Decluttering the Camera Roll

Redesigning Apple's iOS 18 Photos app to fix a cluttered layout, unreliable search, and missing onboarding.

PROJECT

Academic

Duration

4 months, 2025

TEAM

Dason Luo (Me)
Rordon Crawley
Justin Foley
Sasha Zecs

ROLE

Lead Designer
UX Researcher
Project Manager

Overview

Background

Apple Photos is the default photo app on iPhone, iPad, and Mac. iOS 18 swapped its tabbed navigation for one continuous scroll — a poorly received that Apple reversed some changes in iOS 18.2.

Background

Apple Photos is the default photo app on iPhone, iPad, and Mac. iOS 18 swapped its tabbed navigation for one continuous scroll — a poorly received that Apple reversed some changes in iOS 18.2.

Problem

The redesign restructured the app's information architecture around algorithmic suggestions instead of the content people need. This change buried their most-accessed photos and collections, and left them to figure out what had changed and how to use the app.

Problem

Solution

We researched why the redesign wasn't working, then rebuilt it around three ideas: bring back tabbed navigation, separate recommended content from user content, and bring structure to search.

Solution

We researched why the redesign wasn't working, then rebuilt it around three ideas: bring back tabbed navigation, separate recommended content from user content, and bring structure to search.

Impact

We tested our prototype with 3 participants. Every task we measured came back easier than the iOS 18 experience, and testers described the result as something Apple could have shipped.

9-10/10

album finding experience across all 3 testers

8–10/10

library navigation experience across all 3 testers

7–8/10

overall navigation experience across all 3 testers

CONSTRAINTS

What We Were Up Against

This project was shaped by what we could change about the product, and who were making the changes.

Apple Intelligence Was Off-Limits

Apple Intelligence features like Clean Up and Genmoji had to stay fully intact, so even when users found the app cluttered, we could only reorganize what existed, never remove a feature outright.

Make it Feel Like Apple

Our redesign had to match Apple's existing visual language — typography, color, interaction conventions. The bar wasn't just fixing the problem; it was fixing it in the way Apple would.

Leading Designers Who Weren't

My teammates weren't design majors and lacked design fundamentals. I owned design decisions, managed the project's direction, and guided my teammates through the design process.

BUSINESS CONTEXT

Why the Redesign in the First Place

Apple had clear goals for this redesign. Our research showed they didn't hold up.

Automation at the Forefront

Apple's WWDC24 announcement said the goal was to help users "easily find and relive special moments" through automatic, themed collections instead of albums. Craig Federighi, Apple's SVP of Software Engineering, connected the redesign to Apple Intelligence, describing both as part of the same push for "personal intelligence."

Reducing the Friction

Billy Sorrentino, Apple's SD of HID, stated: "As our features, users and libraries have grown, so has the density of the Photos app. Rather than hunt and peck throughout, we've created a simple streamlined single view... based on deep intelligence. Ultimately, we wanted to remove friction."

Friction Came Right Back

The single view did the opposite of what Apple promised, burying the content users wanted behind algorithmic suggestions — a frustration echoed in thousands of upvotes on r/ios.

Apple seemed to agree. iOS 18.2 restored swipe-to-go-back in Collections — the same navigation friction our research surfaced — and later updates added back library filtering. While Apple never undid the redesign, they incrementally added back friction-reducers they removed.

RESEARCH

Coming Up Again & Again

We ran a heuristic evaluation, 8 interviews, and 8 usability tests (2 each), spanning high schoolers to adults in their 50s. Interviews revealed user needs; tests showed system issues. Heuristic evaluation backed all four insights; most valuable for what went unspoken.

Cluttered By Default

One long scroll forces users to process algorithmic and personal content together, a violation of the minimalist design and efficiency of use heuristics.

6 participants ignored or actively disliked suggested content (memories, trips, wallpapers), rating its importance as low as 3/10.
4 participants called that same content clutter: "I should only see the stuff I've made, not all this default stuff it creates" another said "this should be its own section. One continuous scroll is too much."
One participant deleted recommended content mid-test, calling it "just really stupid"; another asked to "reorder that layout"; a third moved auto-generated albums to the bottom of the screen.

Recommended and user-created content need to live in separate spaces, with users able to customize what they see, cutting the cognitive load of mixing the two.

Lost Without a Map

Search Without Structure

Too Many Taps

Cluttered By Default

One long scroll forces users to process algorithmic and personal content together, a violation of the minimalist design and efficiency of use heuristics.

6 participants ignored or actively disliked suggested content (memories, trips, wallpapers), rating its importance as low as 3/10.
4 participants called that same content clutter: "I should only see the stuff I've made, not all this default stuff it creates" another said "this should be its own section. One continuous scroll is too much."
One participant deleted recommended content mid-test, calling it "just really stupid"; another asked to "reorder that layout"; a third moved auto-generated albums to the bottom of the screen.

Recommended and user-created content need to live in separate spaces, with users able to customize what they see, cutting the cognitive load of mixing the two.

Lost Without a Map

Search Without Structure

Too Many Taps

Cluttered By Default

One long scroll forces users to process algorithmic and personal content together, a violation of the minimalist design and efficiency of use heuristics.

6 participants ignored or actively disliked suggested content (memories, trips, wallpapers), rating its importance as low as 3/10.
4 participants called that same content clutter: "I should only see the stuff I've made, not all this default stuff it creates" another said "this should be its own section. One continuous scroll is too much."
One participant deleted recommended content mid-test, calling it "just really stupid"; another asked to "reorder that layout"; a third moved auto-generated albums to the bottom of the screen.

