# Vidmob Creative Analytics Dashboard > Vidmob is a creative intelligence platform that uses AI to analyze the performance of video and image ads. - Year: 2019 - Disciplines: Product - Role: Lead Product Designer - Designer: Rafael Cespedes (https://rafaelcespedes.com) Vidmob is a creative intelligence platform that uses AI to analyze the performance of video and image ads. It helps brands and agencies understand why creative content performs well or poorly by connecting visual elements (color, pacing, text placement, faces) to ad metrics like click-through rates and conversions. This project was completed using Vidmob’s existing design system and high-level user experience frameworks. ## The Problem Before the dashboard existed, users landed directly into raw data-centric tools that lacked clear next steps. Although the data was powerful, there was no entry point that surfaced why a user should care, or gave them a quick read on what was happening across their account and campaigns. ## My Role As Lead Product Designer, I was responsible for: - Defining the requirements, scope, timeline, information architecture and widget framework - Translating the requirements into low-fi wireframes, prototypes, high-fidelity designs, and spec-sheets - Running and synthesizing user testing sessions - Collaborating with engineering and product management - Driving alignment across product, data science, and client success stakeholders ## Understanding the Users and Media The work started with understanding how clients actually run media. How a business splits into accounts, campaigns, ad sets, and individual creative. Each level carries its own budget, its own optimization goal. Getting that wrong means surfacing data that makes no sense in context. The harder problem was doing this across nine platforms at once. Facebook, Snap, Pinterest, YouTube, LinkedIn, Twitter, AdWords, DV360 — each with its own naming conventions, metric definitions, and attribution logic, none of them built to talk to each other. Pulling that into one view required a normalization layer underneath the dashboard: a common taxonomy that could hold all of them without flattening what made each platform distinct. UX audits across ad platforms and interviews with clients and internal Vidmob teams shaped what data actually mattered and where it needed to live. ## 7 Widgets. Every Platform. No Analyst Required. The core idea: every widget answers a question on the surface, but leads somewhere deeper if you need it to. Nothing is a dead end. Insights connect to Vidmob's production offering. Data connects to action. Everything across Facebook, Instagram, Snapchat, Twitter, Pinterest, YouTube, LinkedIn, AdWords, and DV360 rolls up into a single brand view without losing campaign-level detail, with dedicated Ad Account and Campaign views for those who need to go deeper. The dashboard was also built modular by design, making it easy to test different information architectures and build custom views for different user types. ## Key Feature: Insights Widget Insights are surfaced automatically, drawing from the client's account data and Vidmob's broader dataset. That combination unlocks signals no single account could produce on its own. Analysts can also upload reports directly into the widget, replacing the off-platform back and forth that had become the default workflow. Clicking an insight opens a modal with the full report and two actions: view the report in detail, or hit Create Project to hand off directly to Vidmob's production team with a brief auto-generated from the insight. ## Key Feature: Top Creative Elements Widget The widget uses AI to scan every element in a piece of creative, emotions, objects, people, text, motion, and maps each one against real ad performance data. The result is a ranked view of exactly which creative elements are driving results and which aren't, broken down by KPI and platform. A "Swipe" element might lift CPM on Snapchat but hurt it on Instagram. "Excited" emotion reads well on Facebook and Twitter but underperforms on Snap. That kind of signal is impossible to see inside any single platform's native reporting. Here it's a single table. Clicking on the creative elements allows for a deeper dive. ## Key Feature: Account Comparison Widget Most performance data tells you how you're doing. This tells you how you're doing relative to everyone else. The widget takes the distribution of a client's media, objectives, placements, and media types, and benchmarks it against Vidmob's dataset of advertisers in the same vertical. The result is a fast read on where a brand is over-indexing, on-par, or leaving spend in the wrong places compared to what's actually working across the industry. ## Key Feature: Campaign Snapshot Widget A quick read on how your creative is performing within a campaign, ranked against the account average. The widget surfaces top and bottom performers at a glance. Hover over any thumbnail to preview the creative without leaving the view. Click to go deeper, pulling up the full breakdown of how that specific asset is performing and why. ## Closing The full dashboard for this initiative also included: - Platform Fitness Score: an overall health rating for how well a campaign's creative adheres to platform best practices across duration, branding, messaging, and key visuals. - Performance Norms: benchmarks a KPI against Vidmob's full dataset, showing whether performance is above or below industry or platform averages over a selected time period. - Creative Element Visualizer: maps every element in an ad account as a bubble, sized by relevancy and positioned by KPI impact. Click any bubble to see how that element performs across creatives. ## Takeaways - Data normalization across nine platforms is only half the problem. The other half is educating users on how the numbers are compiled. Without that transparency, the data loses credibility. - A dashboard is only as good as what it links to. Inconsistent UX downstream undercuts the experience upstream. - The Vidmob design system wasn't built for this level of complexity and data vizualisation. Working within those constraints showed typography, color, spacing, and other areas needed massive improvement. - One dashboard for every user type is a compromise. The research pointed clearly toward role-based views, but the infrastructure wasn't there yet. That's the next problem to solve. ## Images and video - https://rafaelcespedes.com/projects/vidmob-dashboard/carousel/01_vidmob_cover_01.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/carousel/02_vidmob_dashboard_00.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/carousel/03_vidmob_widget_03-account-comparison_00.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/carousel/04_vidmob_widget_05-campaign-snapshot_02.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/01_vidmob_cover_05.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/02_vidmob_nike-account-set-up_02.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/03_vidmob_widgets_00.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/04_vidmob_widget_01-insights_00.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/05_vidmob_widget_02-creative-elements_00.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/06_vidmob_widget_03-account-comparison_00.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/07_vidmob_widget_05-campaign-snapshot_00.webp - https://rafaelcespedes.com/projects/vidmob-dashboard/case-study/08_vidmob_dashboard-full_02.webp --- From the portfolio of Rafael Cespedes. Everything about him in one file: https://rafaelcespedes.com/llms-full.txt · Updated 2026-11-08