Gleb Polyakov, Nylas

All right, here we are, another episode of Breaking Changes. Today I have Gleb Polyakov, the CEO of Nylas, with me. Welcome, Gleb.

Thank you very much. I’m glad to be here.

Well, I’m excited to talk to you today, because Nylas has been coming up on my radar in several different ways lately, most notably the fun recent funding you had, but also just in a COVID world, you all kind of represent, I think, the core resources that folks are needing in their business. So I’m really excited to talk about how you all do what you do, but also how APIs fit into that. So let’s start with the fundamentals for the audience: what is Nylas?

Yeah, absolutely. So at the highest, most abstract level, Nylas is a platform that allows developers to access that super high value, high signal but hard to get to and structured data that exists within business interactions, and then also provides the front-end components that let you develop rapidly against your end goal. So you mentioned some of the acceleration on the digital transformation side. Given the past year and a half, two years, and everything going on with COVID and remote work, folks have used our platform to do things like drive the rollout of over 25 million vaccines across several states, several companies, where we’re powering everything from that front-end user interface across several hundred point of care centers around the country. Different flows, different signup, branding experiences, where our front-end components allow you as a developer to really flexibly decompose, deconstruct and then put into your flow the calendaring and scheduling view, the availability times, the composer for different messages. So folks have used that to schedule vaccine appointments. We make sure we power the email and SMS messages that go back and forth to the folks getting the vaccines, to make sure it’s being de-thawed before they’re on their way, no vaccine is wasted. And then we also do the automation, the logistics, the business logic and routing of putting all that together for those companies that are our customers, that allow them to not only deploy these kinds of use cases quickly but also securely in an ISO, SOC 2, HIPAA certified way without having to worry about the ongoing maintenance or security load or how to extract the necessary business information from the actual API functionality.

And this is all primarily built on the existing services that I’m using to run my business, right?

That’s right.

And so you’re aggregating data and moving data around. Are you all polishing and enhancing that data as part of the process as well?

Yes, we do. So the data types that we give you access to, for example calendars or your contacts folders, the contents of your inbox and the content of whatever messages are going back and forth, the files and attachments and documents that you’re working with in a business context, all of these represent actually a surprisingly unstructured data set. I think that makes sense for a lot of people, like, oh, a PDF is going to be unstructured because it’s a physical document. But then a lot of folks might have the idea, oh, but email’s a solved problem, right? It’s something that’s easy to work with, there’s easy to use APIs out there already. And the reality is that no, there aren’t. The most common comment we get is, Nylas, I thought this is how this type of data was accessed already. And it’s like, yeah, we just have a pretty amazing engineering team led by Christine Spang, my CTO, that over the past seven years has done a great job of making that happen. But we actually launched our service before the Gmail API existed, and the existence of the Gmail API has only made it a tiny bit better to access Gmail data. In some ways the actual contents, the actual things that you’re trying to get out of the messages, plus the use cases you’re trying to drive at the end of the day still have to be very much custom built, and we try to make that easier by abstracting away, by structuring a lot of that unstructured data and providing our developers with a universal API and the components they need to build robust applications.

Yeah, I mean I can imagine what’s in my inbox is pretty insane, and that’s just one inbox. You all have to do it across many different entire teams and all of that, so that’s pretty significant, even with the Gmail API. Well, before we dive into the details of all this, I like to get to know you a little bit better. Why did you do Nylas? What brought you to starting this company?

Yeah, absolutely. I joined the team, as I like to say, the first evil business guy. So I was the first non-developer professional background that came on the team to take the technology that had already been built and some of the community that we had already started working with, and take it to the next level, build a really world-class, broad reach business out of that. So originally I have a physics background and worked in PE for a little bit. I had a startup right before Nylas that was an IoT hardware play in the coffee space, of all things, where we made for 300 bucks an espresso machine for the home that was just as good as a $20,000 coffee shop machine. We did it as open source hardware and software, which, long story short, none of the regulatory industries loved very much. And after three or four tooling cycles that we didn’t expect, we effectively ran out of money and sold off to a larger manufacturer. And when I was looking for my next gig, through a friend at MIT I got connected with Christine Spang. She impressed the hell out of me, not only when it came to her raw intelligence and the ideas she was working on. She’s like a lightsaber, she just gets things done and explains them so clearly. She’s an absolute legend in the back end community. But she’s also just a great human being, a good, empathetic, quality type of person that you want to hang out with. And that’s kind of who you want to start businesses with: people who are super smart, working on huge ideas that can change the world, that are also like humans you want to spend time with and build around as you grow your company.

Yeah, well I have to say, us tech folks need a good evil business person around to move this forward and keep the bills paid and grow and scale things, because as we’re paying attention to the technical details we’re not so good at those other pieces. So I’m thankful for folks like you, and I don’t consider you an evil presence.

I appreciate you.

So let’s dive into the impact of Nylas on the overall economy. In the space we tend to use a phrase called the API economy, which if you talk to the API management providers of the last decade, it’s very much like you build an API and people come and you generate direct revenue from that API, and that’s how you make your money. I see Nylas more as about what my vision of the API economy is, and it’s more about enabling entirely new businesses and new types of APIs and flows and processes. So how do you see the phrase API economy and what Nylas does?

