Nikita Shamgunov - founder of Neon: storytelling, pricing and hiring execs

And that was the era of PHP MySQL. And so I wrote a lot of PHP code in my life. All the programming competitions were in C plus plus and eventually I became a systems engineer. And then the Nassar is so when you do programming computations, it's about algorithms and data structures and elegant solutions. And databases, they kind of have all that.

So databases is computer science in a box where there's a piece of operating system piece of, algorithms and data structures on this data structures typically be trees and then there's a query processor with a bunch of algorithms there's and the hard problems inside the database of where it was an optimizer. And so I didn't quite realize all the spectrum at the time, but they, that that was kind of gravitational pull for me, kind of translating from programming competitions. Then I took a pause. So I went to do the grad school, worked on, compilers and then, also worked at the same time and worked on, like, navigation systems, which is yeah. It was just an app.

It's an app, written in C plus plus of all the things, but it's an app. And then I joined Microsoft, and and in the app, I used SQL Server. And then, I got an offer to work for SQL Server. And so that was exciting. Microsoft flew me over, gave me $20,000 budget to relocate.

Lots of scars when you build system software, the cycle is so long. And the cost of of wrong decisions, which you believe in, is high. So because, like, backtracking is tricky. And there's one thing I learned later at Coastal Ventures is to constantly provide a narrative of what you're doing at the high level. So somebody wakes you up at night and is like, what are you doing, Ann?

And then you're like, well, working on this and that in the grand context of this. And you can explain what the grand context of this is, and being able to defend it and segment it into, well, these are the obvious things. Right? If we're building that, we need to do that. And these are the secrets. These are the things that we believe in, the world doesn't. But if I explain it to you why we believe into this, I'll wheel you over. Right? Because logic. And so, we didn't have as much of that.

We have a lot of rah rah, let's go build system stuff at SingleStore, and we had an incredible assistance team. But that piece is what I learned over there at SingleStore. When I tell the Nian story, people are like, yeah, makes sense then. Like, go do it. They they they root for us. They root for these things to exist.

Right? And I think in this particular case, I have passion for both. SingleStore mostly like, SingleStore is a broad platform, but it can do a lot of tricks. Dollars are mostly coming from data and analytics, or real time analytics. And that's why having this bot platform helps because some foundation of it allows you to do real time analytics.

But the user of that is some sort of big data person. And while I love them to death, the the connection that I have with actual software developers is higher. Yeah. Right? People who are running code every day, that people who are deploying stuff, the people who are using all this modern dev tools, the people who use you know, these days, they use cursor and open and co Copilot and all this, like, crazy AI tools.

I have more connection and and sympathy to that type of user, and Nian is built for that. And Nian is sophisticated from the system's perspective. So it sits on an on them, pretty, hardcore. I think hardcore is the right word because complex can mean good or bad. So it's not that complex, but it's very hardcore.

Right? And in places where when you, you you know, reliability and performance, we we go down the rabbit hole very, very deeply. And on that hardcore piece of software sits very, very nice developer experience. And that is not something that we were obsessing at Singlesorm. We are obsessing about this at EAN.

And the reason we're doing that is that because of the user, because who are we building that for? So so, yeah, that that that's that feels pretty perfect for for for me personally, the kinda the kinda technology and the kinda user and the kinda, and what you obsessing about every day. So that's pretty good. Also, Vee doesn't have a sales team, really. So we have a partnership team and, like, a tiny bit of sales.

It's like one salesperson. And it grows on itself double digit percent, months and months. And so that's nice because you can translate that growth to to venture dollars, which you translate into engineer product and engineering work, which translates to better growth. And, that's that's also pretty exciting, world to live in where in this world, there's a lot more predictability in your business.

Right? So if you're only up to, you know, obviously, like, the scale is relative, but let's just, you know, pull some numbers out of the air. Let's just say, like, if it's point six or something like this, then you will grow at that rate, and you will cap, your revenue at some other rate. And the reason you cap the revenue is because that's when your attrition will match your growth. Mhmm.

And so as you approve your product, you approve the product market fit, you reduce attrition, you increase retention. And that allows you because you every day, you're onboarding, you know, fixed number of users. And then, hopefully, that number of users keeps growing, as you get the word-of-mouth thing going. And then you either retain them or you don't. And you retain them or you don't based on the quality of your product.

And so as you keep improving those things, that's why I have a growth team. That's why you have observability. That's why you obsess about latencies. That's what that whatever you have, you obsess about instant gratification and good experiences. So, that improves that motion. You know what I mean?

