Consulting with a global broadcaster’s engineering team to encourage more user-centric algorithm design & development.
The BBC asked me to join their content recommendations team, made up predominantly of engineers working in remote locations, to encourage a more user-centric culture of product development. By introducing UX practices such as collaborative workshops, user research and persona development, the team started developing algorithms based on user needs and viewing scenarios.
Process
1: Discovery
2: Definition
3: Delivery
Discovery
Understanding the BBC’s overall vision and mission, project goals, usable data, information architecture and other relevant factors.
Map of key data used for recommendations. Audience data fields censored for security & confidentiality
Existing recommendation engine architecture gathered from multiple Miro workshops. BBC data is used to personalise multiple products including iPlayer and the news app
I carried out high level qualitative user research to get a basic understanding of people’s attitudes to the BBC’s current content recommendations system
I also carried out some theoretical research by exploring relevant behavioural psychology research papers
Definition
Hosting collaborative workshops based on best practice UX to define the problem and agree on solutions.
“Mow Might We” problem solving ideation workshop
Encouraging creative, open thinking in a “no idea is a bad idea” type workshop + Introducing the concept of MVP and subsequent iterations
Translating workshop outcomes into user stories, to represent a progress milestone and a point of reference for product development
Delivery
Prioritising stories, grouping similar ones into “epics”, unpacking items into granular tasks and contributing creatively to algorithm development.
User story prioritisation workshop. All stakeholders were invited to rate each story for UX value add and development effort
As features were delivered as code, I created presentations to help explain new functionality to senior BBC stakeholders
Explaining how colour is used to influence recommendations
Explaining how a viewer’s devices and social connections could be used
