Publicis PeopleCloud
UX / UI / Research
Publicis PeopleCloud is an integrated cloud-based platform that aggregates various data sources in order to identify sources of client growth at an ID level.
Problem
The initial PeopleCloud platform that I inherited had several issues. The main being the reliance of user-inputted data. Instead of providing an automated system, users were expected to find and manually enter their information. This was both tedious and unreliable. Besides this issue, stakeholders were guiding the product development without doing any sort of user research which made PeopleCloud focus on problems other than the users’.
The process of creating such a complex platform comes with a lot of complicated problems. Acquiring, integrating and permissioning data sets was essential to the success of the product. All of this had to be completed while still maintaining and adhering to GDPR compliances.
Role & stakeholders
I assisted in various stages of the design process - from start to release to new iterations. I frequently worked with members from UX Research, Product Management, Data Science, Data Engineering, Product Engineering and QA.
Process
Generative research
3x Google Venture “Design Sprints” (NYC, Chicago, London)
Persona Building
User Flows
Whiteboarding
Prototyping
Concept Testing
High fidelity design
Implementation
Usability Testing
Noteworthy features
Feature importance
When expanding an audience using a modeling program, there tends to be a lot done behind the scenes. The Feature Importance tool enables the user to see which items are the most important.
Clustering
The clustering module was created from a data science exploratory project. It is an easy way to identify the natural groupings of a user’s audience.
Insights and Demographics
Aggregated data enables various different demographics to be visualized, filtered and interacted with.
Comparisons
Based on user feedback, the comparisons feature was created. There was a need to have a more flexible way of comparing segments to one another.
Reporting
In order to ensure that PeopleCloud’s ID counts and data sources are healthy, a reporting screen was vital.
Interests Indexing
Besides knowing demographics, it is far more interesting to see what an audience is interested in to cater creative and media buying to accurately target audiences.
Measuring success
+ 62% user adoption in first 6 months
Reduction of data load times from 5s to 0.25s
+14% on satisfaction survey post-launch (from 72 to 86)