Yeah, absolutely. And I think your way of viewing the world matches pretty well with mine. So some analyst reports you read, some people who aren’t in tech will hear APIs and they think, oh, that’s a way that I can access Salesforce, like APIs are connectors to data. And it’s like, way not really. I mean, yes and no. Or some people who are a little bit more nuanced are like, APIs are a building block that you can OEM into what you’re doing. It’s like, again, yeah, but also no. The analogy I really like to use is, I mean we know what API stands for, but it’s like a wrapper around a more complex functionality that allows you to abstract away the complexity and do something more simple. So for example in a car, the interface to make the car go faster is a gas pedal. You step down on the gas pedal, car goes faster. Under the surface though, what’s happening? The throttle valve is opening, the valve timings are changing, the fuel pump is accelerating its rate, on and on and on, right? Multiple changes happen that make the cylinders fire faster, the exhaust pressures change, so much is going on that you never have to worry about and you shouldn’t have to worry about as a driver, because all you need to think about is, okay, the thing I want to happen is car go faster, pedal go down. And that’s, I think, a slightly better model for an API. And the general trend kind of matches that. If you look at those cars from 1895, 1903, before the Model T, they had a bunch of levers and gears and knobs and dials, because you had to do a lot of that stuff on your own. But as the interfaces got more straightforward, as the level of technical and engineering complexity reached a point where those could be abstracted away, the ability to drive a car became much simpler and the utility of car driving expanded into the trucking and logistics industry, into the private taking a joyride sector, into car races and on and on in terms of use cases. I think a similar thing is happening now with APIs, where, yeah, you could use it to access Salesforce, and I don’t mean that in any disparaging way against Salesforce obviously, but the more interesting thing to me is APIs and the API economy as an extension of that fundamental metaphor of software development overall. Going from literal binary code to higher order languages to, now we have SDKs that let you build whole mobile apps in a day or two. I see that as being the continued trend that, one, is why people view software in the first place. The whole point of it is to provide additional leverage to manual processes that can and should be automated, or that humans aren’t very good at doing in the first place. And it just happens to be that they’re also kind of really cool to work with.

Yeah, I love the automobile industry analogy. I just did an episode with John Musser from Ford about the car as an IoT API device. But I think your analogy as far as the business of it, I mean in cars you have the direct, you have Ford, but then you have the entire supply chain for the Ford ecosystem, you have the aftermarket, you have all these other areas of business in and around and after you buy your car. That, I think, that’s the API economy that I’m looking at and thinking about. And then you mentioned it, the entirely new workflows and processes and being able to be agile, like during COVID. This isn’t a standard business process we need to set up, but we need to react to what’s happening and set up vaccination clinics and do all of that. So it’s having access to the data and those resources, like you said, but it’s really the wider being able to define new workflows, new processes. And no one likes, I mean a few of us API nerds like working directly with APIs, but really we want that hard stuff abstracted away. So that’s what really gets me going. But when you think about APIs and the spread of services that Nylas offers, what are the biggest APIs that come to mind as far as offering the most value to what Nylas does?

Yeah, at the highest level we are here to help you with automating human to human interactions, by which I mean the engagement aspect, the can I have a meaningful conversation with someone quickly through software. And like some of what you’re saying, yeah, some people really love databases, they love them, they love the theory of databases, but most people, I would wager, interact with databases or use them to accomplish a goal as a more instrumental part of their tech stack than a pure intellectual dopamine hit on its own. And so for us, the value we drive is not necessarily, yes we connect and provide automations and logistics for XYZ thing, but at the end of the day we’re making it so that you can talk to your users in the channels that they are using without them having to leave your app, download a new thing, without them having to either reduce your product’s experience or force them to do additional steps on their end. We’re helping you surface the insights and information from the content of that communication in a very high signal and easily usable way, and we’re helping you automate a lot of the process that would have to otherwise go into it or be manual if you didn’t have software built around it. And so that’s the, what was it in 2020, I want to say it was $2.8 trillion that was lost in the US alone on things that can and should be automated. It’s the four and a half million people whose only job right now is data entry. It’s every doctor’s office who’s trying to schedule appointments. It’s every company in any industry in any sector who’s trying to make sure that they’re talking to candidates quickly and giving them a great experience and also managing their own team’s time and scheduling the interview loops. It’s every car dealership that’s sending out invoices or repair stuff to their customers. Those are the end goals of human interaction, and that’s the highest level value that we’re aiming for. Because if you look at the vast majority of software, it all has to build the same three things. You’ve got to have an authentication identity path, so that’s where folks like Auth0 play. You’ve got to have a payments piece, so Stripe, Plaid, those companies play in that way. And you have to be around interactions, you have to have some sort of interaction component in your software to help humans use it. And that’s the real estate that we really want to stake our claim in. Traditionally folks like the Twilios of the world have focused on that really outbound broadcast one-to-many communication on the API side, because it’s a little bit the lowest hanging fruit out there, especially when you can monetize on that super high volume sending. But our belief at Nylas is that the value we want to drive, the real differentiator and help we want to provide our customers, is to build those one-on-one interaction flows at scale and make them super streamlined and easy to work with so that you can honestly not have to think about it, so you can focus on the thing that you’re actually trying to accomplish at the end of the day.

