From 20+ hypotheses to 5 working formulas: how we built a content machine generating 50 video clips

Testing stages

• Stage 1. Identifying baseline content formats in the logistics niche
  We started by testing basic content formats to understand what could actually work in logistics. We experimented with memes, parking footage, accident-related content, posting timing, and drone shots. Even at this early stage, performance varied significantly from 100 to 26K views. Key takeaway: memes delivered stable reach, while accident-related content was not suitable due to platform restrictions.
• Stage 2. Shifting from formats to emotional audience triggers
  We moved from format-based testing to analyzing emotional triggers. We tested themes like income, fear, disputes, immigration, and personal stories. It became clear that triggers — not production quality — drive reach and engagement. The strongest performance came from income and immigration-related topics, reaching up to 26K views.
• Stage 3. Platform-specific strategy: YouTube, TikTok, Instagram
  We tested whether the same content could work across multiple platforms. Results showed clear differences: YouTube, TikTok, and Instagram require distinct approaches. We also found that matching titles on TikTok increased views by +190%. Key takeaway: there is no universal content format.
• Stage 4. Optimization and scaling of working formats
  At this stage, we optimized existing winning formats instead of searching for new ones. We tested thumbnails, titles, trigger words, and combinations of approaches. This resulted in CTR growth of up to +30% and stable performance up to 10K views. Conclusion: scaling comes from precise optimization, not constant new hypotheses.