Q2 2024 - Q3 2025
My role: Product designer
My partners: Two squads of engineers, 2PMs, EM, PMM
Added the ability to pay international suppliers; with 12% of eligible customers adopting three months post-GA, contributing to a −57% gap to the projected revenue. Led the design from beginning to end under tight resource constraints due to restructuring. Ran UX experiments to improve adoption.
Previously, Spendesk did not offer customers the ability to pay their international suppliers via wire transfer. As a result, finance teams needed to handle these payments with a competitor or with a traditional bank, resulting in fragmented workflows and missed revenue opportunities for Spendesk. This project aimed to integrate international payments directly into Spendesk via a partnership with Wise, enabling finance teams to manage global supplier payments in one platform.
I was on the project from the beginning, originally leading the user research to discover what our users needed to manage their international payments, and later to handle the UX of the flow, bridging the technical requirements with the requirements with Wise.
During discovery, it was clear there were many differences in experience between the current domestic payment flow, and the new international flow, including additional forms that the user needed to fill and additional details on fees. I prototyped several different iterations, focusing on entry points to the new forms and hierarchy of information and then did external and internal user tests with different finance teams. Based on the feedback I received, I went with a version that remained close to the original flow, but still included the necessary additions for making international transfers.
In the beta phase, we began to notice trends of poor adoption. To understand what was happening, the PM and I spoke with the beta customers to hear their feedback. It became clear that the primary barriers to adoption was the increased workload due to the required amount of recipient details and fees that were not competitive enough with traditional banks. As the pricing and packaging was out of my control, I passed the feedback on to the responsible team and focused on improving the recipient form to reduce the manual work required and improve the understanding of the requirements. This included adding some helper text on transfer types and auto-populating any information we had file for the supplier.
Several weeks before launch, we learned from Wise that our KYC process was not aligned to theirs, requiring adopting users to provide additional documents in order to access the feature. I flagged this immediately as a barrier to adoption, as we had learned in a former initiative that customers are very reluctant to perform extra tasks that interrupt their workflows. Initially, I designed a totally in-app workflow to collect the information while minimally disrupting workflows, but due to a lack of engineering staffing and inflexibility of launch date, we had to reduce to the bare minimum flow.
After the feature went GA and we continued to have adoption problems, we noticed that indeed, the KYC steps were preventing customers from adopting, with only about 10 customers remediating per week. In order to increase this number, I did a series of experiments using in-app nudges and banners. After they were deployed, we have had an average of 115 customers remediate per week, which is an increase of 1050%. As we continue, this number should go down as more users have access to international wire transfers.
At the beginning of the challenging implementation, Spendesk went through a restructuring, laying off a third of its staff. This resulted in a great lack of morale among the product teams, and a greatly reduced number of engineers, as our workforce shrank more than had been intended. As a result, we needed to ship a less-than-ideal MVP and weren’t able to commit to all the improvements we had hoped.
At three months post-GA, adoption among approved customers is 12.0% (88 of 731); 87.96% remain non‑adopting. I am still working with the PM and PMM to run experiments to improve these numbers, the next ones are focused on measuring the ROI of the feature so we can improve the messaging on the benefits it will give to users.