Yeah, I mean you touched on, 70, 80% of my day is emailing, calendaring, setting up, working on making the show happen, as well as the other teams I manage and other people that I manage. But what I do as chief evangelist at Postman is thinking about the important ideas in the space, the important people, brands, and how it all fits together, and I don’t get a lot of time to think about that. And even in the COVID world, I would say it’s gotten worse. I’m doing more calendaring, more scheduling, more Zoom calls. I think I’m able to do more, but I’m a little worried of going back into the regular world where I have actual traffic and driving places and physical things interrupting it. And when I was working my way through your website, I noticed you used the phrase remote first. So in that world, how do you see the future of work? I don’t think we’re going to go entirely back to normal, whatever that is, but how does Nylas see the future of work now as we’re hitting the end of the pandemic, hopefully?

Hopefully, I really hope so, knock on wood. But yeah, remote first I feel is the logical way to go, because otherwise you’re sort of cutting yourself off from a lot of talent, and especially from folks who just have different styles of work, or who for the past two years have adjusted to a new style of work and found that hey, this actually fits their lifestyle a little bit better. I mean we still at Nylas for ourselves have offices, I don’t see offices going away anytime soon, but just my personal experience is that I would say probably 40% of people do really well in the remote work world and kind of have that as their preference. Another, let’s say 30% really prefer being in the office, they need that human connection. It tends to be a lot more of the interrupt-driven functions, like sales want to be more on the human side, and a lot of the more heads-down functions, so obviously engineering or marketing, the ones that interrupts and context switches are super expensive for, they tend to seem to want the remote work side. But that’s a generalization. There’s plenty of folks I think on all teams that have one way or the other of working. And then there’s a group of folks that is sort of ambivalent, they can do it one way or the other, they just want to get the job done and get to work on the stuff they want to work on. So my thinking is, it feels like the obvious pragmatic answer of, let folks work remotely if they want to work remotely. Worst case it just reduces the amount of investment you have to make as a business in real estate. You’ll need less space because not everyone will be in 100% of the time. And I kind of believe that culturally it’s a forcing function for a lot of really healthy organizational behaviors. So when the pandemic first started, I’m sure you saw all the old school managers freaking out about, oh, what about the loss of in-person XYZ thing, how will people be productive? And well, people were productive after all. And my personal thought around it is that it’s because they kind of sucked as managers in the first place. If you feel called out, good. I think it’s because you were micromanaging too much and you didn’t know how to set clear goals. And our philosophy at Nylas is that leadership’s job, and I sort of mean that positionally, positional leadership, whether you’re the CEO or whether you’re the tech lead on a project with no one reporting to you per se, your job is to outline where you’re going and why. Here’s the direction, here’s the context around why I think that’s the right direction, how is up to you, you tell me how we get there. For a few reasons. One, it actually gives people autonomy and agency and impact and ability to make their own choices, which, surprise, people react better to. Also it gives a little bit of a feedback cycle for folks to be more collaborative and go, okay, that direction is great and I get that context, what about if we went 10 degrees to the right, wouldn’t that be better? You get to go, okay yeah, that would be better. But mainly in the context of the remote work function, it really forces leadership and management and organizations to get super clear about what they’re trying to accomplish and why. What are the goals you’re trying to get at, how are you measuring those goals? And if those goals are accomplished, who cares how long someone spends in an office, or even if they put in more than two hours of work a week? Goals accomplished. That’s on leadership to set a different goal next time if they feel like they’re not being efficient enough with their staff.

Yeah, it feels like folks who really don’t have a handle on their workforce and what people are doing are going to have a tougher time automating a lot of that, because they’re just unaware of what happens and how it happens. So where do you recommend folks start thinking about business process automation? Where are the most common ways that, if it’s kind of chaos right now, where do we start?

Start at the top of the funnel I guess. Whenever you’re optimizing these things, figure out where the most of the time right now is being spent, what the biggest lift might be that will give you the biggest, most outsized win. When it comes to things like the business process automation piece, I would say really focus in on how you’re interacting either externally with potential customers, users, prospects, potential team members that you’re trying to recruit, or look at how folks are interacting internally. Because like what you were saying before, your day is spent a lot on communication, and I would make the argument that everyone’s is. We are humans in meat space, as much as we want to deny that, or as much as I personally sometimes want to. But that means that day to day our monkey brains are just there trying to reach out and connect with other monkey brains and influence them, whether it’s on TikTok or on GitHub. It’s a weird sentence, but those are the two extremes of the spectrum I’ve just laid out. And whether you’re a marketer, an engineer, a salesperson, your day is ultimately spent talking to other people and collaborating with them and trying to create useful, meaningful experiences together.

So when it comes to automating those things that happen in my daily world, is automation just about scheduling things, saying at this time do some sort of automated task, or is there more to automation than just scheduling?

