annjalina.k

get in touch →

annjalina.k

about

work

approach

contact

get in touch →

← back to work

case study · 04 of 05

research-led enhancement · 2024
1 month · gov

research-led enhancement · 2024 · 1 month · gov

Making Invisible Features
Findable Again.

Making Invisible Features Findable Again.

Confidentiality Note: To respect confidentiality, certain visuals have been adjusted. The overall process, design challenges, and results are based on real work.

+20%

Task Completions

+15 Points

CSAT

client

gov agency · USCIS

role

UX Lead

timeline

1 month

team

3 cross-functional

tools

figma · mural · matomo

project overview

When I joined the team, the notification center had been live for a year. Matomo and SQL queries showed low engagement, but the numbers didn't tell us why. A survey and usability sessions did: the feature had a visibility and relevance problem, not an implementation one.

the challenge

Making notifications more visible without making them more annoying. Push everything forward and you flood the UI with low-value alerts that erode trust. Users already felt the notification center was cluttered and cranking up the volume would only make it worse.

voices from research

"I didn't know about the notification center until now."

survey respondent · usability session

survey respondent,
usability session

key usability findings

The numbers told us something, the users told us why.

Behavioral data from Matomo showed low engagement. A 30-person survey and moderated usability sessions told us exactly where the feature was breaking down.

52%

viewed their notifications through email, not inside the product.

62%

were unaware the notification center even existed.

38%

knew about the notification center and never used it.

70%

didn't find the notification center helpful when they did use it.

50%

felt there were simply too many notifications to sort through.

82%

couldn't access the notification center through email links.

the process · 04 phases

An end-to-end cycle from research to rollout

One month, three people, four phases. We let the research steer, aligned stakeholders early, iterated on design in the open, and shipped only what usability testing backed up.

phase 01

Conduct survey & synthesize data.

Ran a 30-person survey and moderated usability sessions, then synthesized the qualitative and quantitative data side-by-side to find the real breakdown points.

what we did

— Synthesized data by highlighting trends and patterns across Matomo events, survey responses, and session recordings.

— Prioritized recommendations based on impact and technical feasibility before we took anything to stakeholders.

phase 02

Get buy-in.

Walked stakeholders through the findings and proposed scope. In a constrained 1-month timeline, getting alignment before the first pixel saved us from expensive reversals later.

how we aligned

— Met with engineering, product, and CX leads to share findings and proposed recommendations.

— Framed every recommendation around both business impact and user need so no one had to choose between the two.

phase 03

Design and test.

Brainstormed multiple directions, narrowed to the strongest candidates, and validated every choice with moderated usability sessions before anything left the design file.

how we designed

— Brainstormed and iterated on design options in Mural before pulling anything into Figma.

— Conducted moderated usability sessions to validate each design solution against the original research.

phase 04

Implement.

Clean handoff to engineering with annotated specs, post-launch tracking, and a plan to measure whether the problems we set out to solve actually stayed solved.

how we shipped

— Handoff to engineers with annotated specs, behavior documentation, and interaction details.

— Set up Matomo events and SQL queries in advance so we could measure adoption the moment the release went live.

what needed fixing · 05 opportunities

Synthesizing the research revealed five distinct problems and five places to intervene.

opportunity 01

Notification clutter.

The opt-in promised visibility, but delivered an unfiltered flood of notifications users weren't prepared for.

opportunity 02

Communication and feedback.

No clear hierarchy between urgent action-required alerts and purely informational ones; everything looked the same.

opportunity 03

Email disconnect.

Users weren't being brought back into the product through email notifications and the deep links weren't landing where people expected.

opportunity 04

Visibility.

The notification center existed but users weren't finding it. Discoverability was a core problem, not a surface one.

opportunity 05

Engagement.

Users may not have understood what the notification center was for, or why it mattered to their day-to-day work.

IDEATION WORKSHOP

From braindump to clear workflows.

From braindump to
clear workflows.

Userflow · login to notification center

From sign-in to a read notification.

the enhancements · 04 surfaces

Four surfaces, one notification system that actually gets seen.

Each surface received targeted improvements based on research, prioritizing better notifications over more. Clear visibility where it matters, strong hierarchy where it counts, and user control throughout.

fig. 01 · home screen

A prominent notification indicator on the home screen that is discoverable without being disruptive.

A prominent notification indicator on the home screen that is discoverable without
being disruptive.

fig. 02 · notification menu

A restructured menu that separates urgent action-required alerts from informational updates at a glance.

A restructured menu that separates urgent action-required alerts from informational updates
at a glance.

fig. 03 · notification center

User controls for customization, filtering, and muting because trust is built when users feel in control of what reaches them.

User controls for customization, filtering, and muting because trust is built when users feel
in control of what reaches them.

fig. 04 · notification center

User filtering controls that build trust by letting users decide what reaches them.

fig. 05 · email notification

Redesigned emails that deep-link users straight into the relevant product context to close the loop between inbox and action.

Redesigned emails that deep-link users straight into the relevant product context to close the
loop between inbox and action.

outcome and impact

The enhancements worked and the numbers that prove it.

After the rollout we measured adoption through Matomo, a follow-up feedback survey, and SQL queries evaluating whether the notifications feature was being used as intended. The before/after story couldn't be clearer.

+20%

task completion rate

task completion
rate

More users finishing what they started, measured across core workflows.

+40%

notification open rate

notification
open rate

Users now open notifications when they reach them because they expect them to matter.

+45%

opt-in rate

Users opting into notifications signals confidence in what they now receive.

+45%

notification center visits

notification
center visits

Discovery problem solved: users can find and return to the feature on their own.

+15

points · CSAT

User satisfaction lifted across all surveyed cohorts after rollout.

80%

daily reliance

of users now rely on notifications to plan their daily tasks, which has become a core part of their workflow.

key insights · what we learned

Three takeaways that now shape how I approach every research cycle.

Design for value, not discoverability.

Increasing visibility was necessary, but not sufficient on its own. Relevance and timing mattered far more than raw prominence.

User control builds trust.

Giving users the ability to customize, filter, and mute notifications drove both the opt-in lift and the CSAT climb. Control is a feature, not a config option.