How Two French Engineers In New York Built The Company That Monitors The Entire Cloud

My co-founder stood in front of the whole engineering team a couple of months ago saying, "Hey, in 2 quarters we're not writing any code anymore." And we saw, like, literally absolutely excellent developers. Like, not beginners, but people who rebuild whole systems in a couple of days alone, that were absolutely amazing.

Things that would have been possible but, you know, done by a team of 6 over 6 months.

Uh, welcome, Olivier.

Thank you.

By the way, it looks like an IQ test of sorts here, like there's like—

Like which one, uh—

—6 microphones of different colors on it, you know, so.

We always like tests.

Yeah.

Uh, as we were preparing that session, Olivier reminded me about something here. I'm sure you wanted to say that,right?

Yes. So, uh, we actually did not get into Y Combinator. Datadog did not get into Y Combinator.

I checked. They applied.

Yes. We applied. We got an interview. I've got a rejection email from Paul Graham still, you know.

You put that on the wall, hand it, and like, uh—

Yeah, yeah, yeah.

So it's thanks to us that you became such a big—

Exactly, you know. Like, we had to prove you wrong, so.

Okay, so if any of you get ever rejected, here's the outcome.

Yes, yes. And then, and then, and then, here's the thing. So then I received, like, when I agreed to do this, you know, I received an invite to a YC dinner in Paris. And then I said, "Oops, sorry, we are not into YC, so you're not invited anymore."

So I got rejected twice by YC.

Well, the second time, hopefully, it didn't hurt as much. Uh, well, but you're here. And, uh, we'll make up for that, hopefully.

Yes.

Allright. Um, you and your co-founder, Alexis, you met—so you were in Paris,right, at Centrale?

Yeah.

Uh, that's how you met. And then you ended up working in New York. So, uh, why did you leave? Like, was that a deliberate choice back then?

So, no, it was not a deliberate choice. So we—so I graduated in '99. I had an end-of-study internship, and I did it at IBM Research in upstate New York, which at the time was a, you know, fairly legit place for research.

Not anymore?

Uh, I mean, I mean, now, you know, look, now there's, there's, uh, you know, DeepMind and OpenAI and—

More sexy places to do research.

Yes, yes.

Were you doing research there?

I was doing research on internet protocols, you know, so at the time sending emails. So, you know, not quite AI, but, you know, it's, uh, email works. Um, so, but anyway, Alexis, my co-founder, was a couple of years ahead of me in school.

He had already been to that internship and was still there. He had started working at IBM after that. And so we met there, uh, for real. Like, we met a little bit in school, in weird ways. I can come back to that.

And then we, we, uh, we spent some time at IBM. I thought I would stay in New York 6 months, and I'm still there. You know, it's like 26, 27 years later. So, uh, so that part was a, uh, was a happy accident, I would say.

When, um, halfway through the experience at IBM, um, it was the dot-com boom. So startups everywhere in New York, super exciting, lots going on. I started working for startups. We actually started working together with Alexis.

How many here know about the dot-com boom? Well, they may have heard a little bit about that.

And then after that, it was a dot-com bust. So I got a little bit of the dot-com boom and the entirety of the dot-com bust. So a lot of lessons learned about what to do and what not to do in a startup for that reason.

But you were not building a startup back then.

I was, I was working in startups, not mine. But I was working in startups. One of them was actually should have been pretty amazing, but made a lot of mistakes, you know, in terms of, you know, spending too much on the wrong things, not shipping product when they should have shipped product.

Like, a number of things to remember.

Okay. Is that where you got the idea for Datadog?

So later on. So after that, the things got worse in New York. There was the September 11th and everything else. So it was pretty dark. But I decided to stay there for personal reasons. And I started working again with Alexis in an educational software startup.

And at the time, it was after the bust, so nobody thought technology would be worth anything ever again. And so we started working there. That company was pretty successful. We built the tech teams there. So I was building the dev team, Alexis was building the ops team, like tech ops team.

It was pretty successful. It grew. It was SaaS. And that's what led us to Datadog.

How big was the company?

When we left, it was like, I don't know, 800 people, something like that.

Okay, so it's pretty big. So you had, like, a lot of DevOps issues, operation issues.