I would say that automation as a general concept, and I am keeping it high level and vague right now, applies beyond scheduling. It applies to any process. It means simplifying that interaction surface area. So like driving back to that car analogy, or going from machine code to Python or other languages, how can you make it simpler to achieve your end goal, do the thing with fewer inputs? That has cultural context that needs to come into play, in terms of you have to constrain the possibility space of what you’re trying to accomplish if you’re trying to also constrain the input space, for the most part, or you can get really complex around the functions that you’re going through. But when it comes to Nylas specifically as a company right now, the reason I’m not jumping in with a bunch of stories of like XYZ this person did that, is because our focus is on being that API economy player. We’re not the ones doing the cool, our customers are the ones out there doing the heavy lifting, sending out all the vaccines, empowering folks from underrepresented minority groups to have access to financial capital markets, or for some of our more fun use cases in the IoT space in automotive, or showing LCD picture frames, or selling caster wheels, or whatever the situation might be. We’re providing the infrastructure. And so I sort of keep it vague because I want to focus on our customers that are doing the hard work and just giving them the most flexible tools possible to get their job done.

And are these developers you’re giving these tools to, are these average business folk developers?

Yeah, so our customers are always the developers, the product teams that are trying to build things.

But you’re really abstracting away all those friction points that they normally encounter, having to deal with the different APIs, deal with the different SDKs and all of that. You’re looking to make our life as easy as possible when it comes to that.

That’s right, yeah. I mean that’s, if software’s eating the world and developers are building that software, I want to be in the hands of every developer helping them get their job done better, faster, easier, cheaper. I kind of see that as being the necessary place to play if you want to have the most impactful business possible. How can you accelerate the pace of software? But that goal itself is an instrumental step towards how can you accelerate the pace of, you could phrase it as a business pace of opportunity, pace of the ability for technology to provide people with better access to those human experiences or better access to whatever type of opportunity or life they’re trying to build for themselves, which is why humans have been making tools for the past 100,000 years.

So how are you seeing the concept of an application evolving? When I first started tuning into the API world in 2010, it’s very much you build it they’ll come, you build an API, developers come and they build a web or a mobile app, and then eventually an IoT app. How do you see the concept of an application evolving in a Nylas-powered world?

Great question, let me think about that for a second. I don’t know whether I see the role of a discrete application fundamentally changing all that much, though the interface may change. The way I think about that landscape is in three vertical buckets, where applications are at the very top, they’re the top layer, the cream that most people see. And I’m defining application here as anything where the experience and the value you get from that software is driven primarily or almost entirely through that end user visual interface. So there I would include everything from Salesforce to this Restream to things like Honeycomb, a log diving tool, where one way or the other you’re using it in an interface, going in and visually pressing buttons to live in that experience they created for you. Then below it I see there being, at the very bottom, an infrastructure layer. So these are the folks that are laying down the fiber optic cable that sends signals across oceans and sending up Starlink satellites and running the bare metal servers themselves for the cloud providers or what have you. In the middle, that’s where I think that API economy piece, as we were talking about it, lives. I think of it as the abstraction layer. The folks, API economy, digital transformation, low code no code, I think that all fits in that abstraction layer bucket, and the common theme is companies like Nylas, Stripe, Twilio that are making more powerful tools, more powerful Lego blocks, IKEA kits for folks to go with. And you can kind of think of that application, abstraction, infrastructure layer as mapping directly to a restaurant, grocery store, farm type idea. Where I really like the API economy, I like the abstraction layer space, because if you’re like the grocery store, you’re able to sell not only to the restaurants of the world but also to home chefs and to companies that have cafeterias, and you provide little meal kits that make it easier to get the job done while still giving people all the flexibility they need. And from a business perspective, as my brand of evil business guy that I’m establishing here, that’s the much more valuable place to be. There is no one right answer to restaurant, there is no one right answer to video streaming. Zoom’s doing a great job, Skype did a fine job before, there’ll be some other video streaming tool after, we’re on Restream right now, right? But the reason for that is that when you constrain the possibility space, the business advantage you get is rapid distribution. Because if you really nail the value you’re driving for a specific user segment, then you’re able to really just step on the gas in terms of getting it out in people’s hands. There are viral loops you can use, there are huge paid partnerships that you can do, or just dump money in ads. As long as the constrained experience is relevant to the users, you can get broad scale distribution quickly. You can open McDonald’s in a lot of cities. It’s harder to open a French Laundry in every city, and it’s harder to say that McDonald’s is the perfect answer to all your food needs.

Yeah, well you used one of my favorite analogies for APIs in the space, legos, but I think you took it one step further that I think is really important, you used the IKEA. And I would say that’s very powerful, because especially in the gig or sharing economy and the empowerment you talked about with your customers, is you, Nylas, isn’t super concerned with whether you’re in the restaurant business and building IKEA furniture or IKEA templates, API driven templates for the restaurant industry, or if it’s healthcare. You’re enabling those folks who are building those cool types of kits that can be rapidly applied and used and iterated, and then Nylas evolves your platform. And for someone like me, I’ve worked and lived and breathed in a tech space, so I move to New York, I get rid of all my furniture, I move my art, my books and my clothes, and I order an entirely new house from IKEA. I move to Seattle, I do the same thing. I move to LA, and that’s how I live my life. So if you think about that from a business standpoint, how do you automate, how do you do that in a business world? I think that IKEA approach is pretty powerful to think about.

