founder of Techcrunch.
When you're launching a company, you have this very orchestrated notion of what your launch is going to look like. You have this idea in your head and it's all embargoed. And there's a shroud of mystery around what you do. You’re coming out and suddenly you have a big splash in the media. You have this notion that this is what a launch looks like — and then he Rickrolls Mike Arrington.
At the time, Jeff thought that Mike wouldn't know where it came from. But Dave does something else, he posts on Mike's Facebook wall that the Rickroll came from him via Twilio.
But Jeff still felt like it was going to be alright. After all, he still had his official meeting with Techcrunch to brief them on the launch.
10 minutes later, a post goes up on Techcrunch: "I just go Rickrolled by Dave McClure using something called Twilio. God, I hope its more useful than this."
But the stunt didn't backfire on Jeff as he feared. The placeholder site for Twilio blew up, getting huge amounts of traffic. When they saw the amount of traffic, they scrambled to get a sign-up form on the site.
And the next day, Jeff had his meeting with Techcrunch, who ended up writing about their actual launch.
Dave works in mysterious ways. And that little PR stunt was a cool way to kind of set the stage for the Twilio launch a couple of days later.
Scaling the Cloud
Having previously worked at Amazon, Jeff learned an important lesson from the company's S3 web service, which at the time was designed to scale at a set number. But the design team got sent back to the drawing board after a meeting with Jeff Bezos. That's because they assumed they'd just re-architect the service to scale linearly if it gained traction, said Jeff. The lesson he learned that was if a cloud infrastructure takes off, you won’t have time to re-architect it, he added.
If you fall over because of scale, you're going to put a nail in the coffin in the cloud infrastructure. The whole promise of the cloud is scale. If you can't live up to the challenge, then there’s no point in the product in a lot of ways.
It took another two years before S3 was ready. And that lesson stuck with Jeff, which is why Twilio invested in scale and capabilities very early on.
Twilio also helps developers not worry about scale by having them pay for what they need, not an entire circuit or channel like in the old school telecom model.
Another way that Twilio has scaled is by forming partnerships, with other services such as Parse and KDDI in Japan. Which has helped Twilio break out of the United States. It’s now in 40 countries worldwide.
Just last year alone, its customer base skyrocketed by 400%! Some 150,000 developers use Twilio, which processes an average of 1.5 million calls a day.
Twilio, through its partnerships, is certainly creating an ecosystem that is the best of breeds rather than the old juggernaut all-in-one services, such as Microsoft.
But scaling isn't just about users or how many countries you are in. There's much more than that. There's the company itself and the people who work for it.
At a certain scale, you have to think about the company as a product, said Jeff.
Give a lot of my thought to the company as a product.
Jeff put it this way: you have a certain experience with a product. Working at a company is the same way — it's a product, you include things as well as exclude them.
A company should be easy to use like a product is.
And that comes down to culture. That's why Twilio has made it a point to write down its core beliefs, which have evolved over time with the company, which they call their "Nine Things." It’s called that because when you think of core values, people want to barf, said Jeff.
Think core values as a thing on the wall in a nice frame with the words [like] "integrity."
The tenants had to not just be words on the wall. Because if they are you just go about your day and do what you want to do anyway, said Jeff. "That's not culture," he added.
You don't create them, you articulate them because they have to be something that's already there. If you create them, then they're just nonsense on the wall. If you articulate them, they're real and all you're doing is stating what’s there. Because if you don't put a spotlight on them, you're at risk at losing them.
Our conversation continued with Jeff as he took questions from the audience. We like to thank Jeff and all those who attended.
Ryan: How many of you remember this tune? All right. Well I just want to play a little joke. Sorry about that. My sense of humor. The year was 2007, Jeff Lawson built in an application that would basically rickroll [sounds like] people through their phones. It wasn't a prank.
It was Jeff actually demonstrating his then new cloud based telecommunications service Twilio. Five years later, Twilio is in some 40 countries. Its customer base has basically exploded by 400 percent in the last year. 150,000 developers are using the service which processes 1.5 million phone calls a day. Certainly Twilio is growing at an amazing rate.
I want to get into all of that, but first, please give a warm welcome to Jeff Lawson CEO and co-founder of Twilio.
Jeff Lawson: Thank you.
Ryan: Thank you very much Jeff for being here. I'm really excited to have you. And I just wanted to start off by having you give us kind of like, almost the elevator pitch. What really separates Twilio from other technologies that are in this space. What makes Twilio different from stuff like Skype, Google Voice and all these other applications that sort of came before.
Jeff Lawson: Yes, obviously I mean telecom is not new. Telecom is over 100 years old and has been around for a very long time, except that for a long time, communications were sort of like this notion of well, we have this network out there of copper and fibre under our streets and then you could plug a phone in to that and that was what people considered communications.
What Twilio is doing is building this set of building blocks that put the power of communications into the hands of developers. To combine those building blocks into unique and interesting applications that solve business problems for their clients.
