I designed SoftShield from scratch — research, UX, UI and the design system — then used the metrics to turn a distrusted little VPN into a product people pay for and recommend.



The challenge
SoftShield was a tiny VPN with a working product and a stalling business. Trials barely converted, week-one churn was brutal, and the top App Store complaint wasn’t a bug — it was a feeling: “I have no idea if this thing is actually doing anything.”
The category didn’t help. Competitors buried users in servers, protocols and settings; the free-VPN space had trained people to expect shady upsells. My job wasn’t to add features — it was to make protection feel obvious, trustworthy and worth paying for on a small iOS team.
Discovery
Is it even working?
The #1 anxiety. With nothing to confirm protection, users assumed the worst and bounced. A trust gap, not a feature gap.
Too technical for me
Server lists, protocols and settings read as “not for normal people.” Choice and jargon created hesitation at the worst moment.
I’ve been burned before
Years of sketchy free VPNs made any paywall feel like a trap. Trust had to be earned before we asked for money.
Grounded in 12 user interviews, a competitive teardown of 8 VPN apps, and mining of ~2,400 App Store reviews for recurring language.
How I decided
Every screen answers “are you safe right now?” before anything else.
Defaults that just work. Power-user depth hidden until asked for.
No dark patterns. Free is free, premium is clear, exits are obvious.
Key decisions
I shipped the first version, watched the funnels, and redesigned the moments that leaked. Every call below is mine — and each one moved a number.

Off

Live

Leak Test

Servers

Speed Test

Onboarding

Offer

Settings
In a category that sells fear, SoftShield won by selling calm — and proving it.
Impact
Measured over the 8 weeks post-launch against the previous version, via Amplitude funnels and App Store Connect. Trust tools (speed & leak test) were used by ~38% of active users and correlated with higher conversion.
Reflection
It threatened the one-tap simplicity that was the whole bet. I shipped it in v1.2 behind an “Advanced” shelf, once the core had earned the right to add depth.
The low intro price lifted trials, yet pulled lifetime value down. Next time I’d run price-elasticity tests instead of defaulting to the cheapest offer.
I’d add a short in-product trust survey to measure the feeling directly, rather than inferring it from reviews weeks later.
Design system
Palette
Typography — Nunito, rounded & warm
Key components
The hero control — reads on/off in an instant.
Flag, place and tier in a calm white card.
Three jobs — protect, measure, verify.