Creator Optimization Guide

This guide explains how the ProFans Discovery Feed recommendation system works. By understanding these rules, creators can optimize content strategy and improve visibility, engagement, and long-term growth.

1. Algorithm Overview

ProFans uses a composite scoring system to rank content in the Discovery Feed. Each post receives a final score based on freshness, engagement, user interest, follow relationships, and exposure control.

2. Final Ranking Formula

final_score =
  user_similarity_score
  × base_score
  × followed_boost
  × decay_factor

Higher final scores result in higher priority placement in the Discovery Feed.

3. Base Score Calculation

base_score =
  α × recency_score
  + β × engagement_score

Default weights for recent content:

α = 0.7   (freshness weight)
β = 0.3   (engagement weight)

For content older than 60 days, weights are adjusted:

base_score =
  0.4 × recency_score
  + 0.6 × engagement_score

4. Recency Score (Freshness)

recency_score =
  e^(-0.1 × hours_since_posted)
  × new_content_bonus

New content bonus values:

• Less than 7 days: 1.8
• 7–30 days: 1.5
• 30–60 days: 1.2
• More than 60 days: 1.0

This exponential decay ensures that newer content receives priority while older content fades gradually instead of disappearing suddenly.

5. Engagement Score

total_engagement =
  likes + comments + favorites + views + base_engagement

engagement_raw =
  (likes × 4)
  + (comments × 6)
  + (favorites × 3)
  + (views × 1)
  + (base_engagement × 2)

engagement_score =
  engagement_raw
  ÷ (1 + e^(-0.1 × (total_engagement - 5)))

Comments have the highest impact, followed by likes and favorites. A base engagement value is added to prevent cold-start issues for new content.

6. User Interest Matching

user_similarity_score =
  min(
    0.9,
    max(
      0.1,
      (tag_match_score + content_type_match_score) ÷ 2
    )
  )

Interest matching is calculated using historical user behavior such as views, likes, comments, and favorites.

• Tag interest increases by 0.1 per interaction
• Content type interest increases by 0.05 per interaction

For new or inactive users, the default similarity score is 0.5.

7. Followed Creator Boost

if user_follows_creator:
  followed_boost = 1.8
else:
  followed_boost = 1.0

Content from followed or subscribed creators receives an 80% ranking boost. This makes follower growth a key driver of visibility.

8. Exposure Decay Control

if exposure_count > 8:
  decay_factor = e^(-0.2 × (exposure_count - 8))
else:
  decay_factor = 1.0

A single post will not be shown endlessly to the same user. After 8 impressions, ranking weight gradually decreases to prevent fatigue.

9. Ranking and Diversity Strategy

• Posts are ranked by final score
• Top 70% are selected by score
• Remaining 30% are randomly selected from qualified content
• Posts with similar scores may be randomly reordered
• Initial feed load includes partial randomization

This approach balances quality ranking with content diversity and discovery.

10. Optimization Recommendations

• Post consistently to benefit from freshness scoring
• Encourage comments and meaningful interaction
• Use accurate and focused tags
• Build long-term follower relationships
• Avoid repetitive or duplicate content

Conclusion

The ProFans Discovery Feed algorithm is designed to reward high-quality, relevant, and engaging content. Creators who focus on consistency, interaction, and audience alignment will benefit most over time.