Media Audience Builder Tool

Designing a privacy-first targeting solution - and evolving it through user insight

Main dashboard view of an audience created within the tool

Part 1: Designing the Tool

Context & Challenge

With increasing restrictions on cookies and device IDs, media teams needed a new way to identify and target audiences - one that respected privacy without sacrificing precision.

We set out to build a tool that surfaces audience propensity by geography, using anonymised location indicators and multiple datasets - all without relying on personally identifiable information.

My Role

As the lead UX designer, I was responsible for:

• Scoping and defining the tool’s functionality and user experience

• Leading user research and usability testing

• Collaborating with external developers, product managers, and the ethics team

• Designing a flexible, privacy-compliant interface for audience building

  
  

Approach

We began by mapping out the landscape - both in terms of user needs and upcoming privacy legislation. I ran workshops with media planners and data scientists to understand how audiences were being built, what tools they used, and where friction existed.

Key themes emerged:

• Existing tools were either too rigid or too technical

• Users needed flexibility, but also intuitive guidance - especially given high turnover and varying levels of experience

• Slide-ready audience summaries were in demand to help tell audience stories in client decks

We worked closely with the ethics team to ensure our approach aligned with privacy standards, and I partnered with engineering to prototype and test core components early.

Outcome

The tool was successfully rolled out across the group, enabling users to:

• Build audiences based on geographic propensity

• Combine multiple datasets without relying on IDs

• Navigate a clean, intuitive interface that balanced power and simplicity

It positioned the team ahead of the curve in privacy-first targeting - and laid the foundation for ongoing iteration.

Part 2: Evolving the Expression Builder

The Feature

One of the tool’s core components was the Expression Builder - a UI for constructing complex boolean queries using AND, OR, NOT operators and parentheses. It allowed users to define custom segments based on demographic, interest, and behavioural criteria.

Version 1 of the Expression Builder. with dropdowns within panels indicating their position within the parentheses

Initial Design

We launched with two views:

• Simple View for guided selection

• Advanced View for manual query building

This split was based on early user interviews, which suggested a divide between novice and expert users.

Fullsize version of the V1 expression builder once several variables have been selected

But once the feature was in use, the reality was more nuanced.

Research-Led Pivot

Usability testing revealed that the dual-view approach was causing confusion. Users struggled to understand which view to use - and switching between them disrupted their flow.

Through:

• Observational testing

• Feedback workshops

• Usage analytics

I uncovered a clear need for a more unified, flexible interface.

Redesign Highlights

I scoped and prototyped a new version of the Expression Builder that:

• Merged Simple and Advanced views into a single panel

• Introduced a powerful search field to locate variables quickly

• Allowed confident users to manually edit expressions directly

Improved version of the Expression Builder, integrating Simple and Advanced together along with a single powerful Search field to locate all variables in one place.

We ran prioritisation workshops remotely using dot voting in FigJam, and this redesign emerged as a clear favourite among users.

Some of the dot voting that was carried out over FigJam showing considerable support for the improved feature suggestions here

Collaborative Delivery

I worked closely with the lead developer to refine the component - iterating on expanders, layout, and interaction patterns until we landed on a solution that felt intuitive and robust.

Outcome

The improved Expression Builder gave users:

• Greater control over how they build audiences

• A smoother, more intuitive experience

• The ability to work at their own level of confidence and expertise

It’s now live across the platform, and continues to evolve based on ongoing feedback.

Reflection

This project was a reminder that good design isn’t just about getting it right the first time - it’s about listening, learning, and adapting. The initial tool delivered real value, but it was through continuous research and iteration that we unlocked its full potential.

An in-progress shot of the improved expression builder, now with variables entered and a query being written.