It came from an observation that we had which is, I'm a serial entrepreneur and this is the fourth company that I've started. Actually, I dropped out of college and started three companies in a row and then went into Product [sounds like] at Amazon web services.
When I left Amazon I did so because I wanted to start my next company again and was looking at a bunch of ideas. One of the things I realized was each of my three previous companies at various points in time, we said, "Wouldn't it be great if our applications were able to reach out and talk to our users, our customers, our employees, our vendors, in some interesting way that would create the customer experience that we were trying to create?"
Communication is really fundamental to how we interact and how we do business both as consumers with our business or B2B and all sorts of things like that. Every time we looked at building those things we wanted to build to create the communications experience we wanted, we found that the world of communications was not designed for software people.
It was essentially... First of all there's is very esoteric carrier stuff that's out there that isn't fundamentally software. Then there were a lot of companies selling software that did various forms of communication stuff like PBXs or call centers and all this kind of stuff.
Yet there were these big monolithic applications so the big hardware, big software, very expensive. Very esoteric and if you wanted to configure them, to mold them, to make them into what you wanted that really wasn't what they were designed to do.
They were designed to do the thing that they built them to do. As a developer, you say, "What I want are modular building blocks that can arrange to build the experience I'm looking to create" and nothing like that existed.
That's why we started Twilio. Is to provide that level of flexibility of those fundamental building blocks that brings communications into the realm of software people.
Software people need building blocks in order to construct the experiences they want and that's why we built Twilio the way we did. It's a set of eight guys. A plurality of each simple building block that then you can put together in arbitrarily interesting and complex ways to solve business problems.
Ryan: It's instead of just having here's this huge big box of stuff you can't change. Let's give you all these little boxes and build what you need out of it.
Jeff Lawson: Yes and it's like, you know, think about when you call your bank's toll free number, right? The machine answers and it says, "What's your account number?" and you enter that. It says, "What's your mom's dog's maiden name?" and you like answer all these questions and then they finally get you to a human being.
The first thing that person asks you is "What's your account number?" Right? We all have had that experience. And it's because these systems that companies use fundamentally aren't for software people.
That is ridiculous to our software people. What they have, what companies have been deploying [sounds like] are these monoliths. Right? Big hardware, big software monoliths that don't talk to each other. That don't talk to other systems very well.
Or, spending the money to do that is so cost prohibitive that most companies don't do it or they try to do it but really the mediocre and integration points that these monoliths have just don't accomplish the task. That's how you end up with these horrible customer experiences. Based on monolithic software.
One of our fundamental beliefs is that we are in the age of software people. Capital "S", capital "P". Software People. That's not just people who write software like, writing code. It's more important than that. It's actually people who have a fundamental belief in the power of software to tackle the world's problems.
The way you do that is by having these agile, nimble pieces that you can composite together into solving problems and so that's why Twilio is built the way it is. It's because these monoliths are not built by or for software people.
Ryan: They're just built by the monoliths for the monoliths basically. In that way. Just what they wanted to do and this is what you get for the price.
Jeff Lawson: Yes, exactly. You see, your iPhone. You can't open up your iPhone. I mean, literally. You can no longer use a service [to] stuff. That's sort of how these big monolithic software and hardware packages are, that traditionally companies use for a lot of different things, right?
That's why people, you know, they spend tons and tons of money on Oracle, right? Because you have to pay Oracle every time you want to change something, you keep shelling out.
That's sort of the experience of big hardware and big software is that notion of going to solve a thing. It does what it does. Yes, there are a few integration points or a few levers you can pull, but for the most part it just does what it does and the end result for consumers is horrible experiences.
Ryan: I kind of want to, you touched a little bit about the origin. I kind of what to go a little bit back as well and I kind of want to first talk about the joke, the prank and why you did that and even Dave McClure [sounds like] used it..
Jeff Lawson: We were never, ever, ever supposed to be talking about that, ever, in the life of the company because it was just a, not at all the use case that we actually want people to associate with Twilio because Twilio is about building business applications.
Ryan: Right. Absolutely.
Jeff Lawson: It was one of those things where we were building Twilio. We actually had like the Alpha version of Twilio built and we handed it, we gave it to a small handful of our alpha customers like, you know, way, way, way in the early days to get customer validation of what we were building. One of the things that I had built was the rickroll app which was you - I'm trying to remember.
It was actually before we had SMS. You would email an address and emailing it was like email@example.com I think it was. It would trigger a call. You would just email a phone number and then it would trigger a call. It was one of the ways I described to my friends, who were like, less technical, about what I was building with Twilio, right?
We'd be at a bar and it'd be like "Here I'll show you." Their phone would ring and they'd get rickrolled. They'd say, "What the hell was that?" and I'd say, "Well I guess that's Twilio, I don't know." It just seemed sort of funny and rickrolling was a thing in 2007.
Again, we never really intended for anyone other than my drunk friends in bars to have that. I showed it to Dave McClure one day and he had been an advisor to the company and then an investor. I showed it to him one day and he said, "Oh that's funny...." and then two days before we launched Twilio, it was November actually, probably 18th. What's today?
