Designing trust as a requirement, not a message
In civic and financial products, credibility has acceptance criteria.
Building Ease Voting made one thing obvious: people do not distrust results, they distrust gaps. The silence between a count and an announcement is where confidence is lost.
Once trust is framed as a requirement, it becomes testable. Can an observer independently verify this number? Can a voter see the same state the administrator sees?
Start with the sceptic
I model the most sceptical persona first. If the platform satisfies someone actively looking for a reason to doubt it, every other persona is already served.
That inversion changes the interface: fewer summaries, more traceable primitives, and no state that exists only in the admin's view.
Working on something like this?
I partner with founders and teams on discovery, requirements and delivery.
Get in touch