Recommended and user-created content need to live in separate spaces, with users able to customize what they see, cutting the cognitive load of mixing the two.

Lost Without a Map

Search Without Structure

Too Many Taps

STRUCTURE

Reintroducing the App's Navigation

This wasn't a new pattern, it's iOS 17's tabbed structure, replaced by iOS 18's single scroll. Users described the app as separate tasks, not one feed, so we restored it and focused effort on the content within.

4 Separate Tabs

We split the single scroll into four spaces: Library, For You, Albums, and Search. For You and Albums use the same components and categories iOS 18 already introduced, Memories, Featured, Media Types, Shared Albums, just split into two separate tabs instead of one mixed feed. This removes the confusion of unrelated content appearing next to their own.

ITERATIONS

Ideas That Didn't Hold Up

We moved from sketches to a tested prototype, revising decisions for Pinned collections and Search.

Casting a Wide Net

I explored layout and interaction directions for each of the four tabs, evaluating their options against research findings before narrowing down.

Sketch

Hi-Fi

Reintroducing Pinned Collections

Pinned collections were buried with other content, making them hard to reach quickly. A tool bar button and a Library tools-menu option both routed back to Albums, but mixed pin interactions with browsing photos.

We landed on a fixed placement beneath the grid instead — the app already opens to the Library by default, so one swipe up displays pinned content. This placement matches Apple's goal of reducing friction through a pattern introduced in their redesign.

Bringing Structure to Search

Participants couldn't find or understand filters, and results came back as one uncategorized list. We first tried a segmented date view like Library's, but it still needed a separate media-type menu and looked too close to that screen. Next we split results into collections and all-photos views, but content stayed mixed and still too much to view at once. We landed on tabbed filters mirroring Music and TV, defaulting to a gallery view organized into sections that open into their own page.

Sketch

Hi-Fi

Bringing Structure to Search

USABILITY TESTING

Putting the Redesign to the Test

We ran three user tests on the prototype, each covering four tasks: customizing For You, searching, finding an album, and pinning a collection. Onboarding wasn't part of this round.

Browsing For You

All testers responded well to the layout, calling it "organized" and "clean, it makes sense." Customizing was equally smooth, with the Edit button "right in front."

Finding an Album

This was the strongest result of any scenario: 2 testers rated their browsing experience 10/10 scores and one 9/10; testers called it "super easy to get to" and "grouped in a way that made sense."

Searching a Photo

Search rated well overall, and navigating to it was simple since "the screens are all at the bottom." One issue came up: preview image size, though pinch-to-zoom was part of the intended design but it's a gesture Figma couldn't simulate.

Pinning a Collection

All three testers struggled to locate and use Pinned Collections. Since Figma can't simulate tap-and-hold, we approximated it with a tap-to-open menu wired to one predefined item — a limitation our instructions never flagged. The prototype also lacked a view/edit page and an overflow-menu option.

Our four scenarios succeeded without a walkthrough, since they leaned on familiar patterns. Pinning was the one truly new interaction, showcasing where onboarding would help most.

FINAL DESIGN

Apple's Language, Reorganized

The current design, followed by what we changed and why.

Current Design

The current layout of Photos.

New Design

Five areas that saw the most significant change from Photos.

For You

Insight 1 — Cluttered By Default

Recommended content lives in its own tab with a visible Edit button, letting users choose the content they want to see instead of scrolling past it.

Albums

Pinned Collections

Search

Onboarding

For You

Insight 1 — Cluttered By Default

Recommended content lives in its own tab with a visible Edit button, letting users choose the content they want to see instead of scrolling past it.

Albums

Pinned Collections

Search

Onboarding

For You

Insight 1 — Cluttered By Default

Recommended content lives in its own tab with a visible Edit button, letting users choose the content they want to see instead of scrolling past it.

Albums

Pinned Collections

Search

Onboarding

REFLECTION

Lessons Moving Forward

The redesign was only one battle. Leading the team through it, and working around the tools we had, taught me just as much.

Back to the Original Goal

Apple wanted to remove friction with a single, intelligence-driven view instead of albums, but our research showed that friction came right back, buried behind algorithmic suggestions. Our redesign meets that same goal: separating recommended from user-created content, restoring navigation, and adding accessible customization.

Leading Without Buy-In

My teammates weren't equally invested, so I ended up leading design and project management. I set expectations, checked in regularly, even asked instructors for advice, but couldn't make people care who didn't want to. In the end, it came down to pushing through and getting the work done without everyone on board.

Prototyping With an Asterisk

Figma couldn't simulate tap-and-hold or pinch-to-zoom, so we tested approximations instead. A tool like ProtoPie or Bolt.new would have let us validate them directly, but the curriculum and project timeline ruled that out.

NEXT STEPS

Where We'd Take It Next

Two things this project didn't have scope to fully address: validating further, and design beyond the phone.

Testing at Scale

A higher-fidelity prototype, tested with a wider group of users, would further confirm whether these solutions hold up beyond three testers. It would also uncover any issues that remain against the actual iOS 18 Photos experience.

Photos Across the Ecosystem

Photos lives across Apple's devices, each with different screen sizes and interaction conventions. Extending this redesign means rethinking how navigation and pinned collections translate to iPad's larger canvas or Mac's windowed interface.