Jeff. Sixteenth, okay. Four years ago Sunday, we were actually having sushi and we were drinking and we walked out of the joint and Dave was a little inebriated. That's uncommon so it's like, it's okay. He's fine with that.
He said, "Hey, what's that email address for the rickroll thing?" I was like, "Oh it's ric@twilio" and he says, "Oh, okay." and he pulls out his phone and he's like, "blah, blah, blah done." He's like "Yes, so I just rickrolled Mike Arrington ."
I felt like, when you are launching a company like you've this very orchestrated notion of like what you're launch is going to look like. Like, you have this idea in your head, it's like all embargoed and nobody knows and there's a shroud of mystery around what you do.
Then you're coming out and now suddenly you have a big splash in the media right, you have this notion of this is what a launch looks like and then he rickrolls Mike Arrington and I was like, "Oh sh..."
You know, Mike's a little, he's got his ways about him. I said, "Well, at least he won't know where that came from, right?" Dave says, "I just posted to his Facebook wall that that was from Dave McClure at Twilio". I was like "Ohh."
"I'm like, "Okay, well it's fine, I've got my meeting with TechCrunch tomorrow to brief them for our launch the next day. Yes. No harm done."
Ten minutes later, a post goes up on TechCrunch. "I just got rickrolled by Dave McClure using something called Twilio. God I hope it's more useful than this."
Meanwhile, we've got a placeholder site that all it had was our logo, hosted on like Slicehost [sounds like] you know, the tiniest like four megabytes of RAM like, whatever, like $0.12 a month kind of hosting and it was like huge amounts of traffic.
We were like "Oh shit" and we like scrambled to put a form up there to capture email addresses from all of these like thousands and thousands of people visiting this like placeholder site. And so that was it.
The next day we went and we briefed TechCrunch and then they wrote about our actual launch the next day. No harm done and it turned out that it was actually great, you know. Dave works in mysterious ways and it turns out that that little PR stunt was a really cool way to kind of like, set the stage for the Twilio launch a couple days later.
Ryan: He was pulling his Jedi mind freak there.
Jeff Lawson: It was. It was a total Jedi PR trick. Yes, but we were kind of freaking out. We were like "Oh my God." We so carefully thought about our messaging and our like launch and how it's going to be and now we're like the "rickroll company".
Ryan: I actually do want to move away from that get into kind of more of the early days, after that launch. You eventually kind of raised like $3.7 million in your first round. You've got Dave McClure advising, investing.
When was it that you guys actually first started looking at scaling in and what were some of those challenges that went into kind of, building this kind of, almost new kind of, way of telecom infrastructure that really, sort of, didn't really exist as much back then.
Jeff Lawson: I mean, scaling the company or scaling the product or....?
Ryan: Scaling the product and then scaling the company after that.
Jeff Lawson: Gotcha. I mean we built for scale from day one. Everything that we did when we built Twilio was designed to scale horizontally. We actually built, we were built on top of Amazon on top of EC2. We built our own cluster orchestration system we call box config and that was used to be able to scale out our systems.
One of the reasons we did this is you know, I learned a really interesting lesson in my time at AWS. My very first day at AWS, it was fall of 2004. I joined the company and they wouldn't tell me anything about what AWS was doing before I joined, big secret.
I joined and got wind of all the stuff that they were starting with S3 and EC2 were really the first two ones that were really under development at that point. The morning of my first day, the S3 team had a design review with Jeff Bezos and it was supposed to be the final design review before a launch that week of the service. Like a beta launch.
It was designed to scale, to I don't remember exactly what - some large number of terabytes or petabytes. I don't remember exactly what it was. It was like a beta that they were launching. The team got basically sent back to the drawing board the week of their launch. It was designed to scale and they said well, "If this thing gets trashed and then we'll re-architect it in a way that really scales linearly with the traffic."
The lesson learned was if a cloud infrastructure service takes off, you will not have time to re-architect it. If you fall over because of scale, you're going to put a nail in the coffin of cloud infrastructure.
Right? The whole premise of the cloud is scaling. If you can't live up to that challenge then there's no point in the product in a lot of ways right? The lesson learned was, you launch with an architecture that will scale horizontally as your usage grows. You just cannot drop the ball when you are an infrastructural provider.
That lesson always stuck with me. I don't know if you guys remember, S3 I believe, launched in 2007. Right? They went back to the drawing board and rebuilt it from scratch and it took over two years. Interesting lesson, right? I think Bezos is a really smart guy and I think we could probably all look back at the graphs of S3 usage and say he was probably pretty right about that assertion, right?
We invested in scale, or at least the capability to scale very early on. That's one of the reasons why we built on AWS.
Ryan: Gotcha, gotcha. In terms of that infrastructure, and I've heard you say before that you were kind of moving away from these on premises, these monolithic models of telecommunication and moving into this more cloud based space. Having this infrastructure that moves horizontally.
