All case studies

If app Case 2 of 4

Creation, improvements and redesign of If self service apps

Policies, claims and invoices in native apps for different countries. I designed them end to end: new features, validation, data and feedback analysis, and a branded native redesign on a shared component library.

Role
Lead Product Designer
Team
POs, business developers, iOS and Android engineers, QA, analysts and a scrum master
Platform
iOS (SwiftUI) and Android (Material Design)
The redesigned If app: home, claims and travel insurance screens

In short

The If app does what our web self service does, faster and native: policies, claims and invoices.

When I took it over, it worked, but every release made it look a little older. I designed new features, validated them with users, used data and feedback to decide what to fix first, and redesigned it as a branded native app in SwiftUI and Material Design.

Before and after

Same tasks, a new app: a structure built around how people look for things, and one colour and component system on iOS and Android.

The old app on the left, the redesign on the right.

Goal

An app that feels native on both platforms and handles the common tasks without the website.

Success metrics

  • App Store and Google Play ratings
  • Monthly active users
  • Completion of key tasks
  • Claims reported in the app
  • Adoption of new features
  • Support contacts about the app

My role

I owned the app’s design end to end, working with POs, business developers, engineers, QA, analysts and a scrum master across the Nordics and the Baltics.

Discovery

Customer feedback and app data showed what to fix first. Interviews, user flows and card sorting showed how people actually look for things.

Research board with user flows, interview notes and card sorting results
Flows, interviews and card sorting.

Workshops

The major flows were worked through in the open, with specialists from across the company.

Onboarding and login flows in light and dark mode, with comments from specialists
Onboarding and login, in light and dark mode.

Decisions that shaped it

Fix and redesign at the same time

Customers didn’t wait a year for fixes, and every fix taught us something the redesign could use.

Structure before screens

The app was reorganised around how people look for things. The common tasks are one tap away: insurances, claims, payments and support.

Validate first, then measure

New features were tested as prototypes before development, and app data and feedback after release showed what to improve next.

Treat sensitive changes as sensitive

Changing contact details is a classic account takeover route. Every change is verified, with a re-check after six months: quick for the owner, hard for anyone else.

Flow chart for changing contact details
Changing contact details: the logic first.
User flow for changing contact details, with a re-check after six months
Then the screens, with a re-check after six months.

Design reviews

Every flow went through reviews with the developers who would build it.

Design review board with the development team
Design review with the development team.
Design review of the contact detail change and verification flow
The contact detail flow, in review.

One system for every app

Colour variables and a component library in Figma, shared by designers and developers, built natively in SwiftUI and Material Design with the If brand on both.

There isn’t one If app: each country has its own, with the features its business and customers need. They share one design, so every new feature had to fit the system and still make sense where it was needed.

Colour variables with light and dark values
Colour variables, light and dark.
The If app component library in Figma, with components marked ready for development
Components with their states, sizes and dark mode, ready for development.
App Store screenshots of three If apps, five for each
Three apps, one design.

Looking back

The best investment was the system: every later feature was faster to design and build, in every country’s app. But it was only a means. The real test was whether someone could check their cover, report a claim or pay an invoice in a few taps, often on a bad day.