Today, I welcome Olivier Pomel, the co-founder and CEO of Datadog. He created Datadog with Alexi back in 2010 just to help engineers figure out what was happening in their cloud. Now, he's built that company into the defining company of the cloud era. Welcome, Olivier.
Great to have you. So, we were chatting before about your first YC experience actually. It was not today. It was back in 2010. Tell me, what happened then?
Yes, so we actually applied to YC in 2010. We got to the interview and we got rejected.
So, I still have, you know, the email from PG at the time. And you know, it turns out so, he was saying at the time that we were building a platform, but the platform was only as successful as its first product.
And you know, a little revenge on life has been that we did make the platform successful in the end.
[laughter]
Do you think it actually helped you in some way to show him wrong?
I think so, first of all, I think many people who start companies go through a lot of rejection initially. I think that's normal. I think it's good, you know, to give you reasons to prove people wrong and to, you know, put a chip on your shoulder. And also, you know, I think the fear to fail that you feel initially, I think at least in our case, it helped really build up the culture of the company, you know, making sure we were so obsessed about building something that was valuable. We were so obsessed about building a good business, a business that could be self-sustainable in case we couldn't raise money. And so, all of that, I think to this day remains the culture of the company and we owe that to the difficulties early on.
I'm so happy to learn that we actually helped you in the end. [laughter] That's awesome. Let's go back to the early days because you were born in France and then you moved to the US after. Was that after college?
Yes, so I had an internship at the end of college in New York in research at IBM at the time. And I thought I would stay for 6 months and I ended up and I'm still there.
There? How long have you been living there?
So it's been like 26 years now. It's been a while.
So you pretty much moved there just at the time of the bubble, the dot com.
Yes. So I lived through the end of the dot com boom and I actually worked in a startup during the dot com boom.
Which was it an experience?
Yeah, it was super, it was weird, you know, because you could tell that things were happening that, you know, didn't make complete business sense. You know, so for example, the CFO of one of our startups every day went to the most expensive pastry shop in town and bought pastries for the engineers.
I think that was a red flag maybe thinking about it.
I guess you don't do that today.
That's not the thing you learned from them.
Yes, but so I think the world was full of promise then, then of course the dot com boom turned into a dot com crash. We've been through all of that. There were few years after that where it didn't look like technology would be worth anything ever again. So I stayed in New York and my co-founder for Datadog, Alexi, also stayed in New York.
But I guess you worked elsewhere.
Yes, we worked elsewhere. Like we started working for an educational software company. It was a tiny company that we, it was not our company but it was less than 10 people when we joined. And I ran the development team there and Alexi ran the tech ops team.
And so did you meet Alexi there or?
No, we knew him from before. Oh, we met actually briefly in school in France.
And the funny story here is I was part of the team that ran the campus network.
And Alexi was caught hacking on the campus network. And so he was court-martialed by the student body and he was sentenced to disconnection and I carried out the sentence.
That's a great way to get to know someone.
So, and we didn't really interact all that much, you know, I thought he was this black hat guy, you know, and then we ended up finding each other again at that IBM internship, basically. So, he was there, too. And then you worked again together at the next company you joined. And then we worked together in a total of four or five companies, you know, so you never worked in any place without him.
Yeah, basically, I mean, we always bring the other, like, you know, maybe one of us would go somewhere and then two or three months later we'd bring the other, you know.
So, that's a strong relationship.
And then you co-founded Datadog together.
And you're still working together there.
Any secret of keeping that kind of marriage lasting like that?
Yeah, well, I mean, look, it takes effort. For one thing, before we started a company because starting a company is pretty hard. Before we started a company, we had known each other and worked together for more than 10 years. And so, we had tested a lot of the, you know, different ways the relationship can go, you know, and we're also good friends. So, and that helped quite a bit. After that, I think as we've grown Datadog, like we've made sure we kept a lot of space for talking to each other.
You know, so we have these regular lunches and things like that where the whole point is just to talk. And I think so.
No agenda, just open-ended conversation.
