Photoshop Express · Content Systems · Multi-year · Global

Scaling Content Systems at
Photoshop Express

Scaling Content Systems at
Photoshop Express

Scaling Content Systems at
Photoshop Express

How systematic guardrails enabled 2.6× growth without quality drift

How systematic guardrails enabled 2.6× growth without quality drift

5.8M

5.8M

MAU (2021)

→

15M

15M

MAU (2024)

ACT I · THE STARTING POINT

A system that worked —
until it had to scale.

A system that worked —
until it had to scale.

In early 2021, Photoshop Express had 5.8M monthly users and a tight content operation. One vendor, one clear vision, consistent output. Then growth arrived — fast and unrelenting.


To serve a global user base doubling in size, we brought in multiple vendor agencies spanning the US, Ireland, China, and India. Same strategy. Multiple teams. Parallel pipelines operating simultaneously across time zones and creative traditions.

What worked cleanly for one agency, under close supervision, now needed to work reliably for four — with minimal friction and no dilution of the standard we'd built.

Context

Context

Context

For Whom

For Whom

For Whom

This feature targets beauty enthusiasts, stylists, and anyone looking to experiment with different hair colors virtually. From Gen Z users to professional stylists, the feature is designed to cater to a diverse audience with various natural hair colors. It allows users to realistically preview hair colors, which can be shared on social media or used as a reference for real-life changes.

This feature targets beauty enthusiasts, stylists, and anyone looking to experiment with different hair colors virtually. From Gen Z users to professional stylists, the feature is designed to cater to a diverse audience with various natural hair colors. It allows users to realistically preview hair colors, which can be shared on social media or used as a reference for real-life changes.

This feature targets beauty enthusiasts, stylists, and anyone looking to experiment with different hair colors virtually. From Gen Z users to professional stylists, the feature is designed to cater to a diverse audience with various natural hair colors. It allows users to realistically preview hair colors, which can be shared on social media or used as a reference for real-life changes.

Objective

Objective

Objective

The Hair Color Feature in Adobe Photoshop Express allows users to transform their hair color seamlessly while preserving natural texture, light, shadow, and shine. By utilizing advanced technology to detect hair and create precise masks, users can experiment with a wide range of hair colors effortlessly. The feature is equipped with proprietary algorithms for natural hair color identification, ensuring realistic results and inclusivity across all hair types.

The Hair Color Feature in Adobe Photoshop Express allows users to transform their hair color seamlessly while preserving natural texture, light, shadow, and shine. By utilizing advanced technology to detect hair and create precise masks, users can experiment with a wide range of hair colors effortlessly. The feature is equipped with proprietary algorithms for natural hair color identification, ensuring realistic results and inclusivity across all hair types.

The Hair Color Feature in Adobe Photoshop Express allows users to transform their hair color seamlessly while preserving natural texture, light, shadow, and shine. By utilizing advanced technology to detect hair and create precise masks, users can experiment with a wide range of hair colors effortlessly. The feature is equipped with proprietary algorithms for natural hair color identification, ensuring realistic results and inclusivity across all hair types.

Growth didn't break the strategy. It tested whether the strategy
could survive without constant hand-holding.

ACT II · THE DRIFT

Scale doesn't break things
all at once. It drifts.

Scale doesn't break things
all at once. It drifts.

No single failure. No obvious breakage. Just a slow, compounding divergence between what the strategy intended and what was actually shipping.

QUALITY
Visual standards varied across vendors despite shared guidelines. "High quality" meant something slightly different to each team.
Visual standards varied across vendors despite shared guidelines. "High quality" meant something slightly different to each team.
QA
The same design and logic issues surfaced cycle after cycle. Problems weren't being solved — they were being repeated.
The same design and logic issues surfaced cycle after cycle. Problems weren't being solved — they were being repeated.
INTERPRETATION
Category definitions were applied inconsistently. Vendors filled ambiguity with their own assumptions.
VOLUME VS. VALUE
Production velocity increased but impact didn't follow. More content, weaker signal.
RELEVANCE
Trend timing and regional resonance fluctuated — content landed differently depending on where it was made.
QUALITY
Visual standards varied across vendors despite shared guidelines. "High quality" meant something slightly different to each team.
QA
The same design and logic issues surfaced cycle after cycle. Problems weren't being solved — they were being repeated.
INTERPRETATION
Category definitions were applied inconsistently. Vendors filled ambiguity with their own assumptions.
VOLUME VS. VALUE
Production velocity increased but impact didn't follow. More content, weaker signal.
RELEVANCE
Trend timing and regional resonance fluctuated — content landed differently depending on where it was made.

The problem wasn't absence of strategy. It was that strategy alone couldn't enforce itself at scale.

Growth didn't break the strategy. It tested whether the strategy could survive without constant hand-holding.

ACT III · MY ROLE

Not defining the vision.
Making it executable.

Not defining the vision.
Making it executable.

I came in as a Product Designer embedded in the content system — the connective layer between organizational strategy and on-the-ground execution across four global vendors.

WHAT I OWNED
  • Translating strategy into clear, actionable frameworks
  • Building structured feedback loops with each vendor

  • Identifying system-level breakdowns before they compounded

  • Solving product-level problems that affected content quality

  • Ensuring intent survived every handoff

WHO I PARTNERED WITH
  • Content Ops managing the full pipeline
  • Four global vendor agencies in execution

  • Engineering teams for system-level solutions

  • Product & Data for measurement and direction

  • Fellow Designers

ACT IV · THE WORK

Systems built.
Problems solved.

Systems built.
Problems solved.

The work wasn't one big intervention — it was a series of targeted fixes, frameworks, and experiments that compounded into a coherent system.