How do you leverage that infrastructure to help with the scale and how do developers that work with your API do it as well?
Jeff Lawson: Our goal is that a developer working with our API doesn't have to think about scale at all. The traditional model in telecom is this. They call it 'per port model', you buy software, you install it on a server, you buy hardware and you buy "X" number of ports which mean simultaneous phone calls. You buy 100 ports. That means that thing can do 100 simultaneous calls. Your 101st person gets a busy signal.
That's how the whole world of telecom has been architected for 100 years, because it goes back to how many like pieces of copper in a circuit switched network. How many pieces of copper do you have to carry a phone call that then go into a piece of equipment that then .. so the whole system is designed around this notion of you're buying circuits.
The new model that we envisioned that you would, that what it meant to bring this into cloud, was that you would never pay for a circuit or a channel or a port. You would just pay for what you used and we would scale as wildly as you need us to.
It does present some challenges because the telephone system that we to our customers that does not present that limitation. A lot of the things that we do engineering and architecturally is we're bridging the gap between what we think makes sense to software people.
Which is, I make a request to API to say make a phone call. and I don't care if it's the first phone call or the billionth phone call, it should work the exact same way as every other request I've ever made to Twilio.
Behind the scenes, we're working with carriers who have this channel model and all this kind of stuff, because they've invested in, they've bought gear that's in their racks that they bought on this channel model or they laid copper that is the whole underlying legacy is in this model. We're abstracting and so we do a lot of things, for example, like we have call queues, right?
If we have you know, 100,000 call requests and we have 75,000 call channels available via carriers, then what we can do is multiplex that and ensure multi-tenancy fairness to all of our customers, based on that.
Building multi-tenancy into a telephone network really hadn't been done before, because you didn't need it when you sold people channels. You said, "I'm going to sell you a channel. Now that's yours. And I'm going to sell another guy some other channel." You have a tremendous amount of waste in that notion because most of those are going unused.
Yes, the guy who actually needs it can't use that capacity that's available because you didn't sell it to him. Our notion was pretty different, about how to build a scalable system so that software people who use Twilio, never have to try to wrap their heads around this esoteric historical implications of how telco was built 100 years ago.
Ryan: Gotcha. It's almost as if you're kind of being the bridge between that for these developers so they don't have to worry about stuff like that. It just kind of happens for them behind the scenes. You guys have this huge engineering engine that you're like really plugging into it.
Jeff Lawson: Yes, exactly. Then figuring out the challenges of, how do you bridge these two basically incompatible models of how you view the world? A lot of our engineering goes into how do you do that. How do you do it in a multi-tenant way so that no customer affects any other customer? You think about that.
Telco, consumer space is not multi-tenant. Your cellphone goes "I have no signal" it's because it filled up. The closest base station to you doesn't have any capacity available and so there's no such thing as that.
Our goal is to say, "We're going to create that multi- tenant capacity on top of a system that is fundamentally not multi-tenant."
Ryan: Are we going to eventually see the change in that system. Are we going to see that old school kind of way go away at any point?
Jeff Lawson: Yes, possibly. I mean, obviously, there's a lot of investments made in the existing infrastructure. Powers it. That infrastructure is more and more going to the world of voice over IP, fiber optics and things like that that aren't essentially circuit switched networks.
What's interesting is, I think that as the industry changed from old model to new from a technology perspective, a lot of people actually kept the economics. The business model's the old one which don't make any sense to me.
Ryan: Speaking of that, I had another question before this but since we're talking about the business models and such, I understand that Twilio has like a cost based model approach. I was kind of wondering how that works for Twilio and how does that work in differential to some of these older business models when it comes to like enterprise product and telecommunications?
Jeff Lawson: It's like a 'pay as you go' model that we have, which is essentially you pay per minute for a phone call, you pay per phone number, per month for phone u pay for message SMS and you know, we did this because it aligns how much you pay. I mean you build a Twilio very closely with your usage of it and it's a few reasons.
Number one, this is, I believe, the prevailing model of the client [sounds like]. This is what Amazon has. Crash 3 for EC2 [sounds like] and this is the new model because it very closely aligns usage patterns with what you should be paying.
It also means that as, say a developer building a product you can experiment very inexpensively. You can build very inexpensively but then when your thing scales, you're build just scales with it, right? It's perfectly elastic along with what you're consuming. That's ultimately the end game of all this stuff and any other business models in the middle are essentially, ultimately, going to go to this. We went straight to this.
Ryan: Just cut out the middle man completely and go right at it.
Also, another interesting thing about Twilio that you know I've read about and so it's almost as if through the cloud you guys are creating in an eco system and you guys to break into international markets you're making partnerships with like AT&T, KDDI [sounds like] in Japan and recently Parse, [sounds like] I believe. I was wondering how these eco systems not only help Twilio provide its service but also continue its growth and how this also leveraged for the product for the developers.
Or how the developers can leverage this with your product.