Yes. And there's one extra thing which is very important, which is anytime there's a big decision to make. So, for example, sell or not sell. Like, you know, as you grow a company, you're going to have folks, you know, every now and then that will try and buy you out. And these are big decisions.
And did you decide after getting a real offer or did you decide to not even open the conversation?
We had those conversations where we, when we had offers or when folks were pretty serious with us. And it's very important in those cases to make sure that you're on the same page. You know, because that's what can cause companies to blow up.
[clears throat and cough]
One of the founders kind of wanted to sell and the other one didn't. But, you know, one maybe didn't really tell the other or, you know, they told them what they thought they wanted to hear, but then, you know, it turns into a problem later on. Like so, it's very important to spend a lot of time talking at those points to make sure you really understand what each other want, you know.
You bet. It's very easy in hindsight now to say, yes, of course that was the right decision not to sell. But I guess at that time it's a lot of money you leave on the table.
Yes, and there's no guarantee that your company is going to be successful.
Yes, and you know, when I started Datadog, I remember seeing like companies that didn't sell for, you know, refused an offer for 200 million. And I thought, well, what kind of, you know, who refuses 200 million? And then, you know, we refused that and then we refused 10 times that and then, you know, refused more than that even. I think the question is always.
When you became public, what was the first cap of the money?
So, we went public at sort of the, the go public price was around 7 and a half billion. And we very quickly traded at 10, 12 up from that.
And then the roller coaster.
And then and then roller coaster.
To kind of like a much more today.
Yes, today we're around 80. But at the time it was around that. And we, you know, we've refused large offers along the way and again the thinking there is always what's the upside? How do we think about the business? Are we done with what we wanted to do? Is there, you know, 5x, 10x more we can do? Can we see? Do we believe in it? Do we want to do it? You know, is it exciting enough for us? And that's what I usually do.
Yes, but that's the consideration.
But you know, it's these are very full-time jobs. So, you have to be really into what you're doing and you also have to believe into the outcomes that come after that.
You bet. But before being that successful, I mean, the early days were not that easy, right? I know when you started it was nearly pre-cloud, the very early days of the cloud, right?
How did you approach that?
Yeah, so it was very hard from the understanding market perspective. So, Alex and I didn't come from systems management. So, most of the people who started companies in our space used to work for a vendor that made software, you know, that sold software for that. We didn't have that. We also didn't come from a hyper scaler. So, we didn't come from a Google or, you know, one of the places that was operating at super large scale and maybe, you know, folks had some glimpse of the technology that can run that or the technology is coming for the rest of the world. So, because of that, it was very, very difficult for us to convince investors initially. So, we mentioned YC, but, you know, pretty much every single other VC also passed on us at the time. So, it was a very humbling experience.
A lot of people will regret today, I guess.
Well, I mean, but some of them invested later, right? You know, so that's the thing with investors. There's always a next round. You know, maybe not for YC, but, you know, for everybody else there's always a next round. And the other part of it though is because we didn't come from the space, we sort of saw things a little bit differently. Our starting point was not, 'Hey, let's take that existing product category and make it better.' Our starting point was, 'Hey, there's this huge problem. Like Dev and Ops don't talk to each other and they fight all the time and can we bring them under one roof and into one platform?'
So, we saw that in a way that I don't think the rest was seeing it because it was on the user side. We, what we didn't really fully understand was the importance of the cloud and the fact that what we were doing was right at the center of it. You know, so, disintermediating, bringing Dev and Ops together, that actually that's a big part of the cloud adoption. And of course, we at the time in 2010, you know, Amazon was a, or AWS was the only game in town and it was a toy. And companies, real companies were telling you that they were never going to use that. Nobody was ever going to move to the cloud, right? Yes, exactly. And of course it happens. Yeah, of course it happens. It was extremely broad, extremely deep. And so that was a big part of our success.
It took a, when I compare that to how fast people are adopting AI, for example, I mean cloud it took 10 years to get to full adoption.
And in a way you were lucky, I mean lucky. You are smart, but like you cannot like end up at the center of that revolution.
And that's probably gave you a unique advantage compared to all the incumbents who didn't see it coming.