Yeah, and right now, talking back to that identity, payments and interactions piece of things, I feel like the identity space is fairly well understood, the payment space is getting there, if only because of the fights with the regulatory space. I feel like this communication sector, this interaction sector, is still at the very beginning part of the curve, because it’s a hard problem. Because you can’t constrain the use cases nearly as much, because we’re talking about the possibility space of human interaction across all potential software in theory, and that’s harder to build a universal layer around. You have to cover at the end of the day every channel. Right now unified communications, if you look at Gartner, it means telephony, it means call centers, it doesn’t mean email, WeChat, everything else. So it’s starting to a little bit. But when I think of universal communication, when I think of universal interaction, that’s where my head goes to, is okay, where are people actually, where do they actually live? And the reason we began with email in large part is because the answer, much to everyone’s chagrin, our own included, is still email. There’s four and a half billion mailboxes out there today, that’s not, I’m sorry, let me rephrase, 4.5 billion people with mailboxes, the number of mailboxes is way higher, but human beings with access to the internet have email addresses. SMS usage is declining, in large part because it’s being broken apart into the WeChats, the Facebook Messengers, into these corporate owned channels. Email usage continues to go up, and that doesn’t seem to be stopping. Part of that I believe is that it’s a decentralized protocol. Every time that the internet is up, your email is also up, type of thing. You can’t say the same thing for Facebook Messenger. But oh, I’m not making the claim that email’s here to stay forever, and I don’t even really have a, I’m agnostic to channel at the end of the day. I’m just making the observation that email’s a big one and growing. Slack or whoever else coming in changing the game and how folks communicate, that’s awesome. Whatever happens, we will still need to make sure that we’re able to build software quickly, to make sure that we’re able to give every developer out there, or every person who’s interested in becoming a developer, the tools they need to accomplish those end outcomes quickly. They’ll need to be able to interface with the depth and breadth of the human possibility space of how people are interacting online, over the phone, in person. How can software help those? And I think that’s going to be the biggest trend we see over the next five years, the way that companies build and buy software, the way that people build and interact with software, is going to see a pretty dramatic acceleration and change. A lot of factors driving into that. A lot of the great work being done by the companies in the abstraction space, a lot of folks that I don’t think we mentioned on this call yet, like the Retools, the Webflows, the Vercels that are out there working on making it easier to build these types of applications. That trend is gonna keep accelerating. On the business end, companies are gonna remain not having as many developers as they’d like, and they have, I hope long ago, it may have been only a year or two ago, finally figured out that, like it or not, they have to have a tech stack and engineers and they are software companies, like it or not. And just imagine once there’s global low latency internet everywhere, what new avenues that opens up for things like IoT, for agricultural processes and sensors in far out locations, for the educational opportunities and legitimate real economic opportunities, not just beyond education of building some useful software quickly and easily that solves a constrained specific problem for someone who could never do that 10 years ago or even now. That’s what gets me excited about the future of the API economy.

Yeah, it feels like you all have a real good handle on, because that human focus, that relationship connection focus, you are tuned into the context, and whether that’s healthcare and HIPAA needs to apply, or what can be very social, or whether it’s a voice and it needs translation. I think that rapid context is pretty key. So how does machine learning and intelligence fit into all of that?

Yeah, I think that context piece that you’re highlighting there is the critical part. We have a phrase at Nylas that we like to use, connecting and collecting data is easy, but putting it to work is hard. It’s getting that end outcome, that end result of what you’re trying to do, that’s tough. And it’s not because you can’t build a pipeline into a database quickly, you can, that part’s not hard. It’s because, one, structuring the data in a way to make it useful at all, and then two, having the context around it to make it useful, that’s far and away the hard part. So for us right now as a business, we work in something like between one and two dozen industries that we have a good amount of traction in, something around 800 customers today with around 20 different industries that get a lot of utility out of us. We’re at a scale where we’re at just over a million transactions a second, like two to three billion API requests a day, which is bananas volumes thinking about where we were just six years ago. But because we have this context, because we see the 50 some terabytes sinking down daily and reinforcing our models, that context engine that we have going, that allows us to then tweak and make ever better the common use cases we see today, and allows us to future proof ourselves for whatever use cases we see coming down the pipeline. And by focusing on that more universal human element of, okay, if you’re trying to get information from someone, you have to first make a connection to them, and then you have to tell them what you’re trying to get or collaborate with them on, and then you have to get their consent and buy-in, amazing, that doesn’t matter if you’re a salesperson or an engineering manager, that’s the same process and the same flow. And so when we’re thinking about that future of software, the what should you automate from a business product automation standpoint or whatever else, it comes down to that human connection and element, the interaction between people and people, and the interaction, even though this one I think is a little bit easier to do today, between people and data, is where a lot of the work of software has been happening, and where the biggest impacts remain to be had in the future.

Well, in this new kind of human and connection driven reality, you’re an API company, you’re servicing developers. Similar to my question around how’s the application evolving, is how’s your definition of a developer evolving? Because in the old days, developers were someone you kept in the basement, you kept behind locked doors, you threw pizza in there, you didn’t really put them in front of your customers. What is your desirable view of what a developer is today in the Nylas API ecosystem?

