Rebuilding Capital One's Mobile Navigation from the Ground Up

A full IA overhaul that reduced navigation depth, improved task success rates, and set a new standard across three product lines.

Client
Capital One
Role
UX Lead & Information Architect
Platform
iOS, Android, Web
Deliverables
IA Audit, Navigation Redesign, Usability Testing
91%
Task Success Rate
↑ from 66%
2
Navigation Levels
↓ from 4
28%
Reduction in Onboarding Drop-off
First 30 days
3
Product Lines Adopted New IA
Post-launch

Six Years of Feature Creep

Capital One's mobile app had grown organically over six years. Each new product launch added features without a corresponding rethink of the navigation structure. The result was a four-level deep hierarchy with inconsistent labeling, redundant pathways, and a 34% task failure rate on core banking actions.

Users couldn't find what they needed. Internal analytics showed that the search function — a last resort — was being used for tasks that should have been one tap away. Customer support call volume for "how do I find X" had increased 18% year over year.

The business knew something was wrong. What they didn't know was why — or how deep the problem went.

Before: Navigation Audit Map

The existing navigation structure mapped across all entry points revealed 47 distinct pathways to complete 12 core tasks.

Understanding the Gap Between Business Labels and User Mental Models

The first step was a comprehensive IA audit — cataloging every navigation label, entry point, and pathway across iOS, Android, and web. What emerged was a structure built around Capital One's internal org chart, not how customers thought about their money.

IA Audit Spreadsheet / Taxonomy Map

We ran card sorting sessions with 200+ users across customer segments — new account holders, long-term customers, and small business users. The findings were stark: 12 navigation labels meant something fundamentally different to users than they did to the product team.

"Accounts" to users meant all their financial relationships. To the app, it meant only deposit accounts. "Services" was a catch-all that users interpreted as customer support. These weren't small misalignments — they were the root cause of the task failure rate.

"I know my credit card is in here somewhere. I just can never remember which section it's under."
— Usability test participant, 4-year Capital One customer

Rebuilding from User Mental Models

With the audit complete and user research in hand, we rebuilt the taxonomy from scratch — starting with how users described their tasks, not how the business categorized its products.

Step 01

IA Audit

Full catalog of all navigation labels, entry points, and pathways across all platforms. Identified redundancies, dead ends, and inconsistencies.

Step 02

Card Sorting

Open and closed card sorting with 200+ users to understand how they naturally grouped and labeled financial tasks.

Step 03

Tree Testing

Validated the new taxonomy with tree testing before any visual design work began. Iterated through three rounds.

Step 04

Prototype & Test

Built interactive prototypes and ran moderated usability sessions to validate navigation flows against real tasks.

Step 05

Design System Integration

Worked with the design systems team to update navigation components across the shared component library.

Step 06

Phased Rollout

Launched via A/B test to 10% of users, monitored task success and support volume, then rolled out fully.

New Navigation Architecture — Before / After

A New Standard Across the Product

The redesigned navigation reduced depth from four levels to two, eliminated 23 redundant pathways, and relabeled the 12 problem areas identified in research. Task success rate climbed from 66% to 91% in post-launch testing.

Onboarding drop-off in the first 30 days fell by 28%. Customer support contacts related to navigation dropped significantly in the first quarter post-launch.

The new IA was subsequently adopted by three additional product lines — credit cards, auto loans, and small business — using the same taxonomy and component patterns.

Post-Launch Analytics Dashboard