Jeff Lawson: Yes, it's different cases and various things. I mean the way we look at partnerships round distribution is just, you know, how do we get... We've 150,000 developers and companies using Twilio today. How do we continue opening up the market making it so more and more and more developers and companies can be customers of Twilio or reap the benefits of what Twilio can offer?
Part of that is by working with other developer focused companies, you know, like Parse or Microsoft Azure we've worked with. Partner on their new azure cloud offering. Because you know, think about you know, the Microsoft, you know, community and historically it's been sort of this, you know, over here's the Microsoft community and over here's the non-Microsoft community and then the arrows sling between the two right?
That's the history of that and actually I think Microsoft is doing a lot to break-out of that mode of thinking. They are, right? They now support all sorts of open source software. they support PHP movie [sounds like], node [sounds like] and everything else like, and so they're trying to break down those walls and as are we, right?
We are trying to reach into that Microsoft community which is historically a little more say corporate than some of the people who say, go to hack-a-thons and things like that. It's just a new audience for us and a new audience for them and so it's a great way to cross pollinate our developer communities.
Then you look at internationally, and KDDI is a great example of, there's a huge demand in Japan for what we do. The Japanese market is a pretty different one than the United States and you really have to have a lot of expertise in order to succeed there.
A very common way to do that is with a partner in Japan. That's what we've done in Japan with KDDI, their web technologies group.
Ryan: Do you think that these kinds of eco systems are kind of the new business model that will be like niche services that this service provides this but we need this other service so that we can continue carrying our service rather than the monolithic kind of Microsoft, kind of corporate model?
Jeff Lawson: Like over here, you have Microsoft and Microsoft is designed to solve every problem you have, versus over here, you're going to have best of breeds of many things. Yes, I think the future is in best of breeds personally. I think, look, do you see Microsoft doing that already.
They are now saying "Look it's not, we're not insular Microsoft world any more. We actually believe that all these other technologies play apart. We're going to solve certain problems for you but you are going to want to bring other solutions to the table." This is a very heterogenous world of technology. I think that is the future.
Ryan: In terms of this, now that you're in 40 countries, 6 continents, how do you guys deal with kind of just the rapid growth of the company on the back end of things? The engineering, the internal stuff? Where do you kind of make the decisions like, okay, you know there was a [Venture "B"] article that said you guys hire like every quarter you know, at least 20 engineers or something of that number. How do you guys deal with that kind of growth?
Jeff Lawson: I think it was we have 20 engineering open wrecks and they extrapolated that to every quarter you hire 20 engineers which would be great. Hard to do that.
Ryan: I think there's some in here, right?
Jeff Lawson: Yes, like you're all hired, now next month.... That's a bit tough, but yes. I think one of the things that happens as a founder and particularly as a founder and CEO is that in the very early days you're thinking a lot about the product. As you get to a certain level and scale you start to give more attention to the company is a product too.
I think Steve Jobs stated somewhat famously, that the thing he was most proud of wasn't any particular product but the company he built that could sustain and build those products. I give a lot of my thought to the company as a product. The company as a product, meaning, when you use a product you have a certain experience associated with it.
You know, when you unwrap your new iPhone there's an experience around that. It's got certain benefits. It's got certain trade-offs. All sorts of things like that and working at a company is the same way, right? It's a product and you include things, you exclude things, you want people to feel a certain way when they interact.
A company should be easy to use, like a product is. I remember you know, one of my first days at Amazon I had to go through like reams and reams of paperwork to figure out the whole like healthcare situation, right? One of the things I asked my team was you know, when we rolled out our health insurance like a couple years ago, I said, "How can we make this simple? Like, how can we make it really easy for you when you come in to understand your choices and make a decision?"
As an example of what it means to create an easy to use company, but the other thing is culture and culture is something people talk about a lot as it relates to company building but the way I've looked at culture and our team has, is the team is, you know, we've got 130 people now.
If you believe Venture, it'll be at 150 tomorrow, 170 on Thursday.
Ryan: It'll be like 240 in any time soon, right?
Jeff Lawson: Yes. December 31st is 1,000. Culture is the collection, you've 130 people, out there making decisions every day. Thousands of decisions are made every day. Culture is how you as, say a leader of the company are confident that every one of those decisions is the right one. In an environment where you say "You know, people aren't allowed to make decisions" Well, that obviously doesn't work. You don't great people and then say 'You can't make decisions.'
In a lot of ways it says, "What is it that all those decisions should add up to being you know, the Twilio way." Here's how Twilio would do these things. That's what culture essentially is. For us, we rolled out earlier this year what we call our nine things, which are sort of like core values and we designed them to, well, a few interesting things about that.
The thing about core values most people think of core values as a thing on the wall and a nice frame with words like, 'integrity' that nobody pays attention to.
Ryan: Old school motivational posters and all that stuff.
Jeff Lawson: Yes, you know it's telling the core values of Microsoft, were identical to the core values of Enron.
Ryan: We all know how that turned out, right?
Jeff Lawson: Yes, it was like word for word the same core values. That's not a knock at Microsoft, it's just saying like, if they're just words on the wall, then you go about your day and you do whatever you want anyway. There's nothing special about that. That's not a culture.