Yes, yes. And also we, I mean, we lived through, you know, developers hate operations, operations hate developers, people point fingers at each other all the time, you know. And it was weird because Alexis and I were very, very good friends.

The team reported to us. We tried not to hire any jerks. And that was the—

So there were already observability companies out there back then,right?

It was not called observability. It was called monitoring.

Okay.

And it was very job-specific and very reactive. So, you know, for network monitoring, you'd have a product. For, you know, you'd have something that the ops team would use, but the developers would never touch it, you know. Developers would be blind to anything that happened in production, pretty much.

And you started targeting that audience,right? You were mostly bottom-up, at least at first.

Yes, bottom-up. And we started in 2010. And what we didn't completely realize at the time was that the premise for the company, which was, let's bring Dev and Ops together into one platform, was actually completely central to the explosion of the cloud.

And we also completely underestimated the explosion of the cloud. We thought, oh, that's interesting. There's this Amazon thing. But, you know, when you talk to real companies, they all say it's a toy and they're never going to use it.

Turns out they all did.

But that was your bet back then? Because you were cloud first.

Yeah. So the bet was Dev and Ops together. And then the first infrastructure we started supporting for that was cloud infrastructure because we talked to smaller companies. And the smaller companies or more modern companies were built on the cloud.

And that's what got us to that.

That became the tailwind of everything. Like, the cloud market exploded.

Cloud market exploded. Dev and Ops were smooshed together. Our product also was, uh, like, we called it infrastructure monitoring initially. We didn't call it observability. The word didn't exist. But it didn't look at all like infrastructure monitoring that came before.

And one of the reasons for us to there was we didn't do that initially. Initially, we had, we called it something different. Like, it was a data platform for Dev and Ops to collaborate and things like that, you know.

And, um, because we thought monitoring was a thing from the '90s and it was boring and it was old. And it turns out, uh, people love the idea of a new thing, but they have no idea what they should pay for it.

And their boss has no idea what they should pay for it. Whereas you say, oh, it's infrastructure monitoring. And then everybody and their boss understands, yes, actually, it's a product. We need that. It's a category. It's real. Let's get it.

How fast did you get there? Because, like, we kind of, like, are nearly, I mean, owning that category, I would even say today. Was that obvious really fast that it would work? Or did you add, like, to go through very difficult times?

It was not obvious at all. I mean, for one, we didn't get into Y Combinator.

Yeah, that's probably—

Yeah.

And did you have any other rejections or any blow times after?

No, no. The first round, we raised an angel round, and it was super painful. We couldn't get any of the smart investors to pay attention. Like, if you were a big VC, we were in New York. You had VCs in New York, VCs in the Bay Area.

Not much going on in Europe. And we'd go to New York, and the VCs would be interested. And then they realized, like, they looked at the CEO competitor slide, and they realized they don't really understand any of these companies.

And they think, okay, it's not for me. I don't know this market. And then we'd go pitch the VCs in the Bay Area. At the time, they only invested in the Bay Area. Like, very, very, you know, focused.

And to them, it was a form of mental impairment not to try to start an infrastructure company, but not in the Bay Area. And so I could even hear them speak slower and louder, you know, like.

Just to make sure you understood them.

Yes, yes. So I think it was.

You couldn't frame that on the wall.

I couldn't frame that. Yeah, you know.

But I guess it's kind of like, uh, you want to prove them wrong.

Yes.

And in a way, you can turn that into crazy motivation to prove them wrong.

Oh, exactly. And I think we, in the way we built the company afterwards, there's a number of things we did that were on purpose different from what we saw Bay Area companies do. Because we thought, hey, we're going to, you know, we're going to do differently.

We're going to be better. We're going to do, uh, so, for example, you know, we never wrote down the culture. Like, you know, at the time, I think a bit less today, but at the time, every company had, you know, five values or eight principles or, you know, nine concepts or, you know, there's a.

So what did you do instead?

So what we didn't, we said, what we said is, hey, look, if you need, if you need, don't be evil to be written on the wall. You probably shouldn't work here. You know, like, uh, and what we said instead is culture flows from the top.

So we exemplify the culture. We hire, we promote, we fire based on the culture. And today, that still works.

So your culture is you.

Well, I mean, the culture is what we do. That's not us. That's not me.

