From drop-off data to activation design

Role: Staff Product Designer (sole design lead)
Company: Plum
Time: 2024–2025
Team: 1 PM · 4 Engineers · 1 EM · 1 Staff Designer

Plum needed to turn sign-ups into savers — redesigning the first experience for 2M+ users across 10 markets.

Featured talk: Presenting learnings from this work at UX London 2026

Plum onboarding — welcome screen and purpose selector

Context

Regulatory complexity

Why the personal details step was broken

Plum was growing fast — 2M+ users across 10 markets, £6.7M revenue in 2024 — but the onboarding funnel had a problem hiding in plain sight.

Personal details carried KYC requirements for 10+ products across UK and EU — each with different data demands. One form was trying to serve every regulatory regime at once.

UK and EU product onboarding requirements — KYC field matrix showing different data demands per product

Business scale

£6.7M revenue

Nearly doubled YoY across 10 markets. Every point of onboarding conversion directly impacted growth.

Funnel health

68% completion

Personal details had a 42% error rate — the single biggest drop-off point in the entire onboarding funnel.

User research

480 respondents

Surveyed across 3 channels. 17% cited unwanted direct debit, 16% cited complexity. Qual matched quant.

Business impact

40% non-depositors

Blocked at account connection. Solving this represented 6,900–14,300 incremental conversions per month.

Diagnosis

Data findings

Funnel performance per screen

Data findings

Funnel performance over time

Mapping the experience

Three problems hiding in one funnel

The data pointed to a single broken step, but the journey map revealed three distinct failure modes converging at the same point:

Account connection failures — ~40% of non-depositors blocked here. Bank not supported, connection failing, or user hesitant to link.
No control over deposit amount — 17% of detractors cited unwanted direct debits.
Commitment asked too early — users giving personal details before choosing the product they came for.

Customer journey map showing user emotional states across onboarding steps

Emotional state mapped against each onboarding step, from sign-up to bank linking.

Immediate fixes

Personal details

Consolidated fields, improved scan-ability with headlines, better input affordance and information grouping.

Purpose selector

Moved earlier so users choose their product before committing. Clearer feature descriptions with card selectors.

Skippable bank linking

Removed bank linking as mandatory. Users who skip are guided through the home screen to complete at their own pace.

Bank linking — before and after showing skippable flow

Completion

68% → 88%

Personal details completion over a 6-month window (Looker)

Error rate

42% → 23%

Error rate on personal details nearly halved

Revenue impact

+40% per customer

Revenue per customer growth, contributing to £6.7M in 2024

Scale

10 markets

End-to-end flow shipped across iOS & Android in all markets

Strategic reframe

From drop-offs to intent

Separating onboarding from activation

The tactical fixes revealed a deeper pattern: users arrived with different goals and trust levels, yet the flow treated everyone the same. We separated the journey into two design problems — onboarding (identity, compliance, setup) and activation (reaching first value) — giving the team a shared language and each metric its own levers.

Onboarding screen sequence — from welcome through personal details and purpose selection to bank linking decision point

Activation design

Redesigning the home screen

Where onboarding ends and engagement begins

The home screen is every user's first landing after signup. It has to welcome, orient, and nudge toward the next meaningful action — but without modularity, there was no way to flex the experience by intent or trust level, and engagement was dropping off.

We restructured it around an inverted hierarchy of needs: see your money first, then discover more. It also needed to serve two trust states — users who linked their bank needed reinforcement, while those who skipped needed to build confidence from scratch.

Home screen analysis — hierarchy of needs, trust states, and information architecture

Before

6 competing sections, ~20 journey prompts, no hierarchy. Users scanned the full screen just to find their balance.

After

Modular layout responding to user intent and trust state: balance always visible, actions contextual to segment, progressive disclosure replacing the wall of prompts.

Result

From sign-up to saver

The redesigned flow went live for 2M+ users on iOS and Android. By separating onboarding from activation and designing for intent, we extended the framework into financial product onboarding — the right product, at the right moment. Completion lifted 20 points, error rates halved, and the team had a durable model for future work.

Plum — shipped onboarding and activation screens

Key learning

Design for intent, not just completion

Separating onboarding from activation gave the team two measurable problems instead of one vague funnel — and unlocked a 20-point completion lift.

Approach

Start where the data is loudest

The strategic reframe didn't come from a workshop — it came from fixing a broken form and noticing the pattern underneath. Tactical work is strategic work when you're paying attention.