WorkOS is trusted by Perplexity and Vercel as well as Workbrew, a home brew management startup that I recently interviewed. I just told Mike that WorkOS is the sponsor and this is what Mike said.

This seems expensive, but then I built an integration with them in about twenty minutes. So I had spent two days banging my head off the wall trying to build it directly with Okta. And then with WorkOS, I then have, like, many, many SSO providers, like, support instead of just one. So, yeah, like, for me, WorkOS is one of the nicest developer experiences I've encountered in the last, like, five years probably. And it's so surprising because a bunch of the developer team are at GitHub and therefore very good at the job.

So our our brand awareness is, you know, is high in some places, but, you know, kind of zilch in some other places. And so so we start with awareness, and then you see where those people cannot self serve. And that's where we introduce humans. So so you need, you need the full machine of sales and marketing with pipeline generation. And then once you have pipeline, then you can have SDRs and, like, all these things, eighties, yada yada yada, all this heavy machinery.

Very expensive. Right? Mhmm. Because all these people, if they're not selling, they're just burning cash. Right? And then he's like, well, let's add five SDRs and, like, 10 aEs and five SEs. And that's, like, 20 people right there. So multiplied by, like, I don't know, hundred digits to hundred k a year, fully loaded cost, that's that's quite expensive.

And then from there, you have Fortune 2,000. And from there, you have big mid market. From there, you have SMB. Then there are more people in the SMB than they are in the mid market, that they are in the enterprise, or at least more companies. People I don't know.

Dollars wise, the pyramid is almost inverted. There are more money in Fortune two thousand than in the rest of the pyramid. And, but you graduate through the pyramid. And in every new segment of this pyramid, the sales motion might be different. So in the Fortune two thousand, the sales motion is there's a Salesforce.

And but you need to get to that Fortune two thousand. And you need to be ready, and you need to be ready from the security and compliance standpoint. And, also, market side is different in every segment. And so we go and keep elevating our game, and being able to get to the next layer of that pyramid.

But reliability and performance are not being tested really because well the workloads Pre chair workloads are small and so you find surprises as you push up and of course you have internal testing and all those things at single store. We actually were much better at testing because we shipped on prem software. We shipped software and we shipped the service. And then a cohort of users, you know, want it want it on prem. I hate on prem with a passion.

Well, the reason to that is, like, the amount of time I spend on, like, at the time it was Webex and then Zoom calls. And then you give, you know, people give you access to the terminal. So you feel like you're doing brain surgery through a keyhole. And so it's just not fun, versus in the cloud world, you control the infrastructure. So you can go and and optimize the layers all the way, and you you manage these things effectively.

You have tools rather than trying to adapt to the tools that exist on prem. Anyway, I'm prem parking tops.

In OLTP, in database operational databases, it's volume. You still wanna charge for storage and compute. And a lot of people like Mongo. Right? If you go look at Mongo pricing, Annapolis, there's, like, a lot of sizes there.

So you choose the size of the instance, and then that's it. And then because the end is elastic for storage and compute, we can just charge people for storage and compute. Don't quite quite do this. We used to, and then we would have a free tier, and then there was a paid tier. And we're, like, start at zero.

And then it was confusing people. And so we decided to have a barrier. And so we now, like, $019, which which includes some consumption, some storage, and compute, and the rest is consumption. We're actually simplifying that. So Really?

We we rate one more time, and we're gonna have a serverless plan, that starts at $5. We decided that zero is not fair for us because, you know, even the Stripe swipe cost you something. So you can't really charge somebody zero. But, we have this advantage because we're serverless. And you can have unlimited number of databases and still not pay us very much because we're serverless.

Maybe if you're not using on the system, so so we decided to compress our 19 and $69 plans make it $5 plan plus consumption. And when you need and then you can buy add ons. Add ons is like Shipa or, you know, compliance this, compliance that, security this, security that, private link this, private, you know, stuff like that of of that nature. And the sum of all add ons is roughly adds up to the business plan, which is, 7 to $700, plus consumption. So that's what we kinda envision to to do with pricing.

And, it's very different between, again, single store and here, single store, like, small number of customers, large contracts. Here, lots and lots of customer. We'd like we have tens of thousands of customers now paying customers. And, lots lots free. And so anything we do with pricing, now there's modeling involved.

So there's, like, a dedicated person, who is doing lots of audits. Sec. Well, I'm I'm giving you the the final product. We're almost fine, and we might might do one more pass. So, you know, the numbers might end up not quite exactly the same that I mentioned, but most likely they will be. But maybe our our modeler will discover something and we're like, no. No. No. Like, we actually need to do it slightly differently. But then we kinda know what's gonna happen ahead of time.