Yeah, yeah. I mean, like, you said it needs to come from the top, but that's cool. I mean, I'm not, uh.

And look, we're fairly, uh, we're pragmatic too,right? Every year, we ask ourselves, hey, is it the year we need to write it down? Because, look, we need to hire 3,000 people. We need to train those 3,000 people.

How many people today?

We are 8,000. 8,000. We need to train those people. What do we tell them? And so far, I think it still works.

Awesome. That actually makes me think about, like, a conversation I had with Pierre Betoin that you know very well, because I've worked with him too. And because of this conversation, I asked him, hey, like, what should I ask Olivier?

And actually, one of the things he was very impressed is how heavily involved you are still on, like, some very small product decisions. And you don't spend your time in that strategic kind of whatever, slide decks and stuff.

How did you keep that? How did you keep that going?

Yeah. So first of all, I hate the S word, you know. So when people use strategic in a sentence, I usually, I replace that mentally by bullshit, you know. So that's, uh, I mean, in some cases, it's meaningful,right?

Not always.

And like everyone in the company knows that, I guess.

Yes. Yes. Well, I mean, not.

That's the culture thing.

Well, people make the mistake once, usually, but, you know.

The, so first of all, I think product, when you, for what we do, like, we're, like, infra and product company, it's very important to be in touch with the product. It's a big part of my job. Of course, we have plenty of people who handle the product.

Like, we have people who spend time with customers. We have people who focus on all the small things. But I think it's very important for me and a lot of the management team to be aware of everything that's happening in product, read all of the new developments.

And as you get further up in the org ending with me, you're in a position where you can really edit. Like, you can say, hey, I think this is too much, or I think it's not enough, or let's stop doing that, or I don't understand that part.

If I don't understand it, how are the customers going to understand it? Like, this is this level of a.

Do you feel you could become a bottleneck at some point?

So it's not blocking, you know. So I don't try to, I'm not a necessary approval step, you know, and everything. But, and part of the game is to understand where to direct the attention and where not. Another thing that I think I encourage everyone to do as soon as you start scaling a little bit in your company is to keep in touch with what's actually happening on the ground.

And by that, I mean the actual support requests, the actual sales transcripts from the sales conversations, the actual customer feedback, the actual employee comments that you get in the employee surveys and things like that. Of course, you can't read everything.

I mean, except for, I still read all of the employee survey comments, but for the rest, I don't read everything. You sample. But when you do that, you get a good sense for what the fabric of the universe actually is in the company.

And you don't just get the beautiful summaries that, you know, find their way up to you, you know, where everything's beautiful. All these new products is great. And they see, but I read two complaints about it from customers.

Like, they say it doesn't work. Like, you know, or it's too expensive. What do you mean?

I'll fight that up. Like, kind of like, you discover something, reading some survey or some feedback from customers. And then you go to the team. Hey, come on. Like, it's not working. Look. And before, they were kind of like trying to message all the message.

Yeah. Well, what I do is I just pick up those random emails or, you know, Slacks or whatever. And I just reply with a question, you know. I say, I don't understand. What's going on? And then.

That's passive-aggressive.

And then, you know, what happens is that two things happen. First one is the person who's receiving that is wondering, oh, why is he looking at that? And then after that, the positive effect is that the whole management chain, instead of looking up and trying to figure out how they're going to manage up, they start looking down and starting to figure out, okay, so what's actually happening?

I better know because if I get asked, I need to understand, you know.

So going back to the culture, showing by example in a way.

Yeah.

You need to demonstrate to live the culture yourself so that your, everyone who reports to you is going to do the same thing.

Yeah. Your job is not to understand what I want as a boss. Your job is to understand what's happening within your span of control. And you need to very, very precisely what's working and not working.

I mean, the company has become huge now and is multi-product. How many products have you? Do you have?

We have around 20, 25.

20, 25 product. I mean, how do you decide how you expand the portfolio? How do you decide the next thing to build?

In a way, so we're kind of platforming as a product. So we see a lot of usage from our customers around our product. So they sort of build things around our product. Like, they have their own kind of extensions, their own workflows that they use us in, you know.

So it gives us an idea of what kind of problems they have and where they think we belong. So we do a lot of that. Like, we build a lot of the things that we see them do. And sometimes we also have some, you know, more, I can say, I'll use the S word, some more strategic, you know, projects.

