Back to stories

Experimentation for the Modern Growth Stack: Our Investment in Eppo

By Naomi Pilosof IonitaJune 22, 2022

Experimentation is a key lever for growth, yet 90% of experimentation workflows today are manual. To get scalable, end-to-end experimentation, companies today have to either allocate scarce engineering resources to build an in-house system, or pay hundreds of thousands of dollars for a commercial tool while still managing their own data infrastructure.

As companies tighten their belts, product teams are under increasing pressure to deliver hard, attributable ROI. Experimentation is the only way to do this, making it critical infrastructure in a recession and a pillar of what I call the Modern Growth Stack—new tooling to accelerate customer metrics such as activation, monetization, and retention. Having spent years running full-stack product growth teams, I understand how success hinges on the velocity of learning, iterating, and driving metrics. Our work was a marriage of quant and qual, with experimentation being core to how we attributed our impact on business KPIs. 

Commercial experimentation technology today is woefully inadequate for modern growth practices. Legacy tools target cosmetic marketing changes (i.e., variations of copy) and base success on click-throughs vs. ultimate business impact like revenue. Poor infrastructure leads to experimentation bottlenecks across a series of personas—growth PMs and engineers have backlogs of experiments to run and data folks are saddled with endless statistics-heavy analysis. And the process doesn’t end with publishing results—often A/B test results drive more questions than answers, and the cycle of ad hoc analysis continues across Excel files and Jupyter notebooks. The result is that large tech companies such as Netflix, Lyft, and Airbnb end up building their own experimentation platforms and reporting dashboards internally. Companies that lack the resources to build end up hacking together existing tools that weren’t purpose-built for this beginning-to-end workflow.

Our newest investment, Eppo, positions itself as “Experimentation for the Modern Data Stack.” Eppo powers fast, reliable, and informative experiments. The platform can be used on its own, end-to-end, or paired with existing feature flagging tools such as Optimizely, LaunchDarkly, or in-house solutions. It integrates directly with data warehouses (Snowflake, Amazon Redshift) so that experiments use the same high quality data as BI tools like Tableau and Looker. Rather than measuring basic click-throughs, Eppo leverages core business metrics such as revenue, margin, and subscriptions, allowing companies to align product/growth teams with the same metrics that the C-suite cares about—an even more urgent need in an economic downturn.

Eppo is also the first commercial platform to offer CUPED, a statistical method for experiment acceleration. CUPED allows product teams to conclude experiments up to 65% faster, thus compressing the cycle time needed to learn and drive impact. This also unlocks companies with lower data volumes, giving them the ability to do rapid A/B testing as effectively as companies with massive user bases. 

Customers appreciate how Eppo is privacy-first. While other experimentation platforms host detailed customer data, Eppo processes data in the customer’s cloud and only stores anonymized test results, thus winning over engineering and security teams as well.

Eppo CEO and Founder Chetan “Che” Sharma was an early employee at Airbnb and one of the data scientists instrumental in standing up the company’s experimentation culture and platform. After Airbnb, Che built the experimentation system at Webflow, recognizing that despite differences in company products, the experiment architecture was the same. The opportunity for a platform like Eppo to help bring modern experimentation capabilities to startups and enterprises alike was clear. I first met Che via my colleague Grace Ge and we had been following his progress together for a year before leading the Series A. Menlo was impressed with Che’s deep technical understanding of the space, product velocity, and innate ability to recruit and sell—all of which led us to jump at the opportunity to partner with him. 

I love technologies that help accelerate growth and Eppo is in great company with some of our other portfolio investments in B2B SaaS such as Endgame, Matik, and Vivun. If you are a founder with expertise in the space, we’d love to hear from you!