← Selected Work

Case Study 01TallyMoney

Fixing the funnel for a radical new kind of money

68% onboarding drop-off. Three user cohorts. One redesign that changed how the product acquired customers — built as the sole designer from an MVP-only starting point.

Role
Product Designer (Solo)
Timeline
Jul 2023 – Present
Platform
iOS · Android · Web App
Tools
Figma · Amplitude · GA · Jira
TallyMoney app home screen alongside a fan of gold-backed Mastercard debit cards.
68%
Drop-off reduction from organic traffic
3
A/B test cohorts designed and validated
16
Design team scaled from sole designer
8+
Product flows redesigned end-to-end

The starting point

A product with a radical premise and a broken front door

TallyMoney is a radically different kind of money — gold-backed, with every balance held as physical gold in Swiss vaults. One tally equals one milligram of LBMA-accredited gold.

But the product had a broken front door. 68% of users arriving from organic traffic were dropping out before completing account creation.

I joined as the sole designer — with no design system, no research practice, and no defined process to build on.

The broken door

Three compounding problems, one broken journey

The drop-off wasn't one failure — it cascaded across three stages of the journey, each compounding the last.

01

The comprehension barrier

Organic users landed on a generic GIF carousel — "Supercharge your savings in minutes" and a button — with no education and nothing to build trust.

02

CAPTCHA on screen one

The very first action screen demanded a CAPTCHA and a phone number, before any value or trust had been established.

03

Journey fragmentation

Website → app download → account creation → KYC → submission felt disconnected. T&Cs appeared too early, and there was no sense of progress.

The old flow asked for trust before earning it. Every screen demanded something from the user — data, agreement, patience — before giving them any reason to continue.

Core insight — Nikita Rohan
The old onboarding flow — GIF carousel, CAPTCHA and phone-first screens, with drop-off points annotated in red.

All about the users

Understanding who was arriving, before designing anything

Before designing anything, I needed the qualitative "why" behind the numbers.

I embedded with the customer-facing team and found three distinct audiences arriving at the same broken door.

Cohort A

The Mainstream Saver

Needs the most education before anything else.

I'm not sure what FIAT currency is

Cohort B

The Gold-Curious

Needs education plus social proof.

Inflation is eating my money — but how does gold protect me?

Cohort C

The Gold-Aware

Crypto-native. No education needed — just a fast path.

I'm stacking sats — how does TallyMoney get me money sovereignty?

A fourth group was already ready to sign up — so they needed an "I'm ready, let's open my account" escape hatch straight through the flow.

The Gold-Awakening segmentation journey and the three user cohorts identified in research.
Key redesigned screens: cohort segmentation, "Loved and Trusted" social proof, step indicators, "Welcome on board," and the "Wrong number" recovery modal.

The turning points

Five decisions that changed everything

01

Education before registration

The single biggest lever: teach before asking. Organic users met tailored education that explained gold-backed money before any form appeared.

↑ Directly addressed the 68% drop-off root cause

The education-first screens shown before any registration form.
02

Self-selection as the UX mechanism

Rather than guess what each visitor already knew, the flow let users self-select their starting point — and that choice routed them to the right depth of education.

↑ Validated via A/B test across 3 cohorts

The cohort self-selection screen routing users to the right depth of education.
03

Social proof as a trust bridge

A "Loved and Trusted" screen surfaced press coverage — FT, City AM, Mail Online, GBN — before any commitment was requested.

↑ Trust established before commitment was requested

The "Loved and Trusted" social-proof screen with press coverage.
04

Progress visibility and momentum throughout

Step-dot indicators and reassuring copy — "Welcome on board · Next step: Account opening information · Typically takes 5 minutes" — replaced the old "Bear with us, 48 hours."

↑ Reduced perceived effort and abandonment at final steps

Step-dot progress indicators and the "Welcome on board" screen.
05

Inline error recovery, not dead ends

"Wrong email?" / "Wrong number?" bottom-sheet modals with a "Confirm and resend" action replaced dead-end screens.

↑ Eliminated dead ends that caused flow abandonment

The "Wrong number?" bottom-sheet recovery modal with "Confirm and resend."

The transformation

The journey, transformed

The redesign resequenced the entire journey — trust, then value, then data, then commitment.

Old flow

  • GIF carousel with no explanation
  • CAPTCHA on the first screen
  • Phone → OTP → email → OTP double verification
  • T&Cs presented before any value
  • Marketing consent bundled with T&Cs
  • 4-digit PIN only
  • No progress indicators
  • "Great things come to those who wait" mid-flow
  • "Bear with us, 48 hours" closing screen
  • No inline error recovery

Redesigned flow

  • Gold-Awakening segmentation up front
  • Tailored education per cohort
  • "Loved and Trusted" social proof
  • Email-first, then mobile
  • T&Cs moved to "Welcome on board"
  • Marketing consent separated and optional
  • Full password with real-time validation
  • Face ID prompt
  • Step indicators throughout
  • "Next step: Account opening · Typically 5 minutes"
  • Inline "Wrong email?" / "Wrong number?" recovery
The old onboarding flow versus the redesigned flow, side by side.
Old flow vs. redesigned flow — resequenced to trust, then value, then data, then commitment.

What moved

What the redesign actually moved

68%

Drop-off reduction from organic traffic

Funnel analysis confirmed the lift — driven by education-first onboarding and cohort segmentation.

Data

Data-validated design decisions

A/B testing across three cohorts proved the segmentation approach.

Scale

Design foundation for scale

The work set the pattern for Homepage, Payments, Open Banking, Dormant accounts, Capital Gains, Cards and the Web App — and the design team scaled from 1 to 6.

Security

Improved security UX

A 4-digit PIN gave way to a full password plus Face ID.

Platform

Multi-platform consistency

A shared design system spanned iOS, Android and the Web App, with seamless web-to-mobile continuation.

Looking back

What I'd do differently

The CAPTCHA was a legal and fraud constraint — but I'd push harder for bot prevention that doesn't signal distrust to first-time users.

I'd build the design system from day one rather than retrofitting it once the team grew.

And I'd fight earlier for cohort-level qualitative interviews instead of inferring the "why" second-hand.