Neon One, 2018
Early-2010s web editors were clunky and routinely produced badly designed pages. Clients with dev resources were happy because they got free rein with HTML/CSS/JS to build their site however they wanted. Everyone else either published something embarrassing or asked our engineers to build their site for them. Engineering time essentially turned into a services-function.
I designed a template-based drag-and-drop site builder — deliberately not a full Wix-style site builder, supporting a set of site designs sharing the same variable elements: fonts, colors, navigation, homepage media, buttons, widget management. The design constraint that made it work was ensuring it was flexible enough that a client could legitimately match their brand, but constrained enough that every fundraising widget (progress bars, leaderboards, countdown timers, featured pages) always rendered correctly with live platform data. The site builder was personalizable, but always on-brand and compliant with the data model.
Adoption hit 100% of the user base, despite requiring customers to migrate off their existing custom sites, which is the strongest signal that the tradeoff was calibrated right. Customization work moved from engineering to the clients themselves, and the dev team got a meaningful chunk of its time back.
Product sense on a problem I identified myself: a real user-and-cost pain, a solution whose core design decision was what not to allow, and an outcome measured in both adoption and internal leverage. We later built on top of this foundation to deliver a Liquid templating system powering white-label sites for a top-5 reseller client and a 2024 integration with a third-party CMS.