Mhmm. And so that's the world of of PLG. That's the world of lots of smaller customers. That's the world of just not really having much one on one conversation with your users, but, talking to your users via, your content and talking to your users via, you know, Discord channels, and and via, like, shipping certain things and testing.

If you have competition, you need to study their pricing. If your technology is doing the same as the competition, but somewhat in a different way, you know, serverless versus, you know, fixed size, something, there's some innovation, then you can anchor your pricing against this innovation to highlight your advantage. You can either do it or you can't. And if you can't, then you price the same way as competition with with a 20% margin over what they charge. Maybe maybe 30.

I mean, in the wild greens, 50, but, like, online.

They care about value. If you're twice as expensive, they kinda care. You may still get away with this, but they kinda care. But if you start looking at tech products and their pricing, you will find that they're remarkably tight vis a vis, the competition's with lane. Right? And then my ChargeGPG versus Anthropic versus, perplexity example, shows us this is what it is in the consumer market.

And over a period of time, in the fullness of time, everybody builds everything, right, unless they die. Because, you know, it's kind of like if somebody comes up with a good idea and the market responds to this good idea, like, then then it's gonna be copied 100%. And and then certain pages, it's just hard to copy for us at storage. It's hard to copy our storage. But, you know, people are working on this, in large clouds, in, competition.

SuperBase working on this thing called, Aureal DB, which has some promise. You know, Microsoft is working on something. Google introduced AlloyDB. Like, separation of storage and computer is a good idea. You just can't put it out there in, like, two days or something like this. That's no person will help you to to build storage from scratch.

Now you lose credibility, with the developer. And so, I try not to do that. And when, people ask me, like, what do you think of, database company x y z? Well, I tell what I think. You know, investors ask for all the time, for example.

I tell what I think, and I try to give credit what credit is due. Mhmm. And then I'm also careful when I say bad things, and I say when I say things like, well, I see this as a deficiency there, which may or may not cost them things. Right? Because I'd like they they they probably recognize that too.

But since I've been in the space, I kinda recognize that is might be an issue. But that's for you to figure out. Like, this is not my job to, to go and dig and and and basically match my observation with what happenings, what's happening in the field and then the reality and deal by deal or something like this. It's like, I I wanna learn from what works mostly, and I want to learn work from, a job well done. So, you know, some companies do a very good job in our space.

You know, soft Snowflake is doing a great job in analytics. Mongo is doing a great job in in operational workloads. SuperBase is doing a great job within the developer ecosystem. We wanna learn from all of this.

Like, that's the flaw. But if in terms of the strength of the engine itself, I think they have a very solid engine. And so if you bring people from that that culture, you adds to your to your toolbox inside your own company. And, of course, you hire people from outside of database industry too because they also come with with, ideas and, and tricks.

So you you need to think about, first of all, compatibility with with the founder. Then you look at what the organization needs. The organization, at at at the very at the various time of their journey, can struggle with a myriad of things. Last year, around the same time when we brought James, our VP of engineering, we struggled with reliability. And, reliability is one of those things that, you know, like, I can tell you the reliability is awesome now, and tomorrow we'll we'll have something.

And so, like, I will look like an idiot. But what I can say is that in the period of time between, you know, February, February this year, if you look at the incidents, the time at which incidents, were on and the, the severity of them and kinda map them into some sort of number function and plot them is like a steep drop.

But in addition to that, James brought in processes, people, attention, you know, team management, and, like, this reliability is not one thing. It's not one bug. It's, it it's like hairy, sophisticated thing where you can improve, improve, improve, and eventually, you get the reward of being a very, very stable system. And then, so compatibility with you, like, this with the kind of style leader you, you want he wants superpowers for, for the, for the leader. I think one of the superpowers for James is very emphatic, and people want to work with him and for him.

And it's kinda a big deal. Mhmm. Right? But there are many many ways to to drive the organization forward. And, you know, one way is is, you know, this is the hill.

We're gonna take the hill. We're gonna lose some people on the roads to the hill. And those people who we're gonna lose, that's their fault. But we're gonna take that hill, and, you know, we're just marching there. And then there's a a way to run the organization where you understand the people and you working towards the hill together in a smart way.

And you don't wanna lose people on the way to that hill. Like, that's unacceptable. Underperform is different story. Right? You know, they're they're not everybody can be at a start up.

