Next project

Next project

All projects

All projects

Micro Learnings

Micro Learnings

How I scaled bite-sized learning at TikTok to cut repeat violations by 66% and reduce suspensions for striked advertisers by 50%

Team

Product Manager

Product Designer

Instructional Designer

SDE (2)

Timeline

Apr 2024 - Jun 2024

Role

Product Design & Research

Context

Advertisers received fragmented policy education with no way to track engagement or understanding. Training content was hard to scale globally and keep aligned with evolving enforcement rules, highlighting the need for a modular, trackable, and policy-specific learning solution.

The Advertiser Benefits Center segments advertisers by trust profiles (e.g., High-Trust T0, Trusted) to offer tailored integrity treatments like content grace periods and moderation prioritization. However, the fragmented system, dependent on multiple integrity tools for list management and treatment execution, results in delayed program launches, manual overhead, and increased operational complexity.

Impact

Strike volume reduced per advertiser

-66%

Ad account suspension rate drop for striked account

-50%

Eduated advertiser activity

+15X

Micro-learning CSAT

4.7/5.0

Design

Quick Learning Structure

A fast, focused learning experience that helps users understand policies through short videos, supported by clear guardrails. Users are required to fully watch the video before progressing—ensuring key information is seen and retained

Encourage Confident Ad Creation

After successfully completing the learning module, users are encouraged to create new ads—reinforcing their confidence and applying what they've learned in a compliant, informed way.

Short Knowledge Test

A quick, diverse set of questions allows users to assess their understanding of policy content. These tests not only reinforce learning but also enable the business to track success rates and correlate results with key metrics like ad suspension and strike reduction.

Learn Anywhere, Anytime

Responsive design ensures the microlearning experience is accessible across all devices—allowing users to engage with training content anytime, anywhere.

How did we get here?

How did we get here?

How did we get here?

Problem

Advertisers who violate TikTok’s policies currently receive fragmented and untrackable educational experiences—typically limited to a rejection notice and a generic training video. This approach lacks policy-specific, bite-sized learning, provides no visibility into engagement or knowledge retention, and fails to support behavior change. Additionally, creating globally scalable training materials is time-intensive, difficult to maintain amidst evolving enforcement rules, and hindered by a lengthy approval process.

The Advertiser Benefits Center segments advertisers by trust profiles (e.g., High-Trust T0, Trusted) to offer tailored integrity treatments like content grace periods and moderation prioritization. However, the fragmented system, dependent on multiple integrity tools for list management and treatment execution, results in delayed program launches, manual overhead, and increased operational complexity.

Business goals

Business goals

Build trust and reduce repeat violations

Build trust and reduce repeat violations

Build trust and reduce repeat violations

Understand advertiser intent and decrease recidivism through targeted, issue-triggered microlearning experiences.

Ensure scalable, high-quality education with minimal effort from training teams

Ensure scalable, high-quality education with minimal effort from training teams

Ensure scalable, high-quality education with minimal effort from training teams

Use existing public policy content and automation to streamline content creation, while enabling teams to focus only on maintaining quiz content and updating videos as policies evolve

Increase awareness, traceability, and effectiveness of training materials

Increase awareness, traceability, and effectiveness of training materials

Increase awareness, traceability, and effectiveness of training materials

Provide compact, comprehensive, and traceable CTAs (like MGA), improve visibility into training quality, and drive traffic to the TikTok Ad Policy account for ongoing self-education.

Process

Challenges

Challenges

Page built as an iframe

Page built as an iframe

Page built as an iframe

Limited development flexibility and design constraints restricted the ability to create a seamless, responsive, and engaging learning experience. Integration with other platform elements (e.g., tracking, navigation) was also restricted.

Existing materials are ineffective

Existing materials are ineffective

Existing materials are ineffective

Prior training materials were generic, lengthy, and not aligned with specific policy violations. They lacked interactivity, failed to drive behavior change, and provided no way to assess advertiser understanding or completion.

Language localization

Language localization

Language localization

Scaling microlearning globally required accurate and timely translation across multiple languages. Ensuring consistency in tone, meaning, and policy interpretation across regions was resource-intensive and slowed down deployment.

Research

We studied how competitors introduce policy-based violation learnings and analyzed early experimentation from the Make Good Action (MGA) program. Our research focused on:

  • Identifying effective learning mediums that promote understanding and retention

  • Understanding behavioral patterns of good-intent advertisers who unintentionally violate policies

  • Differentiating bad-intent advertiser behavior to tailor interventions appropriately

  • Current operational workflows and processes

  • Program development lifecycle and implementation

  • Critical pain points and improvement opportunities

This research helped shape a targeted and scalable micro-learning solution aligned with advertiser intent and learning preferences.

What did we learn?

Good-intent advertisers show remorse through action

Advertisers with good intent often delete policy-violating ads themselves and proactively reach out to support, signaling their willingness to comply and learn.

Current learning materials lack clarity and accessibility

Existing resources like TikTok Academy and Help Center policy pages are too long, text-heavy, and not actionable—leaving advertisers uncertain about how to create compliant ads.

No clear benefit to completing corrective actions

Advertisers who make changes or engage with learning content receive no tangible reward—such as strike reversal or improved account health—leading to reduced motivation to follow through.

Whats the learning journey in the module?

Micro learning structure

Micro learning structure

In collaboration with the Learnings team, we developed a scalable learning structure that could be applied across all policies and easily support language localization.

Testing

Testing

Used dscout to run unmoderated usability tests, gathering insights on user perception, clarity, and effectiveness of the microlearning experience and policy knowledge assessments

Too many questions in knowledge test

Too many questions in knowledge test

Too many questions in knowledge test

Users found long quizzes overwhelming. Limiting knowledge checks to 3–5 targeted questions improves completion and retention without adding friction.

Accessibility issues in visual design

Accessibility issues in visual design

Accessibility issues in visual design

The "Allowed" and "Not Allowed" sections used background colors to visually distinguish do’s and don’ts. However, the chosen colors lacked sufficient contrast, creating accessibility issues for users with visual impairments.

Users appreciated the concise format

Users appreciated the concise format

Users appreciated the concise format

Participants responded positively to the bite-sized, focused learning modules. The short format made it easier to absorb policy content and encouraged completion without fatigue.

Designs

Designs

View designs

Learnings

Learnings

Cross-functional collaboration led to a more effective learning model

Cross-functional collaboration led to a more effective learning model

Cross-functional collaboration led to a more effective learning model

Working closely with subject matter experts in training and education helped us shift to a more concise, action-oriented learning module. Collaborating with Legal and Policy teams ensured the content accurately reflected enforcement rules while remaining user-friendly and easy to understand.

Challenging the existing solution led to greater value

Challenging the existing solution led to greater value

Challenging the existing solution led to greater value

The previous approach focused on hosting training on a separate platform, pulling users out of the TikTok Ads Manager and offering limited value. By questioning this model and proposing an integrated solution within the platform, we delivered a more seamless, scalable, and trackable experience—one that the team readily embraced due to its improved user impact and lower maintenance overhead.

Designed and built by Shibi with 🥤 🍕 💤

Designed and built by Shibi with 🥤 🍕 💤

Designed and built by Shibi with 🥤 🍕 💤