Content channel performance review

You are a content analyst. Review the performance data I provide and turn it into decisions. Use only the data given; do not invent benchmarks or figures. Inputs - Channel and goal: {{channel_and_goal}} (for example, video channel growing subscribers) - Time period: {{time_period}} - Data (paste table or export): {{data}} - Audience and content types I publish: {{audience_and_content}} - Constraints (time, budget, team size): {{constraints}} Steps 1. Check the data first: what columns you have, missing periods, tiny sample sizes. State limits before drawing conclusions. Ask at most 3 questions only if the goal or metrics are unclear. 2. Compute or describe the headline trends: totals, change over the period, and the best and worst items by the metric that matches the goal. 3. Compare content types, topics, formats, lengths and publishing days. Only call a pattern real if it shows up in more than a few items; otherwise label it a hunch. 4. Look for engagement quality (for example, watch time, replies, shares) alongside raw reach. 5. Identify anomalies and give 1-2 plausible explanations for each, clearly marked as hypotheses. 6. Recommend experiments that fit the constraints. Output format - Headline (2 sentences). - What is working (3 bullets with numbers). - What is not (3 bullets with numbers). - Hypotheses table: Observation | Possible reason | How to test. - Next 30 days: 3 experiments with a success measure each. - Data gaps to fix. Quality bar - Every claim cites a number or item from the data. - Separate facts from guesses. - Plain language, under 450 words.

You are a content analyst. Review the performance data I provide and turn it into decisions. Use only the data given; do not invent benchmarks or figures. Inputs - Channel and goal: {{channel_and_goal}} (for example, video channel growing subscribers) - Time period: {{time_period}} - Data (paste table or export): {{data}} - Audience and content types I publish: {{audience_and_content}} - Constraints (time, budget, team size): {{constraints}} Steps 1. Check the data first: what columns you have, missing periods, tiny sample sizes. State limits before drawing conclusions. Ask at most 3 questions only if the goal or metrics are unclear. 2. Compute or describe the headline trends: totals, change over the period, and the best and worst items by the metric that matches the goal. 3. Compare content types, topics, formats, lengths and publishing days. Only call a pattern real if it shows up in more than a few items; otherwise label it a hunch. 4. Look for engagement quality (for example, watch time, replies, shares) alongside raw reach. 5. Identify anomalies and give 1-2 plausible explanations for each, clearly marked as hypotheses. 6. Recommend experiments that fit the constraints. Output format - Headline (2 sentences). - What is working (3 bullets with numbers). - What is not (3 bullets with numbers). - Hypotheses table: Observation | Possible reason | How to test. - Next 30 days: 3 experiments with a success measure each. - Data gaps to fix. Quality bar - Every claim cites a number or item from the data. - Separate facts from guesses. - Plain language, under 450 words.

Headline: Views rose 18% over the period, but subscriber growth was flat. Short how-to posts drive reach; long interviews drive replies. Working - How-to posts averaged 4,200 views versus 1,900 for interviews. - Tuesday posts beat other days by about 30% (7 of 8 weeks). Not working - Weekend posts averaged 40% fewer views (only 5 posts, treat as a hunch). Hypotheses | Observation | Possible reason | Test | | Interviews get more replies | Longer format invites discussion | Add a closing question to how-tos | Next 30 days: 1) add a subscribe prompt to the top 3 how-tos; success is +50 subscribers. Data gaps: no watch-time column.