When we first opened up our New York office, we got a ton of folks, it was an engineering office primarily, who were coming from these super emotionally battle-scarred financial sector type jobs where it was exactly like what you described. It was like they had zero decision making power, and it sounded horrible. I frankly don’t super understand how you run a company that way, but anyway, I guess when you have all the money, literally you can. But how does the role of the developer change, is that the question, over time?

Yeah, I mean just what is, as far as what Nylas is looking to hire, or the talent that you would love to see in your developer ecosystem? What’s the optimal profile of a developer?

Great question. For us, selfishly as a company, we’re hiring pretty aggressively for back-end engineers and folks that have an interest in working and processing super large, rapidly changing data sets on the AI/ML side. So we’re trying to at least 10x a lot of those numbers I said, the number of API calls a day, the data a day, the transactions a second, by the end of the year. We’re making great headway and progress on that, but we’re still a fairly lean team overall. And on our end, I mean we’re an API company, we sell to developers, for developers, to make developers’ lives easier. So for us it’s sort of like step one is putting our engineers in front of our customers’ engineers and having them come up with the ideal solutions. More broadly, Nylas aside, I feel like there’s, I mean there’s what, like 50 million, 50 to 80 million developers out there today, ballpark 50 million or so, yeah, it’s gonna about double the next three years. But even the folks who will be graduating in 2025, the class of 2025, with college degrees, even if they come from completely non-technical, I’m trying not to throw shade on any specific university or major, but even if you’ve only read Chaucer for the past four years in university, you understand so much better what APIs are, how software works, that there’s a shape to it, a logical structure, a Minecraft block, a Zoom integration that has to go in, and the concepts of garbage in garbage out, conditional statements are there in the cultural zeitgeist, and that’s only increasing more and more. So you’ll be able to use a lot of the tools that are in their nascent period today around low code no code type flows. A sort of analogy between digital natives, folks that are just used to using software from birth, I think there’s some sort of equivalent to programming natives that we’re gonna see more and more. My tin foil hypothesis is that five to ten years from now, developers are going to look a lot more like architects do today, where they have to deal with the winding together of components, Lego blocks, IKEA kits, and dealing with all the dependencies, than they will on, okay, from scratch creating those microservices themselves. That’s not to say that those jobs aren’t going to still exist, and those jobs will in fact be higher in number, but because of the pace and need for software development being so much higher than the pace at which those jobs can be developed, that’s what gives the room for the more abstracted API economy abstraction layer, low code no code tools to exist and to allow that new category of developer as architect, architect as developer, to exist more and more.

So I mean for Postman, I’ve been at Postman for two years now, and we’ve gone from 10 to 15 million developers in that time period, and it’s just scaling as well. And so where we have numbers we’re looking at, you know, 100 million developers within the next five to ten years, and like you said, there’s only 50 plus million, so we’ve got to create these developers. And so if it sounds like this next generation developer, you feel like should be more in tune with the business requirements, the business needs?

Oh yeah, needed. Well I think they have to, right? If every company is ultimately a software company and developers are the ones building that software, then the value of a developer within a company is going to be increasingly valuable, and the ones that can best understand the business needs are the ones that are gonna be able to best build the right software. It’s gonna become more and more a requirement, that understanding what to build is understanding why the business needs to build that. And I think the pressure is bidirectional, by the way. So I think I’m also highlighting that it’s no longer going to be okay to just get away with saying, I want some software that does something, and then get a response back, what are you talking about, be a little bit more specific about your requirements. The only people on the business side that are going to be able to survive are the ones that are going to be able to interface collaboratively with their development teams, their engineering and product teams. That’s my strong belief.

Yeah, great. And I want to encourage, I mean people listening, I’m trying to get people into APIs because I feel like we need more of those problem solving people. And when I first got into programming, I won’t date myself by saying when, it was low level, it was pre-internet, and I didn’t like it, I hated programming. I was just like, I’m not going to do this, and I dropped out of school and went in a different direction. But once the internet came around and things were abstracted away, I was like, okay, this is a little more interesting. And I feel like that’s what Nylas is bringing to the table, it’s gonna abstract things away and go, all right, I can actually solve a vaccination clinic problem here, I can actually solve some sort of meaningful workflow in my industry or business without having to learn all those low-level annoyances of the programming language, of the APIs, and all of that is abstracted away. So you’re going to get people focusing, like you said, focusing on the problem, on the thing that they care most about, without the worry about all those technical details. That’s pretty huge.

Yeah, well it’s something that technology is empowering, right? And like for us specifically on the Nylas end, we had the hardest time historically finding traction for fundraising with investors, with people who aren’t our core customers or audience, because they saw, all right, well, at a high level you have documents, emails, calendars, contacts and SMS APIs, like what can you do with that, isn’t that just like five things? It’s like, well no, not really, it’s like a million different things everywhere, because again it’s the universal constant thread in human interaction right now. So the fear was, well, you seem to have a, what was the phrase that was used once, a Chinese menu of options around use cases that you do, and you do this for healthcare, this for logistics, this for health tech, this for sales, this for CRM and real estate. And by the way, we just checked that something like 45% of the country’s real estate transactions are going through our servers at one point or another, which is so cool to see at scale, that happening either on the front end like Realtor.com browsing for houses you can never afford side, or on the mortgage underwriting and lending side, which speaks to the power that I think that API economy and ecosystem has. But I’ve distracted myself from my own point.

