Tags: Design Leadership 路 Product Design 路 Cross-functional 路 International Markets 路 A/B Testing
Context: BOLD (Zety), 2021-2023
Type: Design Leadership + IC contribution
Two CV builder platforms existed within BOLD simultaneously: NGB, built and owned by the Polish product team, and MPR, developed by the American team. The question of which performed better had never been tested head-to-head. France was a new market, which created a rare opportunity to run a direct comparison before committing the French audience to either platform long-term.
The challenge was not to localise an existing product. It was to design a version of NGB that could genuinely compete against MPR with French users, which required understanding what French users actually needed, not just translating the interface.
The two platforms differed in one important respect beyond their design: NGB was developed with user research embedded in the process from the start. Our team conducted market and user research before every significant design decision. MPR was built by a separate team that did not follow the same approach. For France, that difference in methodology would turn out to matter.
I led the design team responsible for the NGB side of the experiment. I was also directly involved in the exploratory phase: working with the team to develop hypotheses about the French market and identifying what product-level changes were needed to give NGB a genuine competitive edge on this specific market.
Before any design work started, we conducted research into the French market: its expectations, conventions and the specific ways French users approach job applications. This happened in close collaboration with the people internally responsible for the French market, who brought domain knowledge we could not have sourced any other way.
One finding stood out: cover letters carry significantly more weight in French professional culture than in most other markets where BOLD operated. French employers expect a well-crafted cover letter as a standard part of any application, not an optional add-on.
This finding became the central design hypothesis for the experiment.