Operational Systems

Growth didn't break the strategy. It tested whether the strategy could survive without constant hand-holding.

Rather than adding rules, we added clarity. Guardrails replaced ambiguity. Feedback became systematic — pattern-based, not one-off. Recurring QA issues got documented and addressed at the root, not patched release by release.

1

CONTENT GUARDRAILS

Sharp category definitions with visual benchmarks and production-tied do's/don'ts that every vendor could reference and apply.

Consistent output across agencies

2

FEEDBACK AS INFRASTRUCTURE

Regular review loops with pattern-based feedback — not "this is wrong" but "here's why this fails and what good looks like."

Fewer corrections, faster cycles

3

QA PATTERN ELIMINATION

Recurring issues documented and resolved at the system level — turned repeated problems into permanent guardrails.

Repeat failures dropped

Localised Content for JAPAC & APAC

Growth didn't break the strategy. It tested whether the strategy could survive without constant hand-holding.

Global frameworks aren't enough for markets as distinct as Japan and Southeast Asia. We built dedicated content streams for JAPAC and APAC — culturally calibrated, not just translated.

Japan-First Content Direction

  • Researched regional aesthetic preferences — minimalism, seasonal sensitivity, typographic culture

  • Developed sticker and overlay styles rooted in local visual language

  • Seasonal moments aligned to Japanese calendar (Sakura, Obon, New Year)

  • Vendor briefs specifically calibrated to avoid Western-default interpretations

  • Validated output against actual JAPAC user behaviour signals

APAC Regional Resonance

  • Identified content gaps where global library under-served APAC users

  • Introduced festival and cultural moment coverage: Diwali, Lunar New Year, Eid

  • Colour and style preferences distinct from Western defaults — richer, more vibrant palettes

  • Worked with regional agency (India-based) for culturally informed execution

  • Created frameworks preventing over-indexing or stereotyping regional aesthetics

Localisation isn't translation — it's understanding what feels native, and building systems that can deliver that at scale.

Growth didn't break the strategy. It tested whether the strategy could survive without constant hand-holding.

Product-Level Problem Solving

Growth didn't break the strategy. It tested whether the strategy could survive without constant hand-holding.

Content problems often live at the product level. I didn't stop at visual direction — I went upstream to fix how content behaved inside the app itself.

⚙️

Automated Blend Mode System

Every overlay in Photoshop Express has a blend mode — the method by which it composites with the photo beneath. The right blend mode is different for each overlay type: a light texture needs Screen, a dark graphic may need Multiply, a colour wash works best on Soft Light.


Manually assigning a default blend mode for each overlay across hundreds of assets was unsustainable and error-prone. I researched the relationship between overlay visual characteristics and optimal blend mode, identified clear patterns, and used those findings to define a logic-based ruleset.


Engineers built an automated system from my research — overlays now get their optimal default blend mode assigned programmatically, eliminating manual effort and improving the out-of-box experience for

every user.

🌸

Japan-First Content Direction

When vendor timelines or budget constraints made external production impractical for time-sensitive seasonal moments, we created content in-house.


Executed the design of seasonal sticker packs, overlays, and themed assets — ensuring they met the same quality bar as vendor content, shipped on time, and felt native to the platform's visual language.


This gave us a responsive, flexible production capability that didn't depend on external agency cycles — critical for culturally specific moments where timing was everything.

🎨

Photo Filter R&D — Trends + Experimentation

Filters are one of the most-used features in Photoshop Express. I worked on developing new photo filter directions aligned to current visual trends — analysing what was gaining traction on social platforms and translating those into filter aesthetics that felt fresh.


Beyond global trends, I experimented with differential filter application — testing how the same filter responds differently when applied to backgrounds versus subjects within a photo. This subject-aware approach created more sophisticated, photography-aware results rather than flat, global adjustments.


These experiments shaped both the filter library expansion and informed future product thinking around intelligent, context-aware filters.

Content that doesn't see its users doesn't keep them. Inclusivity and cultural fluency weren't values we added on top — they were built into how we evaluated everything.

Growth didn't break the strategy. It tested whether the strategy could survive without constant hand-holding.

ACT IV · THE IMPACT

One system. Multiple vendors.
One voice.

One system. Multiple vendors.
One voice.

Product growth doesn't happen in isolation. The content system scaled effectively — and the numbers reflect that.

15M

15M

MAU by mid-2024
(from 5.8M in 2021)

+9.2M

+9.2M

Absolute user growth
over lifecycle

2.6×

Relative growth
(≈158.6%)

+36%

+36%

YoY growth
FY24 (reported)

Curated 6000+ content pieces

across multiple categories to enhance the creative editing experience in Photoshop Express. This included:

Editor Themes

Curated creative themes to enhance photo edits with unique aesthetic styles.

Artistic Themes

Created 10-15 artistic editor themes using free content, later expanding to 20 global themes. Achieved a 60% lift in theme usage throughput within the first week.

Overlays

Overlays

Overlays

Crafted texture and effect overlays to add depth and artistic flair to images.

Collage Themes

Curated cohesive layouts and themes for visually appealing multi-photo collages.

Stickers and Animated Stickers

Curated expressive static and animated stickers to personalize and enhance edits.

Looks & LC Looks

Developed layered color looks that style backgrounds and subjects separately for a refined effect.

Given the confidential nature of this project,
I can't share detailed information publicly.

For insights into design decisions, processes, and key contributions,
feel free to contact me.

Given the confidential nature of this project,
I can't share detailed information publicly.

For insights into design decisions, processes, and key contributions,
feel free to contact me.

© Framer Inc. 2023

Create a free website with Framer, the website builder loved by startups, designers and agencies.