Oh, bullshit.

Yes. Where we, you know, we think, okay, we're not, nobody's asking us to do that, but it looks like the world is kind of going that way. And so we should be doing something there. When you do that, those tend to have potentially a bigger impact because this can be brand new things.

On the other hand, you tend to be wrong often because you just made that up. And I would say one thing that's a bit changing now is that with AI in particular, things are changing so fast that you just can't wait to see the demand materialize from the customer base.

So you need to get ahead of it, which I think us and many others, I think, are going to get wrong more often. And we need to be at ease with that.

How does the fact that you're public now impact how you make decisions then? Like, I mean, I guess, like, yeah, like the price of the share is an important factor for you now, or?

The price of the share is not, like, you know, we, so we've been public since 2019. Our lockup expired the day of the COVID lockdowns. So I don't know if you remember, there was a big market crash that day.

It was also a lockup expiring. One of our biggest investors got scared and he dumped everything that day. So I think, I don't know, we were down like 65% or something. Like, it was.

What was the enterprise?

Sorry?

Like, I know that what your value today, but I don't know how that was when you.

Well, at that time, I think we went down from, you know, being worth, from being worth 10 billion to being worth, you know, I don't know, four or something like that.

Okay. Like a huge drop.

Yes. Especially when, you know, when everybody's seen it at that valuation for and they were, and people are waiting to sell after their, when the lockdown expires, you know. Like, it's real money that people can use to buy real apartments.

Like, you know, it's tangible,right? And so that was pretty bad. And then we saw, and then that crash lasted two weeks. Then the market went up like crazy again. And then it kept going up like crazy in 2021, 2022.

And then it started crashing again. And then it started going up again. So, you know, we've seen, like, we've seen the stock go up and down all the time. And I think that we've gotten used to it.

Did you have to change how you lead the company because you were public?

So not really, no. I think it's a, like, you'll see, like, as you run a company, like a private company, you spend a lot of time talking to investors because you probably will need to raise funding again. And your job when you're the CEO of a company is to make sure you don't run out of money.

And as part of that, every single investor out there should believe that you're the future of sliced bread. And, you know, for that, you need to spend time and make sure that they know you and they understand what you do and they think you're great.

So you spend a lot of time, at least I did, spend a lot of time with investors. When you're public, it's very choreographed. Like, there's a couple of.

Every quarter, you have that call.

Yes. Yeah, you have, you spend a couple of days every quarter on that. You spend maybe a week preparing, you know, to make sure you have theright message.

But then you don't worry about the share price, like how the market is going to react to whatever you announce?

You do. I mean, you, look, the main issue it creates is if you do something that's unpopular with the market, you're going, maybe you're going to lose, you know, 20, 30, 40%. Does it impact the company in the short term?

No, but it creates a comp issue because people are paid in RSUs. So the main things you have to worry about is, okay, so what does it change to comp and how can we manage that?

So the actual risk is not the survival of the company. It's kind of like losing key performers.

Yeah.

Because their RSUs are worth less.

Yes, yes. And that becomes comp management. That becomes, okay, so how do we do? Are there, and also, keep in mind, like, people join at different points in time. So they have, their comp is actually at different levels, you know.

So you have to be more, to look into more details of what's going on.

How about

the AI kind of like revolution in a way? I guess the market also was thinking about that, looking at your company, because obviously, early on, it was not obvious you would be a winner or a loser. How did that impact you?

I mean, we'll see. It changes every couple of months now.

Okay. But like, it looks like you're on a good trend now.

Yeah, now we're on the winner side, which is good. We should.

But you didn't start there.

Yeah. Well, we started in the, uh, don't know side from the investors. And I mean, look, the reality of it is there's a ton of disruption. And for investors in general, but I would say even more so public investors, it's really, really hard to understand, you know, what things look like, you know, three, four, five years from now.

And, you know, something, if you, when you start a company, you're going to pitch VC investors. You're going to have maybe two or three of these VC investors that are on your board or spend time with your company.

And, you know, you might go and, you know, have coffee with your friends and say, oh, my VCs don't understand me. But you'll be, the VCs you have as a private company, they spend so much more time on your company.

