All projects

Vodafone Secure Net: Making the Invisible Feel Powerful

Most security happens silently. This redesign made protection visible, actionable, and trustworthy.

Role

Senior UX/UI Designer (sole designer on feature)

Platform

Android

Timeline

4 months (research to rollout)

Reach

~100,000 users / 4 markets

Result

Drop-off 60% → 20%, NPS +1.5

The Problem No One Talked About

Security was working — but users couldn’t feel it

Most of what Secure Net does, it does silently. Threats are blocked in the background. Suspicious sites are filtered before users notice them.

That is great for security, but bad for perceived value. Users could not feel the product working, and when you cannot feel something, you stop trusting it.

The Cleaning Tool was the exception: scan device, find threats, remove them, see result. But almost nobody finished the flow.

Drop-off rate: 60%.

Secure Net Android dashboard showing Cleaning Tool among security modules. — The Tool sits inside the Secure Net dashboard — one of the few surfaces where users expect visible motion after they tap.

What We Found

Not confusion — unmet expectation

I ran user testing in Turkey before changing the interface. Users started curious, then became uncertain and left.

The flow stacked orientation, virus-definition preparation, scan, and removal without telling people how long each beat would last. Timing swung with backend latency, device tier, and storage volume — so even reliable engineering felt arbitrary.

A six-step onboarding sequence still greeted people before the scan began — useful for education, expensive for momentum when someone only wanted to clean their phone.

Cleaning Tool introductory screen — step 1 of 6 before scanning. — Orientation step 1 of 6: users absorbed security storytelling while their urgency drained.

The Design Decisions

Three coordinated decisions across UX, UI, and system behavior

1. Collapse the flow

I removed the database update from the user path and moved it to a silent background pre-step. The experience became a clean two-step flow: Scan → Remove.

2. Set expectations visually

I redesigned screens to communicate progress clearly: active processing, threat detection, and clean-device confirmation became visually distinct states, not just text changes.

3. Improve backend communication

In parallel, I worked with engineering to optimize backend interaction, reducing real wait times and making the tool feel genuinely more responsive.

Execution screens split “work in flight”, “what we found”, and “what happened next” into distinct visual chapters instead of recycled layouts.

Cleaning Tool screen — active cleaning with progress copy. — Removal phase with explicit progress language — users finally saw the machine working for them.

Scan results listing infections with Clean selected items control. — Scan results as an actionable ledger — threats named, selection surfaced, primary action anchored.

Cleaning Tool success — All Infections cleaned celebration state. — Completion state with celebration micro-motion — closure became memorable, not muted.

Testing and Iteration

Tested, narrowed, iterated, and revalidated

I tested three design variations with users. One clearly outperformed the others, but moderated sessions exposed friction points that metrics alone would not reveal.

Mid-scan states were especially vocal: percentage motion helped, yet headline hierarchy, reassurance copy, and cancellation paths decided whether someone trusted the wait.

Cleaning Tool scan in progress at fifty percent. — Representative mid-scan frame used in sessions — anchoring pace expectations before polishing removal.

I iterated and retested until those points were resolved. Four months after the first research session in Turkey, the redesign was live.

Rollout and Results

Consistent performance across markets

The feature launched first in Turkey (with baseline data), then in UK, Spain, and Italy where Cleaning Tool had not been available before.

Turkey (Q1 → Q3 2025): Drop-off 60% → 20% (–67%), Android app NPS 5.0 → 6.5.

UK, Spain, Italy: new market baseline with post-launch drop-off around 20%, consistent with Turkey’s improved performance.

Total reach: approximately 100,000 users across 4 markets.

Drop-off rate

60% → 20%

Android app NPS

5.0 → 6.5

What I Learned

Users judge reliability through communication

Security products often prioritize reliability over communication, but users experience both at the same time.

A flow can be technically reliable and still be abandoned if it feels broken. The biggest win here was not only simplifying steps — it was designing what users needed to feel, not just what they needed to do.