A next gen A/B experimentation platform that enables an entrepreneurial culture.
Most companies are filled with untapped entrepreneurial potential. They will spend enormous resources bringing in the best and brightest talent, but end up isolating product decision making to a small group with seniority. This type of "HiPPO" or "highest paid person's opinion" decision making isn't the original intent, but it is an inevitable result of an inability to tie business outcomes to product teams.
Fortunately, the internet enables a new, customer centric mode of decision making. This new regime leads to quick product iterations, risk mitigation, and ultimately experiences that customers truly want. The solution is AB product experiments.
As a company grows, entrepreneurial culture becomes inextricably linked to experimentation. Metrics aren't actionable until you can run experiments against them. Without experimentation, teams might glance at a company wide dashboard once in a while. With experimentation, their daily work becomes inexorably linked to metrics, tightly coupling product iterations and business ROI.
Unfortunately, scalable experimentation requires tooling that doesn't yet exist on the market.
Commercial experimentation technology is woefully inadequate for a product experimentation practice. The handful of available tools were built for cosmetic tests around adjusting text or images, and are unusable for today's product teams:
Today's commercial experimentation offerings are feature flagging tools masquerading as experiment platforms. This leads to unexpected agony for companies who pay hundreds of thousands of dollars for an 'experimentation platform,' only to discover that they additionally have to roll their own data infrastructure and hire legions of analysts to do all of the operational work.
With no viable commercial options, companies rebuild the same experimentation tooling over and over. Whether at a unicorn or a company of a few hundred people, organizations will staff an internal team to do piecemeal automations of their experimentation workflows. Across every product led growth company, they will purchase G Suite, purchase Snowflake, purchase a BI tool like Looker, ...and then hire a bunch of engineers to build the same experiment tool from scratch.
Inevitably, this in-house tool turns into a multi year engineering investment that still leaves everyone doing heavy manual operations work. Rather than invest even more resources to accomplish these steps effectively, organizations settle for broken workflows, painful analysis experiences, and experiment velocity at a snail's pace.
This means only mature companies get to enjoy mature experimentation tooling. When you have to build from scratch every time, you never build something fully developed. Most in-house tools involve setting up bespoke dashboards for each experiment, use subpar statistical methodology like t-tests and chi-square tests, and don't take advantage of variance reduction techniques that can reduce runtime by up to 50%.
There is a common set of experiment functionality that suits the vast majority of businesses. This is especially true given secular trends of businesses towards cloud infrastructure, widespread adoption of product led growth strategies, and data teams using a homogenous data stack of SaaS products. The difference between each company's experimentation tooling is not a difference in requirements, but of their level of investment in internal tooling.
We have experience building experimentation practices from scratch, multiple times. After contributing to the experimentation system at Airbnb, the CEO built the same tool at Webflow. Despite Webflow being a subscription SaaS business and Airbnb being a marketplace, the same Airbnb experiment architecture worked in both environments. But in both instances, much more engineering investment was needed to fully automate the operations work that analysts were doing.
Since then, we have spoken with dozens of growth experimentation teams. Companies like Lyft and Facebook had tooling with over a decade of investment, others from earlier companies were trying to manage off of Googledocs. The common thread was a.) the market not having an adequate experimentation offering and b.) internal tools teams not having resources to build the features they know would be valuable. We came away from this experience believing that the ecosystem needs a team that is singly focused on building best in class experimentation tooling.
At Eppo, we are creating a next gen platform that fulfills the aspirations of experimentation tools teams, and brings those capabilities to companies of every stage and industry. Today, companies can't use any of the options on the market, and they can't run experiments until they devote multiple engineers and analysts to a function that is unrelated to their core product. We want to build a world where every team can have entrepreneurial, high experiment velocity culture.
Eppo is an opportunity to change corporate culture everywhere. The mission of every business intelligence company is to bring the voice of users into decision making. As all data teams know well, nothing affects company decision making as much as introducing experimentation to product development.
If our goal of empowering internal entrepreneurs resonates with you, we would love to chat! We are currently looking to build our founding team. If technical problems of distributed systems, large scale data processing, and intuitive interfaces to serve data resonate with you, please get in touch.
We'd love to hear from you at firstname.lastname@example.org