Adapting to change
Replacing destructive edits with a flexible data model so users could evolve their setup without starting over.
- Role
- Product designer
- Team
- Product, Engineering, Support
Our workspace in FigJam
Context
In Funnel, users manually upload data when an automated connection isn't available. To do that, they configure the dataset — fields, types, the shape of the data. That configuration is reused for every future upload.
The challenge
Once set, the configuration couldn't be changed. Renaming a field or fixing a type meant deleting everything and starting over — or filing a support ticket so we could patch it from the backend. Painful and slow for the user; expensive for us.
The idea
Let users edit the configuration. Less frustration for them, fewer support tickets for us.
The scope
Making everything editable in one go meant a long development cycle while customers kept struggling. We chose to ship one editable option first and learn.
The process
I reviewed historic support conversations, feature requests and system events to find the most-needed edit — adding and removing schema fields. I ran interviews with affected users to sharpen the problem and pulled developers in so they could feel the friction first-hand. We prototyped three approaches in Figma and tested them, picking the one that felt closest to the existing configuration setup. I spent a lot of time scenario-mapping in FigJam to nail how edits would behave across dataset states.
Launch & conclusions
We released carefully, with clear copy around which options were editable and which weren't. Inbound tickets on this surface dropped sharply, and users started experimenting with their datasets in ways they previously couldn't — which surfaced the next round of edits worth unlocking.
Eliminated a top-3 driver of support tickets and unlocked a continuous editing roadmap.