What we set out to do was to not just put words on the wall but rather articulate the things that we liked and cherished and felt were fundamental to who we were as Twilio.
At a size, and we started this exercise when we were about 50 some people. To ensure that those things were maintained as we grew the company. And that's what core values meant to us. What is the common thread among us today that we like and cherish and when we are 10,000 people by February what we want to make sure is still here?
And ideally, in 100 years when our grandchildren work at Twilio, you know, they should have the same characteristics, right? When I say we articulated our values, that's exactly what it is. You don't create them, but you articulate them. They have to be something that's already there.
If you create them then they are just nonsense on the wall, but if you articulate them they are real and all you're doing is stating what's there. If you don't put a spot light on it you're at risk of losing it.
We did that and we went through about a six month long process to actually articulate those things, starting with an all hands when we broke up into teams and brain stormed what was our favorite part about Twilio, or what's special, why do we get up in the morning and come to work every day. All that kind of stuff.
Then we curated that list down to about 20. We had a number of focus groups inside the company and really took it very seriously as this exercise of self discovery.
In a lot of ways it's interesting. It's almost like when you're a child, like you just run around and play. That's like the earliest stages of a start up. When you're a teenager, you go through a whole lot of fits and starts of figuring out who you are. That's sort of like the middle stages of scaling, maybe the series "A" state of a company. Where you might experiment.
You might be figuring out like, oh, is this core, is that core. Is this important, is that important. Then you get to a certain stage, past say, your teenage years and you are able to finally articulate yes. You feel comfortable in your own shoes. Here's who I am and that's sort of the exercise we did and we did that as 50 some people is when we started that.
Ryan: It's like, you can almost force the culture in a company before it even exists and then when it starts to exist you force that culture on them. It's something that almost has to be bred with the company, grown with the company. It can't just be, this is it, this is what our culture is going to be. You know, here it is.
Jeff Lawson: It's interesting, I'll tell you a funny story. We actually, very early on when we were like eight people I think, we did articulate what we thought at the time were our core values, and we came up with these five things; continuous improvement, detail oriented, lifelong learning, humble and hungry. Okay, five core values. Which are cool and we hired to those things. We said "These are the things we are looking for in people that we're hiring to join the company."
And then a really interesting thing happened. I was at a Union Square Venture CEO summit, like a little over a year ago. They had like these on conference break-outs and one of the breakouts was talking about company culture and so I chose to go to that one and I was chatting with a number of other CEOs and I was telling them about our five core values.
They said, "That's really interesting, that's very cool, but I have a question. If we were to ask any random employee of yours do you think they would know what those core values are?" I said, "Oh, that's interesting. I don't know the answer to that. Let's find out."
And so I called an outfit that we had written, where actually you dial a phone number and it calls everyone in Twilio and dumps them all into a conference. And so I did that, and it
Ryan: We're going to implement that here at ZURB I think..
Jeff Lawson: You know, and I don't know, maybe half the people answered. I had like 30 or 40 people in this conference room. I said, "Hey, I need a volunteer." and so someone volunteered, and I said, "I'm going to call you back." I called her back and I put her on speakerphone and I said, "You're on speakerphone with about ten Union Square Venture CEOs and the question we have is, can you name our core values?"
She said, "Easy, simple, and she went [inaudible 32:28]" I was like "Oh, that's really interesting, because I was describing our people, right? Humble, hungry, all this kind of stuff. She was describing our product. You wouldn't describe our people as simple and easy.
Ryan: That opens up a can of worms when you do that.
Jeff Lawson: Indeed it does. It makes the holiday party a lot more fun, but...I was like, "that's an interesting disconnect" in that we had articulated these things that we called our values, but they only described people. They didn't describe our product, our customers, what we build, how we build it, how we interact with each other on a daily basis. How we talk to our customers.
All sorts of things that these things that we thought were our core values, they just talked to this small slice of what it means to be a company, not the grander picture.
That's when we set out to articulate our core values that would ideally encompass all the things we do.
Ryan: Very good. Well, those are all my questions and I appreciate you answering them all and I want to throw it out to the audience now, in the last ten minutes here. And answer some of your questions. Who's got a question first?
Way in the back.
Audience: [inaudible 34:01]
Jeff Lawson: The question is, essentially how do we reach out to developers to market Twilio?
Great question. When we started Twilio a lot of investors I talked to said it's impossible to attract developers, they're like this slippery market. No one knows how to reach them. We took it as a personal challenge. Basically we set out with one mission in mind, 'Be everywhere, be awesome'. It's sort of the best way to describe it. I think it's a Jason Calacanis quote.
We just set out to be wherever developers are, we're going to be there. And we're just going to be ourselves, because we are also developers.
We're not going to be there hitting people over the head with marketing. We're just going to be authentic, we're going to be ourselves, we're going to be there to help and that's what it meant to reach developers.
