How to Design Effective Content Distribution for Platforms: A LOFTER Case Study

Background

After supporting LOFTER for over six months, the interaction designer identified the platform’s main bottleneck: high‑quality original works are not effectively distributed and rely on users to discover creators manually.

Research Approach

A research study was conducted, analyzing mainstream content platforms (Weibo, Bilibili, Zhihu, Jike, Instagram) through desktop research, competitive analysis, and user surveys to derive a suitable distribution model for LOFTER.

General Content Distribution Model

The model consists of three core elements:

Content : What kind of content deserves distribution.

Channel : Through which methods and pathways the content is delivered.

User : Who consumes the content.

Most products follow this pattern.

Content Organization Structures

Effective distribution starts with a solid content organization architecture. Three common structures are:

Hierarchical : Top‑down classification from broad categories to sub‑categories (e.g., Bilibili channels).

User‑self‑organized : Users collect related content into folders or collections (e.g., Xiaohongshu favorites).

Relational : Linking items across the same level (e.g., Zhihu topics linking to related topics).

Content Production Control

Ensuring new content enters the organized system can be achieved by:

Manual Classification : Creators assign content to predefined categories during submission (e.g., Bilibili’s required tags).

User‑self‑classification : Users tag or folder content themselves.

Machine Recognition : Automated analysis and categorization when manual methods fail.

Content Value Judgment

Two main factors determine whether content is worth distributing:

Timeliness : Critical for news‑driven platforms like Weibo.

Content Hotness : Measured by user interactions (likes, shares, comments) and used by platforms such as Douyin.

Distribution Channels

Valuable content can be delivered through several channel types:

Aggregation : Directly push content from organized pools (e.g., Jike topics, Bilibili channel pages).

Related : Recommend content related to what the user is currently consuming (e.g., Instagram’s Explore page, Weibo related posts).

Hotspot : Highlight top‑valued content (e.g., Bilibili’s 24‑hour hot videos, Weibo hot searches).

User‑Relationship Based Channels

Personalized recommendation builds a user profile from interests and behavior to serve tailored content. Additionally, follower relationships create streams where content from followed creators is shown, often filtered by content value.

LOFTER Case Study

LOFTER has already implemented many aspects of the model:

Creator certification (illustrators, photographers, etc.).

Domain and tag organization on the discovery page.

User‑selected interests during registration.

Follower feed on the homepage.

However, gaps remain:

Only a small fraction of content is automatically managed after certification; most lacks quality control.

Many tags are invisible to users, limiting discoverability.

User profiles are not fully leveraged for personalized distribution.

Follower relationships are one‑way and lack richer social‑graph based distribution.

To address these issues, LOFTER plans a major redesign in early 2019, aiming to improve content organization, enhance tag visibility, refine user profiling, and expand relationship‑based distribution.

Conclusion

Interaction designers should step back from detailed UI work to consider higher‑level product challenges, such as content distribution, to uncover new insights and drive business value.