Hugo Magalhaes

Hugo Magalhães

Product Design

Index

About

Let's collaborate

Index

About

Optimizing moderation logs

Problem

The Trust & Safety team oversees all moderation on the platform, defining policies and guidelines while managing moderators. They use an internal tool to log all moderation actions from human and automated systems, allowing them to assess policy effectiveness and troubleshoot errors. However, the previous tool felt outdated and difficult to navigate. The team identified two major issues: understanding cause-and-effect relationships between entries and the lack of standardization in information from various automated systems.

The Trust & Safety team is responsible for all moderation on the platform. Their role includes defining policies and guidelines, as well as training and overseeing moderators. To understand the various steps in the moderation process, they use an internal tool that logs all moderation actions taken by both human and automated systems. Reviewing these logs enables the team to assess policy effectiveness and troubleshoot potential errors.

Role

Product Designer

SCOPE

Data entry

B2B

TEAM

Trust & Safety
Internal tooling
Product Management
7 people

VIDEO /

Co-design session

VIDEO /

Co-design session

VIDEO /

Co-design session

Challenge

How can we make the Moderation logs easier to understand?

IMAGE /

Co-design session

IMAGE /

Co-design session

IMAGE /

Co-design session

Design process

Before starting the redesign process, I interviewed each member of the Trust & Safety (T&S) team to gain insights into their workflows and identify pain points with the current tool. These user research sessions helped me understand the various automated systems used for content moderation and the key data points they to extract from the logs.

Solution

Following the interviews, I conducted a series of co-design sessions with the team to revise the data structure, hierarchy, and nomenclature. The outcome of these was a redesigned UI that was easier to read and navigate, making it simpler for the team to extract key data points with minimal effort.