It meant we started going to a lot of meet ups. The hack-a- thon think kind of took off, actually right as we were starting Twilio and then we started investing a lot and actually helping people organize hack-a-thons and sponsoring them.
We started just going to all these events, and again, we have evangelists who are developers. One of the big mistakes is to send you know, marketing people to a developer event and so we built a developer evangelism team that now numbers over ten people and they're developers.
It's just that from a career perspective, a lot of them decided that, you know what? I actually really enjoy interacting with people more than I enjoy spending my whole day in front of an IDE. They write code, they are great developers, but they love the human interaction associated with developer evangelism.
That's what they do. They go to hack-a-thons, they go to all sorts of events and they're just honest and humble and helpful to people and as a part of doing that, they are introducing them to Twilio. That's the best way to do it, which is to start with being developer, be yourself, figure out where you would reach yourself and then go there.
Also that's offline and also online too. A lot of the communities developers participate in online to be involved in as well. Stack overflow, Hacker News, yada yada.
Ryan: Brian [SP] in the back.
Brian: [inaudible 36:11]
Jeff Lawson: Yes, nine things.
Brian: Nine things, okay.
Jeff Lawson: We called them that because like, 'core values' people barf at.
Brian: [inaudible 36:25].
Jeff Lawson: That's a great question. I mean, that's sort of like the ultimate thing. Which is you have to roll them into your day to day life. Or else they are at risk of just being words on the wall.
One of the things we do with our core values is we made them real. So, integrity, people don't sit around saying 'Is this integrity?' you know, "Am I being integrity-us today?" I don't know. It's like these words that just don't fit in to your day to day life as part of it.
Ours are very natural. Again, they bubbled up from who we are. Ours isn't integrity, we say, 'No shenanigans'. You can like, we can imagine what that means. What's great, the perfect implementation of core values is when you hear decisions being made by teams that reference the core values. Right?
You'll hear people talking about "Well, shall we do this or shall we do that?" and someone will say "You know, this, that feels like shenanigans to me." and people don't say, "That doesn't feel like integrity to me". Like that just doesn't happen. But people will say, "That doesn't feel like shenanigans to me because of 'x, y and z'." Then people are able to converse around those topics and make decisions. And that's ultimately the goal.
We rolled them into our hiring process, it's one of the things we're looking at when we interview people, is how do we think they fit into these nine things. It's part of our performance review process. We ask employees, "How do you think you've done at these? Which core values do you think you did a really good job at living this year? And which ones might you think you want to work on next year?"
That's been a part of it. We've avoided, like, we haven't put the posters on the wall, just because that sort of felt like blunt instrument. We haven't done coasters or that kind of stuff.
We have an internal Facebook thing that I wrote in a weekend called 'Bacefook' that's ... it was like, when we got to 75 people, I started to realize that as people were coming into the company, they no longer could meet everybody very easily. It was "Yes you've got to go talk to that person." "Who the hell's that person?"
A weekend project, I learned Python and I learned app engine , was one of the side benefits to me was I built this Bacefook thing. One of the things that I built into it was you could give people badges, to say like, "Hey you, you led to a core value."
"Hey, great job, 'draw the owl'." Is one of our core values. I saw an owl drawn in the bathroom here. Draw the owl. Have you guys ever seen the internet meme, the reddit meme, how to draw an owl. "Step one, draw some circles" and it's got like these sketched out circles. "Step two, draw the rest of the fucking owl."
And it's got a fully drawn beautiful owl.
To us, like, that represents kind of what it means to be a doer or to be at a start off. Like your job every day is to figure out how to do it. There's no instruction book. If there was an instruction book for everything we did, it would be boring. It would already be done.
Like, if you want to work at a company where there's a process and a guide book for everything you do, you go work at Boeing.
If you want to invent the future, where there is no guidebook and you just have to draw the owl, you come work at Twilio. That we sort of adopted that as one of our core values. There's a badge in the Bacefook which is like,'Draw the Owl' badge. It's like, you can say, great job, drawing the owl. You just figured that thing out. You know, you took ambitious information, made sense of it and shipped something. So those are some of the ways in which we do it.
Ryan: Next question. Yes sir.
Audience: [inaudible 40:11]
Jeff Lawson: It's interesting.
Audience: [inaudible 40:25]
Jeff Lawson: Yes, yes. What we do is we work with tier one carriers to terminate and originate calls to and from their phone networks. It's sort of a business development exercise on that side. What you're asking is for carriers to devote some of their switching capacity to us as a customer of theirs. That's a business development to create the relationships with the carriers that you need in order to have capacity in order to satisfy customers' needs.
What we do is, we maintain like an order of magnitude greater capacity with our carriers than what we typically expect we're going to need at any point in time, to be able to handle those bursts, those spikes, all that kind of stuff.
Audience: [inaudible 41:09]
Jeff Lawson: Yes. Again, so we've always... our tendency for us has always been over invest. To make sure that we deliver a stellar customer experience. Just like Amazon did. They going to spend two and a half years engineering this thing. It's actually contrary. I mean, Eric Ries would like puke all over this, right?