That is, of course, you know, every start up must, part ways with people who are not, kinda, like, pulling weight in the same way as the top performers. Otherwise, the classic, you know, a players, b players, c players comes in. But you certainly don't wanna be an asshole about it, and you don't want to, burn the organization and throw people under, you know, machine guns because, like, there's no victory there. And so that that that kind of leadership, I think, works better with me. It's kinda nerdy.

Right? So thinking about the organization as an engineering systems, recognizing that it's driven by people and, giving them credits, making sure that you listen to the people, hear their concerns, explain them what you're doing, like, taking that time. And in in return, the organization has high overall, and at the end of the day, moving faster.

So the business is not working. It needs to be a different business. What the engineering team is not working, and it needs to be a new new engineering team. And the new engineering team, either because the the overall bar is low or because the composition of the team must be different based on the change in the technology landscape. Right?

We had on prem people. Now we have cloud people. We had website people. Now we have mobile people. And, you know, once the market to either, adapts to the market or you don't. Right? And sometimes the market changes so fast, you need to change the team. Mhmm. Different kind of leader needs to be here in in this type situation. Or it could be, you know, we're doing some deep research, and therefore, you need to bring a leader that will bring on the researchers.

So now that becomes the higher a bit. So I I think it's very situational. Mhmm. What I can recommend is to always, start with what is the situation. Tell yourself, like, you know, like, the narrative thing. Like, explain it to yourself. What are you doing? Then explain it to your board. Then explain it to a whole explain it to your friend. Then explain it to your board.

Explain it to other execs. And then act and then try to hire the best recruiting firm you can possibly have. I'll do a plug here. I like working with Alex Zagrebowski. He used to be at Artisanal. Now he started his own firm, Fusion Talent. I think that's what it's called.

You know, there's opportunity cost there, and you're doing diligence on them, and they're doing diligence on you. And actually having an intermediary helps them, helps them to, to make the process a little bit more formal. Now it is expensive, but in the grand scheme of things, it's not. And you need the cash constraint or you're you're not cash constraint. So it's, you know, it's a little bit of an art.

Right? But the other thing and sometimes we see tell you that it's like, we're gonna help you with recruiting. Not really. I mean, they will all tell you that, and they will have agreed teams. And then agreed teams will will will do work for you. It's just not gonna be as high quality work most likely

With the support of everybody involved. So you go to the board. I'm hiring this search firm. Here's the man. And that first was this specific person, in the search firm, man or woman. The board some of the people in the board wants to raise their hands. Like, I don't know him. I would just wanna just for for for reliability, let me talk to this person. The board signs off. You hire this person.

Then it's intense few weeks to few months of of searching for the specific, yeah, executive, understanding of the what their process is. But everything is kinda formalized. And and and and also, like, those people, if you do it the first time, those people will help you to to interview as well. Right? You're not gonna just chat.

Right? You can have a plan, and they will help you devise that plan. So I think all of that is is like, making those those things a little bit more formal actually helps.

I don't know. I thought that was a very good answer. I haven't I haven't an observation. I don't know if it's, I'm like squinting at the world and what's working, what's not in the modern world. And you we we talked about how Cursor, you know, ramped the revenue very, very quickly.

I think you can in dev tools, you can sell a dev tool. And a dev tool would be, like, oh, Visual Studio Code or something like this or or, GitHub. And then you benefit if, like, the whole planet is using them. You can also sell storage and compute. So you're in the DevOps budget. So you're now selling databases. You're selling compute. You're selling Kubernetes. You're selling Kafka. You're selling MongoDB.

Or recently, you can sell tokens. And if you sell tokens, you sell in Inference or you sell in Cursor. Right? Because you're, you know, you're running those tokens. You're selling, you know, Perplexity, you're selling tokens.

ChachiPT, you're selling tokens. Vzero is selling tokens, right? And I think tokens scale a lot faster than the other two. So I was of an opinion that unless you're like GitHub, and you like scorcher a swan and everything, it's like, it's actually really hard to make money in dev tools. Mhmm.

Really freaking hard because developers all like to pay for tools. They like to use open source tools. So there's constant commoditization. So what you end up monetizing is either seeds, but then there's always a next open source tool that's free for you and developers build other things for developers. Then he monetizes a service, but that service usually hosts something.

And therefore, it's a proxy on storage and like, you see, there's seats to storage and compute. Right? You know, on the markup, added value on top of storage and compute. And then, that big business could be built. Right? Snowflake's simulated run rate, so on storage and compute. And then tokens are, like, next level. So my advice is, like, understand that framework, to whatever you're building and apply it. And just, like, whatever you choose, be aware of what's going on. Mhmm. So