When I first envisioned Gamchung.com, it wasn’t part of some grand plan. It began, as many meaningful things do, with a simple conversation with Paljor, one half of the popular Paljor & Palkyi YouTube channel. During our chat about their growing channel, I asked about their YouTube income. His response was surprising: many Tibetan YouTubers struggle to earn from their videos due to copyright strikes. Music owners often claim the ad revenue from their vlogs, leaving creators with nothing. “Can you create copyright-free music for us using those machine learning technologies you always talk about?” he asked.

I casually agreed, but like many good intentions, it initially fell by the wayside.

A Wake-Up Call: Protecting Tibetan Voices

The urgency of this project became clear when I read about China actively banning Tibetan-language content on platforms like Douyin (the Chinese version of TikTok). While creators in Tibet face direct censorship, those in exile struggle with technical barriers and limited resources. This digital silencing of Tibetan voices demanded action.

I couldn’t stop thinking about it. If creators in Tibet are being silenced and creators in exile are discouraged from sharing their stories, what happens to the Tibetan voice in the digital age? That’s when I remembered my promise to Paljor and decided to act.

The First Steps: From One Creator to Many

I dusted off my technical skills and started experimenting with machine learning models from platforms like Hugging Face. Slowly but surely, I began producing royalty-free music that creators could use without fear of copyright claims. When I sent the first batch of tracks to Paljor, he was thrilled. It allowed him to focus on storytelling without worrying about YouTube monetization or copyright strikes.

Then my wife Phuntsok, always the voice of reason, asked me, “Why stop at just Paljor? Why not share this music with all Tibetan content creators?” Her insight was the catalyst that transformed a simple music generation project into what would become Gamchung.com!—not just a resource, but a platform built to empower all Tibetan creators.

From Music Generation to Comprehensive Analytics

What began as a project to generate royalty-free music using Hugging Face models has evolved into Gamchung.com, a comprehensive platform built specifically for Tibetan content creators. Today, we offer:

• Royalty-Free Music: A growing library of copyright-free tracks, custom-made using machine learning
• Advanced Analytics: Live, daily-updated insights powered by the YouTube API
• Graphics Tools: Production enhancers like lower-thirds generators
• Content Strategy: Tailored guidance for Tibetan creators

The Heart of Gamchung: Deep Analytics for Growth

The platform’s core strength lies in its sophisticated analytics engine. Each creator gets access to:

Actionable Channel Insights

• Upload Frequency Analysis: Comparing current output (e.g., 0.08 videos/month) against targets (4 videos/month)
• Viewer Engagement Metrics: Tracking metrics like views per subscriber (229.41) to gauge content quality
• Growth Rate Monitoring: Setting clear monthly growth targets (from 0.73% to 5% monthly)
• Yearly Content Planning: Helping creators scale from current (0.95 videos/year) to target output (48 videos/year)

Comprehensive Performance Metrics

• Views per Video: Detailed breakdowns (e.g., 131,181 views averaging 342 daily views)
• Subscriber Conversion: Tracking new subscribers per video (572) to measure content impact
• View/Subscriber Ratio: Deep engagement metrics (229.41 views per subscriber)
• Growth Rankings: Channel positioning among peers (e.g., #35 among 72 tracked channels)

Content Success Analysis

We analyze each creator’s most successful content. For instance, when a video achieves:

• 742,401 views
• 6,474 likes
• 279 comments
• Optimal length (3 minutes 35 seconds)
• Strategic upload timing (Monday at 12:45 AM NY time)

This data helps creators understand exactly what resonates with their audience.

Advanced Algorithms for Fair Rankings

One of our biggest challenges was creating a fair and meaningful trending section. Initially, we struggled with channels dominating the trending list with multiple videos, which didn’t reflect the diverse creativity in our community. Once again, it was Phuntsok who proposed an elegant solution that became the foundation of our current trending algorithm:

1. Weekly Identification: We identify each channel’s most-viewed video from the past 7 days
2. Single Video Rule: Each channel can only have one video in the trending list
3. Engagement-Based Ranking: Videos are ranked based on their engagement score
4. Diversity Priority: The final list features videos from different channels to ensure diversity

This approach is formalized in our ranking formula ¹:

𝑇𝑆(𝑣,𝑐)=𝛼(Δ𝑉+𝑘₁Δ𝐸)⋅𝐷(𝑐)⋅𝑅(𝑡)

Where:

• Δ𝑉: Weekly view growth
• Δ𝐸: Weighted engagement growth (likes and comments)
• 𝐷(𝑐): Channel diversity factor
• 𝑅(𝑡): Time decay factor

Addressing Unique Challenges

1. Music vs. Vlogs: We’ve implemented separate rankings for music and vlog content, recognizing their different viewing patterns while maintaining fair categorization for hybrid channels.

2. Quality vs. Quantity: Our growth algorithm factors in both raw views and engagement metrics, ensuring channels with genuinely engaged audiences rank appropriately.

3. The Ted Kunchok Solution: To prevent single-creator dominance in rankings, we limit each channel to one trending video per leaderboard, ensuring diverse representation.

Our Commitment to the Community

It’s important to note that Gamchung is not a commercial venture. There’s no financial incentive driving this platform; it’s purely a labor of love and a deep commitment to preserving and amplifying Tibetan voices in the digital age. Currently, all operational costs—from servers to algorithm calculations, from data storage to music generation—are covered entirely by me. I promise to continue supporting it as long as I’m able, because I believe in the mission: Tibetan creators deserve tools that match their talent and ambition.

The platform continues to evolve in its beta stage, driven by creator feedback and our commitment to preserving Tibetan voices. Visit gamchung.com to join our growing community of Tibetan creators. Whether you’re a creator looking to grow your channel or someone passionate about preserving Tibetan culture in the digital age, I’d love to hear your thoughts on making the platform even more valuable for our community.

Together, we can ensure Tibetan voices don’t just survive in the digital age—they thrive. Visit Gamchung.com!

The complete formula in natural language:

𝑇𝑆(𝑣,𝑐) = Trending Score (for a video ‘v’ from channel ‘c’) = 𝛼(Δ𝑉+𝑘₁Δ𝐸) ⋅ 𝐷(𝑐) ⋅ 𝑅(𝑡)

Breaking it down:

1. First part (Δ𝑉+𝑘₁Δ𝐸):

• Takes the video’s view growth over the past week
• Adds it to the engagement growth (likes and comments) multiplied by some importance factor k₁
• This gives us raw popularity

2. Multiply by 𝐷(𝑐):

• This is lower for channels that already have videos in trending
• Helps ensure channel diversity
• Prevents one super popular channel from dominating

3. Multiply by 𝑅(𝑡):

• Reduces the score based on how old the engagement is
• Makes newer activity count more than older activity
• Keeps the trending section fresh

So in very simple terms: The formula takes how fast a video is growing (views + engagement), adjusts it to prevent any single channel from dominating, and gives preference to recent activity.