Travelers · Customer portal

One Portal, One Way Around

A navigation model that could carry policies, claims, billing and help without collapsing — presented, argued for, and carried forward.

A note on what you're seeing. I take my clients' confidentiality seriously. The artifacts on this page are re-creations and abstractions of the real work, and the numbers are illustrative of project outcomes rather than audited figures. Happy to walk you through the genuine article — just ask.

Role: Senior UX Designer (via Wongdoody)
Engagement: Multi-month contract
Tools: Sketch, prototypes, customer research
The win: Navigation and Help proposals adopted for the customer portal

The situation

Insurance portals accrete features until nobody can find anything.

Navigation is invisible when it works and infuriating when it doesn't. Travelers needed a global model that could carry policies, claims, billing and help for millions of customers — and keep carrying them as the portal grew, which portals always do. The failure mode isn't a bad menu. It's twenty good decisions made separately over five years.

Add to that: people don't visit an insurance portal casually. They arrive because something happened. Whatever the navigation does, it has to work for someone who is stressed, in a hurry, and has never used the site before.

My role

I owned the information architecture and interaction design for global navigation across the portal, plus proposals for the transactional and Help surfaces that hang off it — grounding the work in customer research and building the case with stakeholders through iterative proposals.

The process

Beat one

Research-grounded ideation

Customer research in, design exploration out. The research mostly confirmed something uncomfortable: people navigated by search and by guessing, because the menu described the company's org chart rather than their problem.

An exploration board of six small navigation-pattern thumbnails — top bars, sidebars and hybrid layouts — with the fifth circled in rust. — Illustrative Six patterns explored, one chosen. Showing the rejected five is the argument: the decision looks obvious only after you've seen what it beat.

Beat two

The navigation model

Five top-level sections, everything reachable in two moves from anywhere, and Help promoted from the basement to the top bar. That last one was the fight worth having: Help had historically been treated as a failure state rather than a feature.

The global navigation pattern annotated across three breakpoints: an open desktop mega-nav, a condensed tablet list, and a mobile drawer. — Illustrative The same model at three densities. The order never changes across breakpoints — that consistency is what lets someone build a mental map on their phone and still use it on a laptop.

Beat three

Transactional and Help UI

Proposals for the moments users are most stressed: filing, paying, and asking for help mid-task. The rule we settled on was that help never takes the screen away from you — you should be able to read the answer and see the form you were stuck on at the same time.

A contextual help panel sliding over a transactional screen without covering it, with four suggested answers and a Call a human button. — Illustrative Contextual help as a panel, not a page. And a permanent, un-buried route to a person — because sometimes the honest answer to a help query is “you need to talk to someone.”

Beat four

Presenting up

Building the case with stakeholders through iterative proposals. Navigation work is unusually political because every team believes their section belongs in the top bar. The way through is to keep showing the customer's path rather than the company's structure.

Stacked highway interchange signs against a dusk sky, reading Policies, Claims, Billing and Help.

2^*

The outcome

Two moves to anything, from anywhere in the portal. A coherent navigation model and a set of Help experience proposals, presented and carried forward — clarity delivered, rather than a metric I can't verify for you.

Reflection

“Nobody compliments a navigation system. That's how you know it's working.”

The work I'd extend is governance. A navigation model is only as good as the rule that stops the next team from adding a seventh top-level item. Designing the menu is a season's work; designing the process that protects it is the thing that makes it last.

Back to the top of the stack · Stealth AI