No, it’s all good, I do that constantly, that’s why we got this nice list of questions here and I kind of route ourselves back. But so, you guys just got a lot of funding, obviously growth is in the cards. What are you going to be investing in, what’s big for 2022 and beyond?

Scaling and growing the engineering team, and scaling and growing the sales team. So basically growing the company. I think right now we have a ton of inbound interest and we want to continue developing and building out that product line. The opportunity we’re heading towards, pointing ourselves at, I believe is a pretty obvious one. Even folks at the application layer level of things are trying to head in that direction, though I believe they’re never gonna get there for that reason that McDonald’s isn’t the only restaurant you’re gonna need, but onions are gonna be the only onion you ever need, so sell the onions, don’t sell the blooming onion, or whatever type of analogy we want to go with. V, sorry I did it again, knocked myself off my own thought. Investment, growing.

I mean, I feel like you guys have a proven formula. People want to be able to solve these problems, build new workflows, and you were already doing it before the pandemic, but I think everything points to, hey, we’ve got to be able to move fast, be agile, we’ve got to work with the services that we’re using already, our email, our CRM. And as you said, what’s in my email inbox is so critical. There’s calendar events, there’s receipts, there’s invoices, there’s all of this activity going on, and I’m in Zoom, I’m in Restream every day. So I think your Chinese menu formula for responding to industries is proven, like you just said in the real estate industry specifically, healthcare, when it comes to being able to adapt fast, move fast and do things.

Absolutely, yeah. I think I can wind those threads back together now. So the argument I’ll make is that it’s not a lot like a Chinese menu, I think it’s a lot more like a Mexican restaurant, in that the common ingredients, tortillas, rice, meat and cheese, whatever you look at, it’s gonna break down to those things. You don’t have to build completely new products, completely new experiences per industry, you just have to make sure you get the core ingredients right. So for us, what that means is our focus right now this year, for the next 12 months, is three things. Developer experience, first and foremost. We want to make it as easy as possible for developers of any skill level, whether you’ve been that deep database engineer for the past 15 years, or whether this is your first time touching a terminal, or you’ve just done a boot camp, I want you to be able to, using our software, in less than a business day build something cool and useful. That’s the core idea around developer experience to me. When I say developer experience, I mean I want folks to be able to do cool stuff quickly. The second piece is scale, so hiring, making sure that our company culture and internal systems are able to keep up with the scale that the business is demanding. And the third being that go to market excellency piece, where I really want to make sure we are building out the Lego instruction sets, the IKEA instruction sets, that actually enable that developer experience, so that folks can come in and say, oh, I really want to build the Eiffel Tower, and we’ll say, oh, I’m sorry, we don’t have Eiffel Tower, but here’s our recipe for the Arc de Triomphe, and here, you use this piece and put it like that, and look, there you go. That’s the core idea, building that out across all of our existing use cases, industries, making sure that we’re empowering the developer community and ecosystem to go into the ones that we haven’t ourselves touched on, and then adding the open source components to empower them to do that on their own. That’s what we want to focus on and deploy the funds against.

Yeah, those templates and blueprints are going to be key, because dumping a massive bucket of legos in front of someone only gets you so far. And that’s what I feel like the API space has been with the Twilios, the Stripes and others. They’re powerful and you can do a lot, but you’ve got to have an imagination. I think a lot of the next wave of developers need those templates, like you said. I need to know what works for selling real estate, I need to know what works for doing email marketing in healthcare, and I need a proven template that works and that scales and that I know is going to be reliable. So that’s pretty key. So, I think we’re approaching the hour here, so we’re getting down to it. One another, so I’m gonna be using your analogies. So the IKEA one I’m gonna explore more and work with and reference you guys, because that’s huge. Another one I want to share with you is, our director of labs and one of the original Postman, Postmanauts as we call them, Shamasis, he runs our labs. And when he came over, I was like, hey, I’m really into Mexican food, and you talked about the different parts, whether it’s, you’ve got a tortilla, you’ve got pork, you’ve got chicken, you’ve got beef, you’ve got different sauce. He’s like, I don’t need Mexican food, I’ve got Indian food. And I was like, oh, that’s kind of snobby. And I didn’t think about it, and then my wife, we’ve been cooking during the pandemic, and she’s like, okay, I want to explore this, because you have naan and you have sauces and you have pork and you have beef. So like, how do you make it, I mean what I see Nylas as is, yes, here’s all these templates, but it still allows for innovation, for you to provide different contexts to international audiences, industries, and you can rapidly adapt existing IKEA templates or blueprints for different industries based upon what works in real estate, but we’re going to try this in shipping and logistics. And it works, but all the tools and everything. So I really feel like that’s what you guys have your finger on, is that abstraction layer that’s going to be pretty key, but it still allows, because you’re developers building for developers, you’re going to be able to innovate within that space, and people are going to be able to do some pretty interesting templates. And I had a question in there and I totally lost it, talking too much.

