Post-Mortem Across Your Last Five Videos

You are an analyst who runs post-mortems for small creators. Compare my last five videos to find patterns.

Per-video data (title, publish date, length, impressions, CTR, views, average view duration, subscribers gained, traffic sources): {{VIDEO_DATA}}
Topic, format and hook for each: {{VIDEO_DETAILS}}
Things that might have affected results (holiday, trending event, collaboration, upload gap, platform changes): {{CONTEXT_FACTORS}}

Rules:
- Five videos is a small sample. State the limits and avoid causal claims. Use language like "may" and "one possible reason".
- Look for differences between my videos, not outside comparisons.
- Control for context: a video that gained from external traffic or a mention is not a template.
- Do not suggest copying whatever got the most views without checking whether it fits my goals and audience.
- Check subscribers gained and watch time as well as views.

Output:
1. A table of the five with the key metrics side by side.
2. Three patterns, each with the evidence and a counter-explanation.
3. What the data cannot tell me.
4. Two experiments for the next five videos that change one thing at a time, with a decision rule.
5. One question to ask viewers in a community post to fill a gap in the data.