In Production
Redesigning participant pricing for data collectors
A data-collector-facing B2B project at Prolific. Redesigned how data collectors set participant pay — tested, iterated, and shipped to 100%. Drove 12.2% revenue uplift.
Product
Web app
Timeline
2026 · 2 months
Role
Lead product designer
Team
1 PM, 1 Data Scientist, 4 Engineers
Overview
A redesign that improved fairness for participants
At Prolific, data collectors set their own pay rates when publishing studies — but many consistently underpay. We redesigned the pay-setting experience to make the recommended rate the default to improve fairness for participants.
I led all UX and visual design as the sole product designer on this project — from early concept through to shipping. This included usability testing and a qualitative research work stream that ran in parallel with the A/B test.
The problem
Pay rates weren't keeping pace with platform growth
As Prolific scaled, the number of studies grew — but pay rates didn't. Data collectors set their own rates, and many defaulted toward the lower end without much guidance. This created a compounding problem: participants were earning less relative to the growing volume of work, and Prolific had no mechanism to nudge researchers toward more competitive pay.
Previous section where data collectors set pay rates
HMW statement
How might we increase pay rates for participants without undermining data collectors’ trust or reducing study volume?
Understanding the problem
Data collectors don’t increase their pay rate because…
No visibility on the marketplace dynamic
Data collectors don’t know what their target participants are seeing — what a competitive pay rate looks like, or how many options they have.
Budget sensitivity
Academics tend to have tighter constraints, while industry data collectors have more flexibility.
Ideating solutions
Soft nudge vs. setting the recommended price as default
I explored two directions with the PM: pre-filling a recommendation as the default, or showing a soft nudge alongside the researcher's own input. Prior behavioural data on our platform suggested defaults drive stronger shifts than nudges, and the trade-off was worth it — with participant earnings directly at stake, a bolder intervention was justified. From there, I iterated on how much context to surface, what language to use, and how to make overriding feel like an informed choice rather than friction.
Recommendation as default
Pre-filling the reward field with a market-rate recommendation, requiring data collectors to actively override it
Soft nudge
Surfacing market context as a passive nudge, without changing the default value.
Usability testing
Three critical issues found
After I designed the potential solution, I conducted moderated usability testing with 8 data collectors — a mix of academics and industry users recruited from the Prolific platform. Three findings shaped the final design:
Clearer CTA
"Change reward" wasn't intuitive enough. Several data collectors weren't immediately sure how to set their own rate.
Misleading copy
The word “recommended” landed badly in an academic context. They reacted negatively to being told what to charge.
More reassurance
They wanted reassurance, not just a number. They wanted more context and detail before accepting a new rate.
The solution
Showing the recommendation as default
We set the recommended rate as the default rather than an opt-in nudge — past behaviour showed data collectors rarely adjusted pre-filled values.
Clicking the CTA takes them to set their own reward
Once data collectors choose to set their own rate, we kept the recommended rate visible — so the reference point stays in view even when they override it.
A/B Testing
Business impacts
Partnering with Data Science, I designed and ran a 4-week A/B experiment across ~16,000 data collectors, split between treatment and control.
+12.2%
Revenue Uplift
Direct result of improved pricing confidence.
12.7%+
Participant Pay Rate Increase
Studies reached their sample size faster.
User survey to understand sentiment
I conducted a survey to understand whether the experiment might be causing broader unintended negative effects that we currently lacked visibility into. The survey showed no negative sentiment, and early signals suggested customers found the new design clearer than the control.
Reflection
What I'd change: invest in alignment earlier
Looking back, the hardest part wasn't the design work — it was aligning a wide set of stakeholders on a decision that directly affected revenue. Because the outcome touched pricing, marketplace pay quality, and researcher trust simultaneously, every design choice had a business owner attached to it. If I did this again, I'd invest earlier in structured alignment: a shared decision document, upfront trade-off framing, and clearer sequencing of when input was needed from whom.
What I've carried forward: autonomy and dual signals
When designing to shift behaviour, preserving user autonomy isn’t optional — it’s what makes the intervention sustainable rather than manipulative. The opt-in nudge only worked because researchers could still override it, and meant it. Quantitative results don’t stand alone. A 12.2% revenue lift would have felt hollow without the sentiment data showing researchers weren’t merely tolerating the change. Both signals had to move in the same direction.
Other projects
New sign up flow for expert participants
Redesigned the expert sign-up flow, cutting a 7-page form to 2 and adding a fast-track path around the waitlist. Shipped and A/B tested.
Redesigning the participant profile
A participant-facing project at Prolific. Redesigned the about you page to drive higher completion — shipped to 100%. Increased completion rate by 25% to 68%.
© 2026 Go Ogata - Product Designer