I'd previously developed a card selector concept and had had it tested in a usability lab three months earlier, where it performed well: a dynamic on-page filter with three parameter dropdowns for card category, reward type, and fee structure, reactively updating the card grid in real time. When the redesign reached its own testing phase, I brought that pattern in as one of the variants. Senior leadership was split on which direction to take into the full redesign with the Design VP in New York favouring the dynamic filter but with the UK VP and other senior stakeholders preferring a simpler static alternative. Comparative testing of the homepage variants was already part of the process, so the disagreement never needed resolving in a meeting room; both directions went into the lab alongside a simpler four-card layout and the existing homepage as a baseline. Tested as part of a full page rather than in isolation, the dynamic filter no longer held up. Users found it cognitively overwhelming and over complex for their needs which was a different response to the same pattern that had tested well on its own. The simpler four-column layout tested better on comprehension, cognitive load, and user preference.