Chat is how most new customers first reach an advisor on Keen. Yet when the $1 trial ended, most of them walked away, on every platform.
58%
trial-to-conversion drop-off across web and mobile app
What did I do
Cut cognitive load at the exact moment the trial ends
01
Found the real problem
The flow treated people like committed buyers while they were still deciding whether to trust the service.
02
Worked layer by layer
First what users needed (confidence, not a hard sell), then which friction to fix first, then the UI.
03
Designed inside hard limits
The timer, backend and pricing were off limits. So: keep users in chat, cut copy, show fewer options, pre-select the lowest amount, enlarge tap targets.
The Result
+9.5%
mobile app deposits
At a glance
Before
High friction at the critical moment
Still building trust, but treated like buyers
A full-screen modal pulled them out of chat
4 dense payment options in 60 seconds
Tap targets too small for thumbs
After
More confident decisions
A bottom sheet keeps the chat in view
3 options with a clear hierarchy
Lowest amount pre-selected
48px tap targets
The button says what you get: “Add 10 more minutes”
02
Context · Research
Paying a psychic online stacks trust barrier on trust barrier
Three trust barriers
Psychic skepticism
Many users arrive doubting psychic readings are legit.
Payment vulnerability
Asking for money amplifies that doubt.
Time pressure
A 60-second window adds urgency, not confidence.
Business model
Turn “5 minutes for $1” trials into paying customers that justify acquisition costs.
The paradox
We ask for money at the exact moment trust is still forming.
The research
The trial’s last minute is the most critical point in the funnel
Why chat mattered
Lower barrier
Chat feels less committal than a phone call.
Privacy and control
Users go at their own pace, no voice required.
Mobile-first
Most people discover Keen on their phone.
What we found
Trend
Mobile converted about 30% lower than web in every cohort, regardless of advisor or time of day.
Mental model
The design assumed committed buyers who only needed payment options.
The analysis
Three exit points killed the conversion
Exit point 1
Context disruption
A full-screen modal pulled users out of chat, away from their advisor, at the moment of decision.
Exit point 2
Cognitive overload
4 look-alike options, dense copy and a disclaimer, all under a 60-second clock.
Exit point 3
Decision paralysis
No default and no clear best value, so the timer often ran out.
The cascade
Each problem fed the next: disruption worsened overload, overload caused paralysis, paralysis ended in timeouts.
03
Framing · Approach
A strategy map from design to dollars
How design drives revenue
Identify
Credibility at the payment moment
Usability under a 60-second clock
Clarity on what it costs
Which enables
More trials converting to paid
Lower acquisition cost per customer
And drives
More annual revenue from mobile
Higher lifetime value from better first impressions
In sum
More confidence and less cognitive load at payment means more trials convert, which means more revenue and a stronger market position.
The root cause
The payment flow spiked cognitive load at the exact moment trust was forming.
What drove the overload
60-second window
Urgency without confidence
4 dense options
More than working memory can hold
Full-screen takeover
Broke the connection with the advisor
Required disclaimer
One more thing to read
Design challenge
How might we streamline the UI so customers can choose with confidence before time runs out?
The constraints
Balancing 3 forces
Business
Same pricing and payment infrastructure
60 seconds, no more (legal requirement)
A measurable lift within 3 months
Technical
No backend or pricing-logic changes
Existing A/B testing framework
iOS, Android and web built differently
User
Clarity under time pressure
Paying a psychic is high-friction by nature
Thumb-zone ergonomics; 80% prefer chat
Key direction
The pain point matrix pointed to mobile-native patterns (easy, high impact) over backend changes (hard, high impact).
What I changed
01 · Decision paralysis
→ Pre-select the lowest amount
02 · Context switching
→ Bottom sheet over the chat
03 · Visual clutter
→ Streamlined UI, less copy
04 · Tiny tap targets
→ 48px minimum
Visual decisions
Minimal noise
Cut anything that didn’t help the decision
Type hierarchy
Price, time and action read differently
Brand color on the CTA only
One place for the eye to land
Subtle elevation
A light shadow lifts the sheet over the chat
The UX principles
Heuristic #1 · Visibility of system status
The timer stays visible, so the decision window is always clear.
Heuristic #6 · Recognition over recall
The button shows the minutes added, so the value is instant.
The prototype
04
Validation · Results
Same design, two outcomes
The hypothesis
Users weren’t against paying. They were overwhelmed by the ask at a vulnerable moment.
The test
Control (A)
Original payment flow
Variant (B)
Optimized for cognitive load
Platforms
Mobile app + responsive web
Key metric
Deposit conversion rate
Locked by the test framework
The 1-minute window
The number of options
Backend price logic
Then the scope changed
The VP of Product expanded the test to every platform. I’d designed for that, so the changes scaled with small tweaks.
The results
Responsive web
No impact
More noncommittal visitors from lower-intent channels. Commitment, not friction, was the barrier.
Mobile app
+9.5%
Downloading an app signals intent, especially for a niche product. Friction was the barrier.
The insight
Reducing cognitive load pays off most when users already have intent.
05
Reflections
What I’d carry forward
Strategy first
Framing it as a business problem, a mobile-web conversion gap, earned buy-in for iteration and a bigger scope.
Systems over features
Diagnosing the mismatch between where users were (forming trust) and how the system treated them (buyers) revealed the real leverage.
