Datadog Co‑Founder Olivier Pomel on Culture, AI Strategy, Scaling

Olivier Pomel, co-founder of Datadog, shared insights into the company's journey, culture, and strategic decisions, including its unexpected path to success and adaptation to the rapidly changing tech landscape.

The Unconventional Path to Datadog

Datadog, a leading monitoring and observability platform, famously did not get into Y Combinator. Pomel recounted receiving a rejection email from Paul Graham, which he jokingly suggested should be framed as motivation. This early rejection, he believes, fueled their determination to prove critics wrong.

Pomel's journey to co-founding Datadog began in Paris, where he met his co-founder, Alexis Lê-Quôc, at CentraleSupélec. Their paths diverged briefly before reuniting at IBM Research in upstate New York, where Pomel initially intended to stay for only six months but ended up remaining for over two decades.

During the dot-com boom and bust, Pomel worked in various startups, gaining valuable lessons on what to do and, more importantly, what not to do. These experiences, coupled with the post-9/11 downturn in New York, led him and Lê-Quôc to an educational software startup. This company, which grew to 800 employees, became the crucible for their future venture. Pomel built the development team, while Lê-Quôc led operations, experiencing firsthand the friction between developers and operations teams. This "DevOps" challenge, where "developers hate operations, operations hate developers," became the core problem Datadog aimed to solve.

The Birth of Datadog and the Cloud Bet

Datadog was founded in 2010 with the premise of unifying development and operations on a single platform. At the time, monitoring was reactive and siloed, with separate tools for different aspects like network monitoring, often used exclusively by operations teams, leaving developers "blind to anything that happened in production."

Their initial bet was on bringing DevOps together, and their first infrastructure support focused on cloud environments. They initially targeted smaller, more modern companies that were already building on the cloud. This decision, though underestimated at the time, proved prescient as the cloud market exploded.

Initially, they struggled with how to brand their product. They avoided the term "monitoring," considering it "boring and old," and instead called it a "data platform for DevOps to collaborate." However, they soon realized that while people liked the idea of a new thing, they didn't know what to pay for it. Rebranding as "infrastructure monitoring" immediately resonated, as "everybody and their boss understands yes actually it's a product we need that it's a category it's a it's real let's get it."

Overcoming Rejections and Building a Unique Culture

Datadog faced significant challenges in its early days, including difficulty raising an angel round. VCs in New York were hesitant, not understanding the market, while Bay Area VCs were skeptical of an infrastructure company not based in Silicon Valley. Pomel recalled VCs speaking "slower and louder" to them, as if they had a "mental impairment."

These rejections, however, fueled a desire to prove them wrong. Datadog intentionally built a company culture distinct from typical Bay Area startups. They never formally wrote down their culture or values, believing that "if you need 'don't be evil' to be written on the wall, you probably shouldn't work here." Instead, they emphasized that "culture flows from the top," exemplified by leadership, and reinforced through hiring, promotion, and even firing decisions. This approach, even with 8,000 employees, continues to work for them.

Deep Involvement in Product and Staying Grounded

Pomel remains deeply involved in product decisions, despite the company's size. He expressed disdain for the term "strategic," often mentally replacing it with "bullshit." He believes it's crucial for leadership to stay connected to the product, reading new developments and providing critical feedback. He sees his role as an editor, asking questions like, "If I don't understand it, how are the customers going to understand it?"

To avoid becoming a bottleneck, he clarifies that he's not a necessary approval step but rather a critical voice. He encourages all leaders to stay in touch with "what's actually happening on the ground" by reviewing support requests, sales transcripts, customer feedback, and employee survey comments. This direct engagement helps him understand the "fabric of the universe" within the company, rather than relying solely on "beautiful summaries" that often sugarcoat reality. When he finds discrepancies, he simply asks questions, which prompts the management chain to investigate issues more thoroughly.

Expanding the Product Portfolio and Navigating Public Markets

Datadog's product portfolio has grown to 20-25 products. Their expansion strategy is largely driven by observing how customers use their existing products, building extensions and workflows around them. They also pursue "strategic" projects based on anticipating market shifts, particularly with the rapid changes brought by AI. Pomel acknowledges that these forward-looking bets carry a higher risk of being wrong, but are necessary to stay ahead.

Becoming a public company in 2019 brought new dynamics. Pomel recounted a dramatic lockup expiration during the COVID-19 market crash, where Datadog's valuation dropped significantly. While the stock market's volatility has become a familiar pattern, he notes that the core leadership approach hasn't drastically changed. As a private company, much time was spent courting investors; as a public company, investor relations are more choreographed, with quarterly calls and dedicated preparation.

The primary concern with stock fluctuations in a public company is not the company's survival, but rather the impact on employee compensation, particularly RSUs. Significant drops can lead to a "comp issue," making it harder to retain key talent.

The AI Revolution and the Future of Development

The advent of AI has presented both challenges and opportunities. Initially, investors were uncertain about Datadog's position in the AI landscape, but the company has since emerged as a "winner." Pomel notes the difficulty for public investors to grasp long-term implications due to their broader portfolio and shorter engagement cycles compared to VCs. Therefore, telling a consistent, compelling story backed by numbers is crucial.

Internally, AI is profoundly changing how Datadog operates. Pomel's co-founder, Alexis, declared that within two quarters, the engineering team would "not be writing any code anymore." While they will still write and read code, the emphasis is shifting from primarily writing to primarily automating. This means taking more shots and being comfortable with being wrong more often, as smaller teams can now explore ideas much faster.

They've witnessed "absolutely excellent developers" rebuild entire systems in days that previously would have taken a team of six months. This shift is leading to a painful but necessary organizational pivot.

Datadog is making several key bets in the AI space: - Agent-driven product usage: Investing in making their product usable by AI agents, reflecting customer trends. - In-product intelligence: Building more intelligence directly into the product, rather than relying on external tools like large language models for operational issues. This aims to fuse capabilities directly onto the data plane within their observability platform.

Datadog serves the top AI companies, numerous AI startups, and large enterprises, giving them a unique vantage point on the evolving AI landscape. While AI labs with "infinite compute" offer a glimpse into the future, Pomel acknowledges that their practices aren't always representative of the broader market.

Co-founder Relationship and Advice for European Startups

Pomel and Lê-Quôc have worked together for over 20 years. Their long-standing relationship, including nearly a decade of working together before co-founding Datadog, helped them establish effective working boundaries. They maintain a standing lunch, typically every two weeks, without a specific agenda, to simply talk and share thoughts. This open communication is vital, as there are few others with whom founders can truly confide.

For European startups, Pomel advises starting locally, as funding and resources are more accessible now than when Datadog began. However, he stresses the critical importance of targeting the US market as soon as a product achieves market fit. This often means one of the founders relocating to the US, as it's difficult to succeed there without a physical presence. Datadog itself established a Paris office later, primarily driven by talent acquisition and accommodating employees who wished to return to France.

His concluding advice to his younger self, and to aspiring founders, is to "always move faster, especially when it comes to hiring and firing." He suggests taking more risks in hiring, even if it means having to fire someone later, as waiting for the "perfect hire" can be detrimental. He notes that while firing is difficult, often the team's reaction is, "What took you so long?"