Hero approached the project with intensive research, speaking to vice presidents of engineering at top enterprise startups and several engineering teams that had built their own code-release tools. We also met with the development, sales, executive, and marketing teams at Split to determine which user flows and features were critical to success. That exercise helped the team hone in on the key pain-points developers were experiencing, how they were addressing those problems internally, and what sort of UX approaches might help developers in the future.
With market validation, customer feedback, and user-persona research complete, the team set to work on design. Focusing first on the Split’s editor, they helped plan out what features it needed to have, how they would work, and how that functionality would be communicated to users. The team also talked through the brand and visual design of Split, deciding that it should be visually inviting and easy-to-use, but still harness the power of an enterprise-grade platform.
In just three months Hero delivered everything Split needed to build the next phase of its service. About 70% of that work was included in the company’s October 2016 release, and the rest will be incorporated into the product over the next several months. Even before those new features are added, the service goes well beyond any alternatives in its capabilities and ease-of-use. Split is now a controlled-rollout platform that goes beyond feature flags, enabling anyone to manage feature rollout for multiple languages, target users, monitor performance, and kill problem features.