Building a culture of experimentation at Stash.
Stash is a New York City-based financial services company that believes everyone should have access to the tools, guidance, and confidence needed to grow personal wealth and live a better life. They offer a do-it-all home for your money, including banking, investing, personalized guidance, simple automation, and education.
Stash is used by over 4 million customers across the US. As a digital financial services company, Stash is constantly looking for new ways to innovate and deliver unmatched customer experiences. As a part of this effort, Stash is building a culture of experimentation across the entire organization. On the customer loyalty engineering team, for example, developers are constantly running tests to measure the effectiveness of customer programs such as referral incentives, Stock-Back® Rewards, and more.
As the company continues to experiment across the platform, it has also worked to modernize its software development practices. The software organization has pursued changes aimed at helping them release new functionality early and continuously, in keeping with Agile principles.
While Stash made great strides in its experimentation and software delivery efforts, LaunchDarkly helped Stash take things a step further by infusing experimentation across their entire stack.
Stash adopted LaunchDarkly to serve a variety of needs. Kahne Raja, an Engineering Manager at Stash, relies on LaunchDarkly to manage phased rollouts and run experiments in production. When running these experiments, the team uses LaunchDarkly to target different user groups with different feature variations.
LaunchDarkly contributed to Stash’s goal of using experimentation and analytics as Stash continues to focus on data-driven results. Stash built a sophisticated internal data platform and intelligence tool called Wisdom that aggregates customer data from all of their experimentation and testing platforms. Before LaunchDarkly, this aggregated view, while valuable, did not include back-end testing and experimentation data. With assistance from LaunchDarkly’s Data Export Add-on, Stash can now transfer all LaunchDarkly feature flag data into Wisdom, including for back-end tests, and receive a full, granular view into customer engagement on the website and mobile app. This synergy, in turn, drives future product enhancements.
In addition to the front-end experiments that Stash engineers have found success with in the past, they can now run more advanced experiments on back-end infrastructure using LaunchDarkly. For example, Stash engineers are better equipped to measure the impact of their new content personalization services that automatically curate content for visitors.
Stash’s engineering team also uses LaunchDarkly to perform gradual rollouts and progressively deliver new features to their customers. LaunchDarkly feature flags enable Stash to deploy features to production while only exposing those features to select groups (e.g., a beta testing cohort, a canary group, etc.). As a result, Stash is more capable of delivering new functionality to customers in a much faster and more secure fashion.
Stash’s engineering team has not only reduced the time it takes to deploy and release new features, but they are also leveraging user targeting and gradual rollouts to deliver more personalized user experiences. Accordingly, customer satisfaction has increased.
Further, by enabling Stash to use feature flags at scale, LaunchDarkly has also helped reduce the burden on the engineering team by assisting those outside of Engineering (e.g., product managers) with determining when to release new features, and to whom. This has helped Stash foster more cross-team collaboration when it comes to building and delivering software.
"In general, LaunchDarkly has allowed us to leverage back-end-driven experimentation across all of our client platforms (iOS, Android, Web, etc.)," said Kahne. "This has fueled valuable improvements across the platform. For example, we now tailor the Stash home screen to each individual user based on past engagement data. We’re providing more value to each person who engages with us, which, in turn, helps us win and retain customers."
Disclosure: This material is for informational purposes only.