Civic non-profit platform

Donation flow design

One Dollar / Day: From a suspected scam to a verified match

One Dollar / Day: From a suspected scam to a verified match

One Dollar / Day: From a suspected scam to a verified match

One Dollar / day is a project of DemocracyFirst, a Dutch non-profit that has built tools for deciding together since 2008. The idea is that everyone gets a dollar a day and chooses to keep it or give it to someone who needs it more. A petition asks each government to fund that dollar. When I joined Democracy First Foundation as a product designer, the MVP had most screens but no logic tying them together. In our first round of user testing, four of five people questioned whether it was legitimate. With the other designers in the team, I rebuilt the giver flow around one question: what would make a stranger trust this enough to come back tomorrow?

Outcome

The giver flow now opens on one verified person and shows where every cent goes. A second round of testing is scheduled to check whether it earns trust.

ROLE

Product Designer

TEAM

Founder, Product Designers, Software Engineer

Matched user profile for the giver

Why it mattered

A dollar a day only works if the giver still trusts it tomorrow.

A dollar a day only works if the giver still trusts it tomorrow.

A dollar a day only works if the giver still trusts it tomorrow.

Every giver makes one choice, keep or give a dollar a day, and it repeats automatically until they they change it. The platform isn't built to make money. It grows when people trust it enough the share the link. Every doubt has a cost.

Giver

They're sending money across borders to someone they've never met. When that feels off, they leave.

Receivers

A dollar a day only adds up if it keeps arriving. A giver who drops off after a week takes the dollar with them.

One Dollar / Day

A giver who doesn't trust it won't share it, and sharing is how it grows.

What I owned

I owned the trust layer, the testing readout, and the handoff to engineering.

I owned the trust layer, the testing readout, and the handoff to engineering.

I owned the trust layer, the testing readout, and the handoff to engineering.

Communication with the engineering team on Linear

Donation flow

I audited the MVP screen by screen, then redesigned matching, giving, and the dashboard. With every receiver capped at a dollar a day, I reframed “Choose amount” as “Set commitment,” so givers choose how long and how many people, not how much.

Testing and readout

I turned five test sessions into per-participant notes, seven insights, and three decisions the founder had to make. I’m running round two as an A/B test of the match page.

Design to build in AI

Before I joined, everything lived in Figma. I moved the team to Claude Design, design system included, and turned the written MVP into the first live prototype. I write implementation tickets in Linear and resolve open questions with the engineer in the comments.

Where it broke

The MVP had every screen. Nothing connected them.

The MVP had every screen. Nothing connected them.

The MVP had every screen. Nothing connected them.

Earlier designers had built screens for the petition, the choice, giving, and a dashboard. The logic between them was missing. I audited every screen, and three issues topped my list.

Key issues:

Key issues:

Key issues:

1

1

1

Three ways to pick a receiver, all competing.

Help me choose and Choose someone sat side by side. Testers used Help me choose like a scroll, asking for someone new and expecting to go back.

2

2

2

Browsing felt like a dating app.

Testing before I joined found that scrolling through receiver profiles made givers pick favorites.

3

3

3

A dashboard competing with itself.

The petition took the top spot, the statistics ran long, and nothing said which step came next.

Freely browsing through receiver profiles made users feel like it's a dating app

A dashboard that lacks a clear layout, and users aren't sure what happens next

Before designing screens, I mapped the giver flow end to end. That map went into the first round of testing.

What testing found

In round one of testing, we showed it to five people with no explanation. Four wondered if it was a scam.

In round one of testing, we showed it to five people with no explanation. Four wondered if it was a scam.

In round one of testing, we showed it to five people with no explanation. Four wondered if it was a scam.

No context was given to external users on purpose and I watched where they hesitated. Four of five questioned whether it was legitimate. All five had questions about the petition. Two asked what solidarity means.

1

1

1

People couldn't tell what it was

"Is this a fundraising app?"

2

2

2

It hadn't earned trust

Testers wanted to know who runs it and whether givers and receivers are verified.

3

3

3

The petition and giving flow blurred together

"Is this the same as petition?"

4

4

4

Wording gets in the way

"What exactly is the surcharge?"

"The English needs to be corrected, it makes me question the credibility of the site."

Translating feedback from user testing into presentable slide decks for the founder

💭

💭

💭

How might we earn a skeptical giver's trust at every moment they hesitate, so they give today and keep giving tomorrow?

How might we earn a skeptical giver's trust at every moment they hesitate, so they give today and keep giving tomorrow?

Key decisions

Four calls that shaped the product.

Four calls that shaped the product.

Four calls that shaped the product.

01) The petition leads after all

In my first pass on the MVP, I flagged the petition's top spot on the dashboard as a flaw. Testing proved me wrong. One tester only understood the idea once they reached the petition page's why-sign section. People need to see the idea before they'll trust the giving, and the petition is where the idea lives.

02) One person first, browsing last

Testing before I joined showed that browsing receiver profiles felt like a dating app. Givers picked favorites. The founder still wanted browsing kept, so instead of cutting it, I changed what comes first. With each iteration, I merged the competing ways to pick a receiver until the screen opened on a person.


I’m running the next round as an A/B test: filters open on top, or tucked behind an icon so the profile comes first. It asks whether givers want control or a simpler first look.

Matched user profile for the giver

03) Proof where the doubt happens

Testers' doubts showed up at specific moments, so I answered each one at that moment, not on a page they'd have to find.

Giver is allowed to endorse receivers to prove legitimacy

Clear financial breakdown during the payment stage

04) Plain words over the founding vocabulary

Solidarity was the founding idea, and testers found it foreign. Connect read as messaging. Both words went. The idea behind Connect stayed - the freedom to support one person directly.

Users are asked what they'd do if they got a dollar

The experience

Current design. Every fix answers something a tester said.

Current design. Every fix answers something a tester said.

Current design. Every fix answers something a tester said.

Dashboard for the giver

Key improvement so far

  • A verified badge and endorsements on every match

  • One suggested match, optional filters, browsing as a fall back

  • A detailed breakdown of finances

  • Step cards with a progress bar

Learnings and reflections

The giver flow is ready for strangers. Round two is the real test.

The giver flow is ready for strangers. Round two is the real test.

The giver flow is ready for strangers. Round two is the real test.

🔍

With the work

In the first pass on the MVP, I asked why the dashboard gave the petition top billing over choosing and giving. Testing answered me. One tester only understood the idea once they reached the petition page's why-sign section. People need the idea before they'll trust the giving, and the petition is where the idea lives. The product now leads with it.

⚖️

With the users

Our first external testers were people I know , and they still called it a scam. Strangers will be harder to win. The receiver side hasn't been a focus yet, and it will need its own round.

What's next

Round two of testing is scheduled. It checks whether the landing explains the idea on its own, how givers choose a receiver, and when and why they'd come back to the dashboard. That last question is our first real read on retention. This page will be updated when the results are in.


Round two is with strangers. That's where this flow has to earn its trust.

Matthew Liang

Product Designer