Designing trust for first-time P2P traders
New users were completing their first P2P trade at half the rate of experienced ones. The UI wasn't broken — users just didn't trust what they didn't understand. Here's how we fixed that.
Role
Users
Scope
Ship
+4pt
Express completion rate lift (69% → 73%) in 14-day A/B test
−21%
Order cancellation rate from timeout down from 22.1% to 17.4%
68.6%
Overall P2P completion rate post-launch, up from baseline
01 / Problem
New users completed their first trade at half the rate of experienced ones
OKX P2P is the sole on-ramp channel for users in offshore and underserved markets no credit card, no bank integration, just peer-to-peer trading. For the platform to grow, new users must successfully complete their first trade.
~40%
First-trade completion rate for new P2P users vs. ~79% for experienced ones a 39-point gap that weakened activation, liquidity, and trust across the entire marketplace.
Why it matters
Low first-trade completion doesn't just hurt activation numbers. It creates a trust spiral: users who abandon a P2P trade rarely return. The problem compounds at market level fewer new takers means thinner liquidity for merchants.
Primary affected segment: first-time P2P users, especially crypto beginners unfamiliar with peer-to-peer mechanics.
P2P requires something fundamentally different from other fintech flows: users must trust a stranger and send real money to them before receiving anything. OKX holds the crypto in escrow but new users didn't know that. They had no mental model for what they were doing.
02 / Discovery
The existing UI wasn't broken. Users just didn't trust what they didn't understand.
I led moderated research with 6 new-to-P2P users across Vietnam, Argentina, Philippines, and Thailand recruited from our existing user panel, incentivised. We combined heuristic audit with direct user sessions.
Blocker 01
Coachmarks were ignored entirely
Users reflexively closed the onboarding overlay same behaviour as any other app's permission prompt. Zero retention. We had passive education, and it was failing invisibly.
= Passive education failed
Blocker 02
P2P terminology was meaningless
New users couldn't define P2P, or explain why "Express" differed from the regular marketplace. They lacked any mental model for what a peer-to-peer trade actually meant.
= No mental model
Blocker 03
Seller matching felt opaque and scary
"Why was I matched with this person? What if they don't pay?" Anxiety before the first action meant users abandoned before placing a single order.
= Distrusted system
““First-time P2P users were not mainly blocked by navigation. They were blocked by uncertainty, timing anxiety, and trust gaps”
Misconception 1
"P2P works by buying crypto directly from OKX"
Users believed OKX was the counterparty. Learning that a stranger held their fiat during the escrow window caused immediate anxiety and a decision to abandon.
Misconception 2
"Payment must be complex I need to be careful"
Users familiar with Alipay or bank transfer didn't realise they'd use the same tools they already trusted. The unfamiliar UI framing made it feel riskier than it was.
Misconception 3
"P2P is instant why is mine pending?"
Users expected crypto to arrive the moment they paid. 70% of trades complete within 20 minutes but no one told them that. The wait triggered fraud anxiety.
Emotional blocker
"What if the seller doesn't release my crypto?"
Fear of losing money to a bad actor was the single highest-frequency anxiety. Escrow exists precisely to prevent this, but it was invisible in the UI.
1. Rebuild user education so that people understand the concept of P2P
Instead of going straight to the coachmark solution that maybe no one pays attention at highlighting the features, decided to relook into the actual user’s fears when they trade P2P. Designed a full one-page scroll education page to inform users.
Letting users experience it for themselves
Part 2 new flow designs:
Trying for the first time is scary. That’s why we want to provide users a safe educational place for them to try it out by themselves. Learn by experiencing it intuitively. Reference on the left it’s duolingo’s.
Initial concepts that are validated for acceptance with quick user interviews
The process of getting the designs are not one-off. With the different design ideas and variations done mocked up, I also personally tested with new-to-P2P users to ask for their comprehension of P2P after
2 & 3. Countdown & limited options
What is expected from PM: From our findings is that one, user don’t know what the countdown means, and two, users, under preview order section wants to have skip or refresh option if the current match is not ideal.
Current flow below:
What as PD I proposed differently: Relook at the journey to place the “Skip” option even before “Preview order” page. Instead, we can give user the ad matching criteria upfront:
1. Best Price
2. Best Completion Rate
3. Online now
See designed solution below:
Adaption from app to web
Working with devs on specific error permulations
Issue: Insufficent error communication on order limit
Besides designing for zero-to-one flows or new screens, small increments matters as well. In this case, several users overlooked the order limit in the P2P marketplace. They either assumed the minimum amount was the same as the 'Buy' section or failed to notice the grey order limit text until after several attempts.
Buy amount: If user enters USD amount and it falls under order limit, error validation for order limit will show.
Then, if that causes USDT conversion to go beyond the merchant’s available balance, the field border will turn red.
Why solutioning with devs matters: As PDs working with PMs, we might not be aware that there are more than 1,2 scenarios and therefore spending time with dev and content solutioning is the best way to get solutions in this scenario.
Post-performance metrics
1. P2P order completion rate up
The recent roll out for past 1 month we have seen a higher order completion rate to 68.6%. More localisation will be done as next steps once results concluded.
2. Reduced order cancellation for first buys
Takers also are able to comprehend and therefore less order cancellations. Overall Order Cancelation Rate from Order Timeout: decreased from 22.1% to 17.4% (-21.30%).
3. Merchant’s check-ins
We also checked in with our business partners and merchants to ensure that solutions aligns from a dealer’s perspective.