Yes, definitely, definitely. I think and look anytime you build a successful company, so yes, there's a ton of hard work, there's probably talent too, but there's a lot of luck.
You have to create your luck. You need to try as many things as you can so that you expand that surface.
Yes, but timing in particular is luck. Like whether something is going to catch now or in two years, I mean it makes the whole difference for a company and it's really hard to predict really.
Yeah. So one thing that came back often when I was discussing with ex-Datadog, I have had quite a few founders came from Datadog or got acquired by you and were on the other side of the M&A after. Mentioned the culture and how you as the founder are still very much heavily involved in like some product decision and like that you don't, yeah, you are not at the strategic whatever PowerPoint level at all. Can you tell us a little more about how you see that and how you make these decisions?
Yes, so there's something that's very important is that I and the rest of the management team have to understand what really happens on the ground. Like, you know, the problem you have as the company grows is that you develop these beautiful stories and then people tend to manage up. They want to be able to tell good news to management, you know, and that's human. That's just the way people are wired, right? And so everything I do there in terms of the systems we put in place is to force people to look down instead of up. So to look at what's actually happening. What's happening with the customers? What's happening with the product? What's happening with the technology? What's working and not working?
And concretely what does that mean day-to-day?
So day-to-day what I do is I sample a lot of the stuff. So I sample, I'm going to read support requests. I'm going to read sales conversations. I'm going to read, actually I read all of the product briefs like all of the product releases and everything we do to make sure I understand exactly what's going on there. And what I will do is occasionally I will reply and I will ask questions like, you know, what's going on there or, you know, I don't understand this part or, and, you know, very often like the first thing that happens after that is the person who wrote that email, putting that conversation in the system says, 'Why is he looking at that?'
And what it does is it kind of jolts people a little bit and also forces the local management chain to look down and feel, and people ask themselves a better understand what's going on there because I maybe asked about it. And on an ongoing basis it forces people to understand exactly the reality under them as opposed to, you know, just reading a beautiful message. And on my end, you know, it sort of trains the machine, you know, like that's a, I get a sense of reality, a sample reality and then I can also figure out what agrees and doesn't agree with the stories I see at the higher level, you know.
That's your own training data.
You're fine tuning your own data.
And I think that's why also it's when you deploy AI in particular, like it's not good to automate everything like it still has to go through your brain at some point for you to get an understanding of what's going on. Yes.
Yes, yes, you cannot delegate decision like.
The to driving the company.
How do you make these product decisions then? Like it feels like you could also become the bottleneck here if people need your opinion on everything.
So we try to set up in a way that, you know, where feedback or executive feedback in particular is possible but not required. So things are going to move along anyway.
And they're going to move along with a number of communication points where, you know, you see what's happening, you see what's going to be released or you see what has been released. And feedback can happen at any point there but people don't wait for it. There's a few specific, you know, hard changes, for example, when you change the pricing and packaging of some products or things that can have a large impact and that are hard to back out of, that, you know, you have some pre-approvals. But for pretty much everything we build, people just push it out. And maybe there's feedback, maybe there's no feedback.
So if it's a life-threatening for the company, of course you have.
You are like a feature. But if it's like relatively low-stake decisions, nobody's waiting for anyone.
Or, you know, high-stakes but, you know, you can reverse them easily or, you know, it's very easy like in B2B, it's very easy also to test some products with some customers, make some changes for some customers, see how it works and actually that's actually how we build in general like we see at a small scale with some customers how it works, you know, whether it improves things for them a lot. And then we roll that up more broadly.
So you started with just infra monitoring. Like I think you were not even calling that observability yet and then you expanded to many other like aspects of observability as a high-level term. Like how do you decide to build a new product? Like what's your process here?
Yeah. So I mean, mostly we see what our customers do with our product. We see what kind of situation they put us in and what kind of extensions and scripting and other things they build around us and that gives us an idea of the problems they have and also where they think we belong. And so that for example, when we expanded from infrastructure monitoring at the time to APM, you know, tracing applications basically. We were seeing our customers build simple versions of that themselves on top of our platform. And so it was pretty clear, I know, they wanted us to do that.
