← Selected work

BBM · Identity & onboarding

BBID.

BBM had about 120 million monthly active users and four competing ways to sign up. I led a 12-person team that rebuilt identity and onboarding around one phone number and one device, and proved it with research on production code before it rolled out.


Role

Department head and lead designer

Team

12: designers, research, product, architecture, data and engineering

Timeline

3 months, 8 iterations

Platforms

iOS and Android

Results over the following two quarters

−67%

Sign-up churn

−84%

Sign-in churn

+14%

New sign-ups

(01) The problem

A huge base and a fragmented front door

Disconnected

Phone, email, UPIN and Enterprise ID: four ways to identify, all competing for the same entry point.

Foreign on every phone

The look and feel matched neither iOS nor Android.

Built for security first

Onboarding opened with a permission request and kept asking, seven times in all.

Local context

Habits, norms and the DANA wallet on the roadmap all shaped what identity had to do.

The original BBM sign-up and sign-in flows, four paths with numbered permission prompts
The original flows: four ways in, each interrupted by permission prompts (numbered). On a base this size, every extra step was expensive.

(02) How we worked

Hypothesis first, then research on production

Framing

  • What does the data tell us, and what are the business goals?
  • Write the hypothesis first, then test it
  • Follow iOS and Android guidelines, and build a new BBM design system to match

Research

  • Competitive analysis and 2 UX focus groups
  • 2 rounds of usability studies on production code
  • 2 rounds of sanity research with Firebase feature flags on 50% of the live app
Whiteboard sketch of the BBID sign-in logic
Early mapping of the sign-in logic.

(03) Findings and decisions

What the data told us, and what we did about it

What we found

What we decided


About 40% of drop-offs happened on permission screens.

Streamline the flow.


Only 20% signed in with email, and 0.3% with UPIN.

Lead with the phone number. Keep email as the fallback when no SIM card is detected.


33% of users were over 60. 62% used Android, 38% iOS.

A familiar, simple UI with fewer elements that feels native on both platforms.


People change numbers, carry two phones, and carriers recycle numbers.

One phone number and one device are the source of identity, with a check for recycled numbers.

(04) The solution

Clear, smart and native on both platforms

Four BBM welcome screens introducing video chat, stickers and calling
A clear, friendly start: four welcome screens, then straight to the phone number.
The new BBID flow branching into sign-up, password sign-in, one-time-code sign-in and email sign-in
One entry point, four outcomes. The back end recognises the number and routes people to sign-up, password, one-time code or email when no SIM is found. A recycled number triggers an account check.
BBM Me tab with profile, phone number, settings and notification screens
Flexibility after sign-up: change your number, edit your profile and manage settings from one Me tab.

(05) Results

Two quarters after launch

−67%

Sign-up churn

−84%

Sign-in churn

+14%

New sign-ups

35%

Password adoption by end of Q2

8%

2FA adoption by end of Q2

Time in app showed no direct link to this project, and no negative effect over the two quarters.

(06) What I took from it

Three lessons I still use

Fix the front door first

Streamlining sign-up and sign-in moved churn far more than anything else we tried.

Plan for adoption after launch

Password and 2FA ramped slowly. Security features need their own nudges.

Get the foundation right

With identity fixed, DANA’s regulations and security measures were easy to build into the app.

Next case study

Project MOPP

A mortgage pre-approval redesign judged on three business OKRs, and the pivot when one of them didn’t move.

Do you like what you see?

Let’s talk.

Clear direction, trusted leadership, and design that moves the business forward.