They're so much more concentrated than the public investors. Like, the public investors have, they're looking at, you know, 20, 30 different names. They come in and out of these names, you know, a couple of times a year, as opposed to a VC that's going to be in like seven, eight companies and, you know, stick around for five to ten years.

So that's a very, very, very different deal. So public investors, they have less, it's for them, it's harder to reason about things. So what you need to do is you need to make sure that you tell theright story to beneficiary going.

And then you produce numbers that actually agree with that story. And if you do that consistently enough throughout the years, that's our job as a public company, then you do fine with them.

And what did that mean for you on the, like, on the company side, not the public investor side? Because I guess you had like to rethink the product. You had to make new bets. In a way, you couldn't stay defensive.

You had to become aggressive on the attack,right?

Yeah. So, I mean, there's a few things. One is working in a bit of a different way. You know, so I mentioned earlier, AI is changing fast. So you need to take more shots and be wrong more often.

And so we pivoted a little bit to that. You know, one example being, look at the way development has changed. You know, it's changed three times. Like, the way you write code or don't write code has changed three times in the last year.

It's impossible to predict that a year ahead of time. So you have to be very reactive and you have to try a lot of things all the time.

So what are your latest bets?

Well, so we're getting earlier into the, so a few things. So one is we're making, we're investing a lot so that the product can be used by agents. Because we see that, like, in our customers, we see a lot of that.

Another one is that we're building a lot of the smarts directly into the product. You know, if you use Claude Code, for example, to work on your ops issues, it's a little bit like, you know, trying to send a postcard by buying a business class ticket for it, you know.

There's a lot you can do and you should be able to do that would be directly fused on the data plane for us inside your observability as being outside of it. So we're building a lot of that so that the system can do everything on its own.

And I guess AI companies who are going super fast also need observability.

Yes.

Is that a big piece of your market now?

Yes. And that's also, you know, in terms of how we find out what to do next and where the world's going. So we have the top 10 AI companies in the world as customers. We have all of the startups that are doing, or not all, but, you know, a good fraction of the startups that are building on AI.

And we also have the big enterprises that make up the normal world, so to speak, like, you know, the big banks, the train companies and whatnot. And so we get a pretty good panel of what people do. We see the AI labs, like, you know, the top, like, two or three AI labs in the world use our product in weird ways because they have infinite compute and infinite, you know, inference and the models that nobody has yet.

And so it, you know.

Does that give you a glimpse of what's coming up after?

Yes, but it's not necessarily the most representative because, again, they have all that. So we see, we get a sense of that, but, you know, the rest of the world doesn't have infinite compute. So they might do things in a way that the rest of the world is not going to do.

Like, you know, most companies are not going to be very happy to, you know, incinerate, you know, a billion in compute to do a lot of internal things in a way as they did what they can do.

How about you? Are you, like, how do you use AI internally? Are you token maxing?

We're not token maxing. But look, my co-founder stood in front of the whole engineering team a couple of months ago saying, "Hey, in two quarters, we're not writing any code anymore," you know. So.

Oh. So directive, no code anymore.

Yes. I mean, we, look, the reality of it is we will write code and we will read code,right? But the point here is it's an inversion. Like, you used to mostly write, sometimes automate, and now you're going to mostly automate, sometimes write.

Does that change how you lead the teams, how you hire them? Like, I don't know, any surprising things you've seen?

It changes everything. But the tricky part is we don't know what the destination looks like yet. So we know, so we know, for example, that smaller teams can do a lot more. And, you know, which is part of the, "Hey, you can do more and be wrong more often," but that's fine because you also, you know, instead of having a team of eight to explore something, you can do it with a team of two or three, you know.

So that's great. On the other hand, nobody knows what the process looks like a year from now. Nobody knows how many PMs, how many designers, how many security people, how many engineers you need to build a thing. And so.

Do you get some benefits on the way? It feels like you probably have already some big ROI on some of these changes.

Oh, yeah. I mean, look, we've seen, like, many companies, like, you know, we had this moment in December. I think in part because the models got really good, the threshold was crossed, but also in part because people had a bit more time on their hand, you know, it's the holiday season.

"Hey, maybe I'll do a bit of a hack project on my own." And we saw, like, literally, like, I mean, absolutely excellent developers, like, not beginners, but people rebuild whole systems in a couple of days alone that were absolutely amazing.

