When a product is large enough to make a complete offline update impractical, should an end-to-end update start with core functionality and high-traffic screens deeper in the system, or from the top and spreading in a contiguous and convex manner?
The goal was to refrain from creating isolated pockets of updated UX, eliminating the need for users to navigate through legacy interfaces to get to updated areas as much as possible while the project was ongoing.
I updated the login experience and entry dashboard first, followed by the global search mechanism and then progressing to deeper levels of the information architecture based on traffic but without jumping over any nodes.
Certain critical components, such as the main workhorse modals, were updated early. These were the only exceptions to the convex approach, so the team wouldn’t need to maintain two versions of the same modals, and to prevent user confusion.
By adopting this strategy, I maintained a seamless user experience update, propagating the design over time without causing jarring switchbacks in the UX.
STRATEGY:
Contiguous Area of Updated UX