Why I’m Joining Eppo
I’m excited to announce that I’m joining Eppo to build the next generation of A/B experimentation tooling. When Che first approached me about joining him in building Eppo, the idea instantly made sense. I’ve spent my entire career trying to foster data-driven cultures, and have come to the conclusion that experimentation is far and away the best way to get there. A mature experimentation practice uniquely enables a company to learn which ideas succeed and which do not. In the global capitalist economy, the organizations that survive will be those that can learn and adapt the fastest. The company that fails to learn and adapt quickly enough will themselves become yet another rejected hypothesis.
The decision to leave Snowflake was tough, especially given the company’s recent blockbuster IPO. At Snowflake, I also had a front-row seat to some of the most exciting developments in the data ecosystem. In my time there, I became more and more convinced that Snowflake had solidified an ubiquitous and indispensable position as the data platform of the future. It was precisely this vantage point that led me to join Eppo. In my customer-facing role, I saw how an effective experimentation process transformed many of Snowflake’s successful customers, and saw an exciting new software architecture paradigm begin to form.
The Next Generation of SaaS
Much has already been written about the Modern Data Stack, but a far more interesting pattern in my mind is what Tomasz Tunguz recently coined “Cloud Prem Architecture”. Snowflake has largely solved the technical challenge of operating data at massive scale, and tools like Fivetran allow companies to integrate all their various data sources into one place. This new generation of Cloud Prem software will operate off of these platforms to provide both best-of-breed capabilities and data interoperability. The latter piece is especially important. With the unbundling of enterprise software into a constellation of SaaS applications, end-users today have access to much better point solutions, but the organization as a whole has less visibility over the full picture. Cloud Prem software offers the promise of the best of both worlds. In fact, I was struck while reading Tomasz’s post by the similarities between his “Data hub-and-spoke” architecture, and this fantastically 90s diagram of the SAP R/3 system:
When the whole of your enterprise data is in one place, that becomes the place to build and run business applications. That was true back in 1991, and remains true in 2021. Today, there is a fairly limited set of applications that take advantage of this new architecture. BI tools like Looker and Mode were the first on the scene, and a new crop of startups is building observability, catalog, and CDP solutions; but by and large, the vast majority of business applications are still waiting to be built. That brings me to Eppo.
At Eppo, we are building the future of Online Experimentation — an experimentation platform that integrates deeply with customers’ data warehouses. We believe that experimentation is the most critical workflow that data teams can unlock for their organizations; and yet, today, only the most mature, well-resourced companies have platforms that truly enable an iterative, entrepreneurial culture. With the rise of the Modern Data Stack described above, the biggest barriers to creating such a platform no longer exist. Business Intelligence tools allow companies to inspect the past, but an experimentation tool like Eppo will allow those same companies to shape their future.
Early Team, Early Dream
Thus far, I’ve talked about industry trends and market opportunity, but the reality is the most important factor in my decision was the team behind Eppo. I knew Che from our time together in the Statistics program at Stanford where we spent many late nights working together on problem sets and group projects. Post-Stanford, we kept in touch as we followed parallel tracks through the industry. Despite working at a completely different types of companies, we shared very similar perspectives on the data landscape. In our early conversations about Eppo, he connected me with some of his partners through the nascent stages of his journey, including Sarah Catanzaro at Amplify Partners, and Keenan Rice, part of Looker’s founding team. It was clear from the outset that there’s a village of data veterans behind us, rooting for our success, and helping in any way they can.
The team thus far is incredible, but it’s only the start. If you’re interested in being part of a world-class team building the future of experimentation, visit us at geteppo.com, or drop me a note at carlin@geteppo.com.
