SafeKeep: a password manager built around how older adults actually share passwords

A simple, secure password manager for people who aren't comfortable with technology, and for the family members who help them.

Role: Sole designer: research, ideation, prototyping and testing
Type: Personal project
Platform: Mobile app
Methods: Surveys, personas, competitive analysis, think-aloud testing

Context

The pandemic pushed many older adults online for groceries, healthcare and banking. Every service needs a password. Writing them in a notebook or reusing one everywhere is convenient but not secure.

Many older parents also rely on their children for help, and those children often live far away. Reading passwords out over the phone or texting them isn't secure either.

Problem statement

How might we make it easy to securely manage passwords, and share them with close family or friends when needed?

What the survey showed

I surveyed older adults and their adult children about how they manage passwords and phone security.

45%record passwords in a physical notebook
54%often need help from someone close with passwords
54%use the same password for every account
73%already use fingerprint or face unlock

Three personas: an older adult new to technology, a busy millennial helping their parents, and a working parent sharing logins with family

Why a phone app

I sketched three possible solutions, Crazy 8s style. I chose a smartphone app because phones are everywhere, they already have fingerprint and face unlock, and sharing between family members is easiest on devices everyone owns.

Popular password managers had three common problems: expensive subscriptions, complicated interfaces full of jargon, and little or no way to share passwords.

Competitive analysis of popular password managers

Low-fidelity sketches of flows and interactions

Usability testing

Testing overturned my own defaults

I tested a low-fidelity Figma prototype with two older adults using the think-aloud method. Things I'd assumed were universal turned out to be anything but.

Sign-up No wizards Testers tapped to jump ahead and asked whether every step was necessary.
Language Jargon got in the way “Master password” and “URL” needed explaining. Testers weren't sure what the website field wanted.
Icons Not everyone understands icons Show/hide password with an eye icon wasn't understood.
Trust Phishing worries One tester asked who verification emails would come from, to be sure they were genuine.
Sharing Whose email is this? Testers confused the invitee's email with their own, and sharing only some passwords needed more explanation.

Usability testing session — A think-aloud testing session

Low-fidelity screens for adding and editing a password — Low-fidelity screens for adding and editing a login

The solution

I refined the design based on what testers told me, then built high-fidelity prototypes in Figma.

Easy to understand

Plain-language directions, no cryptic icons, and larger text throughout.

Share with family

Share a few or all logins with chosen people: healthcare and utilities for older parents, streaming accounts for everyone else.

Secure by default

Suggests strong passwords and uses the phone's own fingerprint or face unlock. Data could be stored encrypted in the user's own iCloud or Google Drive.

Prototype video: sign-up flow

Prototype video: adding a login

Prototype video: editing details

Prototype video: sharing

Retrospective

Familiar to me isn't familiar to everyone

It's easy to assume patterns younger people understand are understood by all. Testing proved otherwise.
Accessibility includes age

Technology should be understandable to older adults and anyone unfamiliar with it, not just usable.
Manage prototype realism

Testers expect every screen to work and every detail to match. A clickable prototype's fidelity has to be planned.

Next project · F5 · NGINX One Designing NGINX One from blank canvas to launch in four months View case study →