Things that would have been possible, but, you know, done by a team of six over six months. And so that convinced us that, yes, you know, there's enough. We have to pivot the organization. It's going to be painful.

It is painful, but we're doing it.

And speaking of your co-founder Alexei, I mean, you've been working together for like 20 years or something. Like, how can you still, like, how did you make that last?

Yeah. I mean, look, we're better than old couple now, I guess, you know. I think there's a couple of things. So one is we worked together without starting a company together for almost 10 years. And so we already had a good, a lot of experience, like, you know, with what's theright way of working together, theright boundary, you know, like, you have to test those things a little bit, you know.

And then as we, when we started the company, like, we also tried to give ourselves a lot of space to talk. So.

Any hack? Any tips?

I mean, you just have, we just have a standing lunch, you know, and there's no.

Like once a week or something?

Yeah. I mean, it's mostly once every two weeks now, but yeah, it's a, and the point here is you don't have a specific agenda. You just talk. And it's important to get what's in.

Got to agree more.

Yeah. And also, like, you know, when you run a company, like, there's not a lot of people you can pitch to or, you know, like, you, so a co-founder is a great person for that.

So most of the people here are not necessarily Paris, but at least Europe. If they start a company, should they start it here or should they go to New York or ASF?

I think it's fine to start here. I think it was not true when I started Datadog. I think you started a little bit, a little bit later.

I started here in 2012.

Yes.

But then we moved to ASF. So it's kind of like, but we kept the team here. So.

And I think it, so when I started, it was, it was not, it was hard to get funded. It was hard to get everything. I think now you can get funded pretty much anywhere in Europe. On the other hand, I think it's absolutely critical to go after the US market as soon as you can.

So as soon as you have a product with some form of market fit, go after the US market. And.

But don't you need to be in the US when you do that?

Yes, but I think the whole company doesn't need to be. I think the, you know, typically if there are two founders, one of the founders moves to the US and you cannot start in the US without having a founder there.

And you should absolutely go to the US, but I don't think you need to move the whole company.

And you did the reverse move,right? You created a team in Paris once you were bigger.

Yes.

Was that mostly talent acquisition?

It was mostly talent. It was most, I mean, in fact, we had a number of people in the US on visas who either couldn't renew their visa or, you know, wanted to go back to France for family reasons.

And that started our office in Paris.

You had a lot of French employees in the US.

Yes. We hired, like, we came there on an internship and that's the thing. We hired a lot of people that way.

Excellent. Allright. Maybe to conclude, is there one piece of advice you would give yourself, like, 15 years ago?

So here's the thing. All of the mistakes and all the hardships and everything, like, this sort of kind of made us successful. Like, you know, I said, not being easy to fundraise initially. I think that that sort of created our values.

From YC is the reason for your success,right?

Yeah. I mean, maybe we should, we could have gotten to YC, you know, and everything went better.

Hopefully, we would not do that mistake again.

Yes. But look, there's a, the one thing, the one thing, the one lesson I keep learning all the time, so is that you should always move faster, especially when it comes to hiring and firing. I think these are the two things that.

Oh, firing I hear all the time.

Yes.

Hiring, I don't hear it that all the time. Why hiring?

Well, I think the two go together. I think it's, you can take more risk and you can hire faster instead of, so for example, one thing I hear often is, oh, we're looking for a head of sales. Because most founders are not, you know, salespeople.

And so, oh, I'm looking for a head of sales. But I'm not sure I found theright profile yet. And my answer to that is, hey, find someone who looks good, hire themright away. And if it doesn't work out, you know, you fire them.

But it's much better to do that than to wait. Like, you'll run more cycles, you'll have more chances. I think it's too hard to know anyway before you've tried. So.

So I guess you've done that yourself now.

Yes. But, you know, I made the mistake initially. I was paralyzed trying to make the perfect hire and, you know.

And then you would have kept them for too long.

And then, and well, that's the other part, you know, it's usually, like, the first time you fire someone in your company, you feel horrible and you think everybody's going to hate you for it. And mostly what you see the rest of the company is like, I mean, what took you so long?

Why did you wait?

Awesome. Thank you, Olivier, for this interview with us.