It is contrary to the Lean start-up thing. I actually think some of the notions of lean start up are, I mean a lot of it's great and I love it, but a lot of it doesn't apply to the role of infrastructure. You can't, it's not like you put out a feature in an API and then a month later be like, "Ehh, we don't like that feature, we're going to remove it."
It's like once you put something in API like, forever, I mean, deprecating in API is really hard. Right? We still support our 2008 API. In fact until I think, about three months ago, we still supported our alpha version of the API because we had two of our alpha customers who were still using it, right? Actually getting two customers to switch off of it was hard.
Like, we basically had to say, "Guys, seriously" and these were like friends of ours, like the very earliest private alpha customers and so it just shows how hard it is to deprecate things. So, infrastructure is sort of a different level of scaling and fore thought that has to go into what you're building and how you're building it and the level of skill you're building it with.
You can't change it.
Audience: [inaudible 42:31]
Jeff Lawson: Yes, it's one of the trickier aspects of getting Twilio off the ground. One of the things we did always say, which was, "We don't want to be beholden to carriers in order to be successful." and so one of the nice things is that there are laws that say carriers have to sell you capacity. Right? to get started, there are things that occur in the United States that made it so that we did have an equal playing field in some ways. That we could get time on their networks.
Ryan: There is a question over here.
Audience: [inaudible 43:19]
Jeff Lawson: Yes, so here's what we said. It is a living document and it isn't. I'll say that if we've done our job well, my hope is that these core values are roughly the same in 100 years. They don't have to be exactly the same, word for word. But it's not the kind of thing that you change every year. It's not like a strategic decision or a tactical thing. This is the core of the company and very rarely should change.
It should survive employees coming and going, leadership coming and going, various product lines launched, customers, everything. That's the goal. It shouldn't be too living. What we said though, when we finally rolled it out and there was a lot of discussion and I did a lot of focus groups just to hear what people had to say about our core values as we were drafting them.
Sort of, what I said when we launched it, because something like that always incurs people being like "I hate that one. That one's stupid." or "Can't we only have this one?" Like that's the only value that matters, right? There is no shortage of opinions on this stuff.
But what I said was, "These are our values and we're going to live with them for the next year. In one year, we're going to revisit. We're going to see how they fit us for the past year." What I sort of said was "Everyone's got opinions, great. Let's live with these for a year before we start asking if there are any changes." Otherwise, you can just keep trying to tweak them.
Audience: [inaudible 44:55]
Jeff Lawson: An hour, yes, I know.
Ryan: All right. One last question. You actually had your hand up earlier, so....
Audience: Yes, I was just waiting for the right moment. I wanted to ask you, those things you [made] about culture, the nine things. How do you reconcile that with more traditional business concepts like corporate vision statements or corporate mission statements? Like, how do you value those against each other or does one replace theother?
Jeff Lawson: It's a great question. The question is how to balance like core values with like vision statements or mission statements and all that kind of stuff.
You know a lot of these things get a bad reputation as like being just corporate BS. I totally understand why, because again, words on the wall is not living them. Our goal with anything we do is to live these things, not just put words on the wall.
Whenever we do talk about these kinds of things, like a mission statement or anything like that, the goal is to make sure that everybody is able to understand, grab it and live that. Not words on the wall. You don't just send a memo to people. You know, "Here's our mission statement" end communication, right?
That's what a lot of people do. That's what a lot of companies, how they do this stuff. We've been thinking about what is the role of a mission or a vision statement, but it's not communicated necessarily as such.
It's a living thing that we talk about constantly, so we do a Monday morning all hands. Every week. We have for like, four years. Where the whole company comes together Monday morning at 10.30 and it's part of that ongoing conversation that we have every week.
Which is, "Why are we here?" and it's just something that you reinforce by the things that you talk about. By the things you highlight for the company. That's really the vision. It comes down to, what is it? What's the conversation about?
At some point you do want to articulate it. So people can be able to respond or answer that question, but I think it's one of those things that it's okay to encourage for every person to have their own interpretation of what something like the vision is.
As long as it's roughly what the actual vision is. But it's not like "Here's our vision statement and everyone has to memorize it" kind of notion. But, "Here's what we're trying to accomplish, and hey, you probably agree with that because that's why you came to work here and it's okay for you to have, and encouraged, for you to have your interpretation. Your personalization, your internalization of that." That's what means they actually bought into it, if they internalize it and actually have their own version of how that works.
Jeff Lawson: I think that's sort of, I don't know if that answers your question.
Ryan: Very good. Well, once again, thank you very much Jeff. Everyone, warm round of applause.
Jeff Lawson: Thank you.
Ryan: Thanks so much for being here.
Jeff Lawson: Pleasure.
Ryan: Just to let folks know, if you haven't caught our former or past speakers you can go to zurbsoapbox.com and listen to all the previous podcasts, transcripts, all that good stuff.
Thank you once again for coming out. I appreciate it.
[inaudible 48:08] awesome.