No worries, I feel ya. But I feel like that’s sort of the challenge right now, what you’re describing of going from tortillas to carbs is the challenge a lot of folks in the API economy are having. Even on the Nylas side, we do the tortillas, rice, meat and cheese super well. The future is figuring out the actual macronutrients portion of it, like how do we go from naan or tortilla or rice or whatever, and level that up to carb? Every culture has some sort of dumpling, some sort of sandwich equivalent. How do we make the flexible, robust, decomposable, modular software framework that empowers all of that? And so on our end, that’s why we’re hiring so much on the engineering and data science side, because that’s where that context has to come from. Right now a lot of our AI is supervised models. How do we generalize that and go to the more unsupervised models and structures that allow for that next leap in abstraction, that next Russian nesting doll to be layered around the potential and possibility of what’s being achieved? And I think we have a pretty huge head start when it comes to that just given the scale of data and access that we have. The contents of your mailbox in a lot of ways is the database of your life.

Absolutely.

The database of your business. It’s where the highest signal lives. It’s where you can get the sentiment analysis for how are your employees doing, what does your real org chart look like, who’s actually meeting with who, who has the positional influence, what are your customers actually feeling about you, why aren’t they taking your meeting, what should you say next, and who can you get an intro from. Again, these very human questions. That’s what the context and data and structure helps you provide, and then really quickly build against to provide new value, to build and create something cool and interesting and just awesome as a developer or as a business person at the end of the day. That’s sort of the goal. And maybe I’ll turn this one around on you, because I see we’re at the end of the hour, but what in your mind, I’ve asked a couple dozen people this question and the answers tend to be pretty consistent, so I’m just gonna free white space it, what has been the consistent theme between the best developers that you’ve worked with?

Developers who aren’t afraid to ask questions, humble developers, ones who listen and ask those questions at the right times, they’re not too talkative, but the ones that can speak to the business problems. I’m a big fan of recruiting evangelists and devrel folks from business circles, not from developer circles, because they’re closer to the business problems. And then once you get them close enough to APIs, and this has been a mission of mine, is explaining APIs to the normals, to the muggles of the world, because I feel like they’re going to be the ones that are going to actually solve problems, where a lot of us developers and tech folks just do APIs for the sake of APIs. Look, it’s right, it’s a perfect API. What problem does it solve? Well, I mean you can add contacts and you can, well no, what problem does it solve? So speaking to that problem and having some sort of domain expertise, either in healthcare or real estate, and grabbing people out of those industries, and then I wouldn’t say just turning them into developers, turning them into tech savvy, curious people who are gonna poke, prod, and try to understand things. That’s what I’m looking for. And a developer of the future for me isn’t someone necessarily coding, because I spend very little of my time coding, and I’m technically a developer, but I spend a lot of time connecting the pipes, seeing how it works, playing with APIs, seeing what’s possible, comparing two different similar APIs and going, okay, what’s the responses look like, and I’m writing very little code. And that’s one of the reasons I’m at Postman, is because Postman allows you to do that. So that curiosity is a pretty key one for me.

Yeah, well I could not agree with you more. I think that’s actually the universal answer that people come back with. The best developers are the ones that listen, that collaborate, that have high EQ, that are able to explain the context of what you’re working on, and that are able to help and mentor people who don’t understand that space, enabling them to move faster. Never has anyone said, oh, they’re the ones that really understand Cobalt, or they’re the ones that can really get Cassandra up and running somewhere. Never. It’s the exact same universal answer to, if you ask someone what makes an awesome marketer, what’s the common theme between every good salesperson you talk to, it’s oh, they listen, they actually help me accomplish things, I felt a connection, and that together we could build something greater. That’s that universal theme. And that’s why focusing on software that connects people and helps them get something done at the end of the day is so exciting for us, why we love, but we’re a Postman customer, we use y’all to help folks navigate our docs, use the software, we also use it internally for internal testing and debugging customer issues, because it allows that next step. And I don’t think we’re there yet, I think it’s five to ten years out, but that next step in the way we think about different roles in business, maybe the role of business overall, in accomplishing things in service of people’s interests.

Yeah, no, agreed. And so moving forward, I think I’m going to be using the IKEA template model and blueprint and referencing you for it. And as I talk to different startups, enterprise organizations about how do you articulate the context and the value, not just business value but value of an API and what it does, and if you can’t explain, like, this is a coffee table that sits in the corner that does XYZ, or it’s this in your kitchen hanging to hold garlic, if you can’t explain what it does in those terms, and then provide some sort of blueprint for someone to put it to use, then it doesn’t bring a lot of value to the table. So I’m gonna probably be pinging y’all for some help on that one, because I think as I work my way through different API ecosystems, I think that’s gonna be a pretty powerful analogy to work with. So I appreciate that one.

Yeah, absolutely.

So I think that’s about it, we’ve gone a little bit over, but that was a great conversation. I appreciate, even though we wandered a bit, that’s what I loved the most about it, because we both lost their way a few times, and I think that’s what this is all about. So I appreciate your time today, Gleb.

100%, thank you so much, it was great being here.

All right, thank you.