Supernova Award Category

Data to Decisions

The Problem

Soylent is based on science and data, but as the business grew, they faced challenges in allowing everyone in their organization to be as data-driven in their work.

All access to data centered around the data team, lead by Henry Hei. Requests for data were quickly outgrowing the data team so getting access to data was often delayed. This meant fewer decisions were made with data because they simply didn’t have the data they needed to do it. When they did receive the data they needed, reports were one-off and managed in spreadsheets across the organization. This lead to mismatching metrics in cross-functional meetings, and worse, a lack of trust in the data they were meant to use to be making decisions.

Henry knew that they either needed to hire many more analysts or they needed to take a new approach to analytics altogether.

The Solution

Soylent was still relatively small but growing quickly. Henry knew they needed to build a data stack from scratch that would empower users in every part of the business, but also establish important layers of governance so everyone was making decisions based on the same numbers. While they were looking at tools, they broke out the options into three categories: Classic BI, Workbook Analytics, and Cloud Analytics. They soon realized that the most scalable and efficient option would be Cloud Analytics. Among Cloud Analytics vendors, Looker stood out as the best choice because it allowed Henry and his team to work more efficiently, and allow everyone at the organization to access the data they needed to do their jobs. This decision also led them to reconsider their database solution, as they wanted to provide the best possible experience for their end users. In the end, they chose Amazon Redshift as that backend because it could meet their needs today, and scale with them in the future.

The results

Since the adoption of Looker as their internal BI tool, Soylent has seen a shift from both a cultural and organizational perspective. Henry equates the change to be night and day: “Before it was night and they didn’t see the data, different teams would light candles to see in the dark but they could only see parts. Now the light is on and everyone can see everything.” But the change didn’t come overnight. Henry and his team invested months into building relationships across the company and creating embedded data ambassadors on each team. They now operate with a hub and spoke model, with the central organizational metrics coming from the data team (the hub) and then ambassadors on each team customize metrics and reports for their specific teams to deliver exactly what they need (the spokes). This new approach has lead to Looker use within every single department, vs. almost zero engagement seen with previous BI tool. Operationally, teams are now working differently with this new, easy access to data. For example, product launches have dashboards that track towards goals and benchmark success metrics in real time, which was impossible before. With decentralized analytics, the data team spent almost half of their time on basic reporting. Now they have automated much of that work, and the data team is spending less time on basic reporting and more time on high-value problems such as increasing infrastructure efficiency and investing in more Data Science.

Metrics

The data team is saving 10 - 20 hours per week on delivering data and reporting requests, time that is now spent on strategic initiatives such as feature building, deep dives into performance, and cross-functional collaboration.

Data efficiency through centralization, org-wide engagement, and self-service capabilities is resulting in 7.5x ROI on their investment in their data stack to date.

The Technology

Looker, Amazon Redshift, and Fivetran

Disruptive Factor

In the consumer packaged goods space innovation is pretty rare from the product and the technology perspective, but Soylent is quickly becoming a disrupter on both sides of the equation. Because the brand started out as an e-commerce business, they have a full e-commerce stack that they’re bringing to their venture into Retail.

Because of the new, open infrastructure and time available, Henry and his team are able to take what is traditionally disparate data from the stores Soylent is in - like 7 Eleven and Target - and deliver that to their team alongside the rest of their data. They have a complete picture of the performance of their products both online and off - a feat that is out of reach for many brands who are doing the traditional switch from retail to digital.

This access to data is allowing Soylent to be more strategic in their shift into the retail world than both their new world and old world counterparts alike.

Shining Moment

Henry says he has been most surprised by the speed and abruptness of the change in the company’s culture. “Before there would be all these requests, then they go down by 90% all of a sudden” he explains. “80% of what people normally ask for is covered by the baseline metrics and reports we built out. Now, I am able to give my team more challenging and fulfilling problems to solve, and it allows us to be better partners to the rest of the organization.”