Leading Design for Scalable cross-platform UX for Ads UX Publisher Ecosystem
Impact: Recovered millions in revenue by translating complex data into simple, scalable user experiences. Improved user satisfaction.
Lead Designer • Desktop UX • Data Tables • Systems Design • Participatory design and research • Organisational Complexity
Summary
PROBLEM
Publishers using Google Advertising products (AdSense, AdMob, AdManager) needed a way to understand and responds to limitations on their earnings placed by Google.
The publishers varied in size and format: they ranged from small, independent web publishers to mega-VIP web publishers, as well as app publishers of various sizes. They needed to understand where in their app or website the problems existed, what was their severity, and how to fix them.
Because apps and websites are structured differently, and the scope and type of issues varied, it was a significant design challenge to design one system to scale to all use cases.
CONTRIBUTIONS
-
Led the design of a large, complex cross-platform product involving 25+ people, 4 products, and 4 locations
-
Defined the product vision and scope
-
Collaborated closely with UX, Research, Product Management, and Engineering partners from inception to implementation
-
Organised and led summits and workshops to create alignment among stakeholders
-
Co-wrote and contributed significantly to the Product Requirements Document
-
Co-designed research and participated in research in various methodologies, incl. participatory design
-
Co-wrote and organised the copy for the product
-
Co-wrote and co-produced supporting videos for the product
IMPACT
-
My proactive leadership ensured alignment among stakeholders and timely delivery of a high-quality product
-
Millions in recovered revenue and improved user experience and satisfaction through user-friendly design
-
Future-proofed scalable designs ensured flexibility and adaptability of the product for future use cases
-
The system which I created for managing and tracking the complex nature of the copy for the product was adopted by other teams
-
Table design contributions were adopted by other teams and products
-
"It used to be difficult to find which app had the violation. I can now easily manage.” — A representative Publisher feedback
PRODUCT HIGHLIGHTS
-
It was critical that one system works across all products with minor adjustments to each products' specific design language variant
-
Overall, there were 16+ classes of issues of varied severity, each of which had its own specific resolution flow
-
Both easy and advanced sorting and filtering on the table were designed to spec
-
All designs were accessible and greenlined
PRODUCT SUPPORT VIDEOS
-
I collaborated closely with the video production team on script and art direction of a series of official support videos for the product, ensuring their accuracy in wording as well as graphic representations of the product
-
The series can be watched here
THE CHALLENGE
Enable Google Publishers to understand, find and resolve issues impeding monetisation of their assets.
Process: Collaboration + Research + Design Explorations
COLLABORATION
-
This project spanned 4 time zones and 4 teams, therefore — collaboration was key
-
I believe that focused, in-person collaboration is the best way to solve complex problems and boost team camaraderie. Therefore, I initiated and led in-person summits and workshops rooted in best design thinking and design sprint practices, as well as regular remote working sessions
-
Through collaboration I ensured alignment on requirements, focused scope, and buy-in from all stakeholders
RESEARCH
-
I collaborated closely with UX Researchers to design research plans, conduct interviews, and distill findings for both exploratory and validative research
-
Co-led a participatory design session with UX Researchers
-
Collaborated closely with Engineering and Product Management on pulling metrics to inform designs and areas of exploration
-
Collaborated with support teams on understanding the users needs and frustrations
Co-design session with the publishers. About 15 publishers participated in this research and co-design session.
DESIGN EXPLORATIONS AND EVOLUTION
-
Alongside participatory design sessions, I explored a variety of models for presenting the data.
-
Key directions explored were: dashboard approach, smart summaries guiding users to actions they need to take (anticipating future AI assistance), separation of issues by severity of impact (scope and / or threat to account).
-
Previously, app and web policy issues existed in separate tabs within the product, and different levels of impact (e.g. site, subdomain, page) existed in separate tables as they were coming from different databases. This meant that publishers had to look at multiple sources to find out what to fix. Through research, we discovered that it would be best for publishers to see all issues in one table This posed a significant design and engineering problem to which we found an elegant solution.
-
Ultimately, the publishers wanted everything in one place. They desired a relatively unguided interface that clearly informed them of their losses and threats, and empowered them to decide which issues to address on their own.
FINAL DESIGNS AND IMPACT
-
Publishers were overwhelmingly happy with the new designs, and their clarity led to millions in recovered revenue, as well as a significant reduction in creation of support tickets
-
The final designs were adapted to the specific design language requirements of the particular products
-
Information architecture was drastically simplified across products, leading to quicker understanding of issues
-
Table designs and component were adopted by other teams
EXAMPLE OF ISSUE RESOLUTION FLOW
EXAMPLE OF BEFORE AND AFTER
EXAMPLE OF FILTERING COMPONENT CONTRIBUTIONS
-
During the design process, we had to both adapt and create new components to serve the desired functionality
-
It was important that the components created would serve numerous use cases across the ecosystem, as opposed to being one-off "unicorns"
-
Filtering is an example of a component that was minimally adapted to the PC use cases. In particular, the design and content of the filtering criteria was made more robust to aid clarity.
