What is Alli Creative Insights?
Alli is PMG’s ad tech intelligence platform, aggregating large volumes of cross-channel performance data to help marketers analyze performance and automate complex workflows.
Creative Insights is Alli’s core experience for analyzing creative asset performance, where users move from high-level signals to asset-level optimization.
I was the primary product designer on Creative Insights, owning:
UX research and synthesis
Interaction and information architecture
End-to-end design for video insights
Ongoing collaboration with data science, engineering, and stakeholders
Following the launch of video insights, Creative Insights engagement increased by 78%, and marketers reported a 60% reduction in time-to-insight, measured through 3 usability tests and a post-launch survey of 25 users.
Creative Insights historically focused on static assets, limiting users’ ability to analyze video performance — a growing blind spot as short-form video became central to paid media strategies.
Introducing video data created a new challenge:
The problem was not data availability, but sense making.
User Groups
Creative Insights serves three primary user groups with varying analytical depth, from data-heavy analysts to high-level client stakeholders. This required designing an experience that adapts to different cognitive loads and exploration behaviors within a single system.
Before designing any new surfaces, I needed to understand how users actually move from raw performance data to decisions — and where that process breaks down. The research focused on uncovering how different users interpret video performance today, what signals they trust, and where friction slows insight discovery.
Competitive Analysis
Competitive analysis of video analytics experiences across TikTok, Meta, Pinterest, and LinkedIn..etc.
Contextual Interview
Cross-functional Domain Discovery
Ongoing collaboration with data scientists to understand:
Available video metrics
Standardized representations of video performance
Constraints around aggregation and accuracy
User Journey
This represents a common exploratory workflow synthesized from research across three user types. Users enter and exit at different points depending on intent, but the experience is designed to support fluid movement from signal detection to question formation, data exploration, insight validation, and action.
Real-World Complexity
This was not a full redesign — all work needed to fit within existing Creative Insights architecture
Backend data structures and timelines limited how aggregation could be surfaced
Business goals required shipping within H1 2025
Rather than restructuring the platform, I focused on progressive UX improvements that enhanced insight discovery while remaining scalable.
The existing flow forced users to click into individual ads sequentially, creating friction and cognitive overload.
I redesigned the experience to support movement between levels of abstraction.
Early iterations surfaced all metrics on a single page. Usability testing revealed this overwhelmed users despite logical grouping. Users also wanted to gain more context on the AI insights provided.
In response, I Introduced a tab-based information architecture, separating the core ad performance metrics and video-specific performance metrics, this increased visual hierarchy and screen space for data interpretation.
Designed interactive video graphs that:
Insights surfaced contextually based on user interaction, supporting direct manipulation and reducing the need for separate explanation layers.
Link performance patterns directly to video timestamps
This required close collaboration with data science and frontend engineering to:
Validate how video performance metrics (e.g., completion percentiles) could be meaningfully represented
Ensure interaction fidelity, accessibility, and performance
Summary Metrics Overview
Enabled scan-first insight discovery through summary metrics at the gallery level
Sortable Data Table
Supported pattern detection and anomaly discovery via sortable, filterable table views
Interactive Video Analysis
Designed an interactive video analysis experience that connects performance patterns directly to creative content, enabling faster validation and action.
Increased Weekly Usage by 30%
Increased weekly usage of Creative Insights by 30% after introducing video insights (Pendo)
Marketers reported a 60% reduction in time-to-insight, validated through 3 moderated usability tests and a post-launch survey of 25 users
The exploratory interaction model now serves as a scalable foundation for future video and creative analytics
01 — Designing for Clarity at Scale
This project allowed me to design for decision-making in data-heavy product. The hardest part wasn’t visualizing data, but deciding what information to surface first, what to defer, and how to help users move between levels of abstraction without losing context.
02 — Using Constraints as a Design Input, Not a Limitation
Working within existing architecture and delivery constraints pushed me to treat limitations as design inputs rather than blockers. By focusing on progressive improvements and interaction patterns instead of structural overhauls, I was able to deliver meaningful impact while keeping the solution scalable.