Spotting Your Highest-Value Content With Clear Data

Look past total views

A video with half the views of another can earn double if its RPM runs high. To find these outliers, sort your videos by estimated revenue per thousand views over a fixed window. That view removes total views from the sort and shows earning power per unit of attention.

Focus on videos that earned real revenue, not zeros. Filter to videos published at least seven days ago and with at least a few hundred monetized views. Small samples lie, so set a minimum before sorting.

Find common traits

Once you have five to ten high-RPM videos, scan for shared traits. Look at length, topic, format, and traffic source. You may find that explainers on a narrow topic pull longer watch time and bring higher RPM. Or you may see that a certain format draws a specific source that pays well.

Do not copy the topic blindly. Copy the constraint. If the narrow scope is what forced clarity and held watch time, use that constraint on a new angle instead of repeating the exact title.

Test without breaking what works

Build a small batch of ideas based only on traits that appear in at least two of your high-value videos. Skip one-off traits. For each idea, predict RPM range and traffic source before making it. Then track actual RPM against that prediction after seven days.

Keep one slot for a test and do not swap it weekly. Give tests time for traffic to settle so you judge against real data, not a burst.

Key takeaways

• Sort by revenue per thousand views over a fixed window.
• Set a minimum for monetized views to avoid noise.
• Look for traits shared by at least two videos.
• Test one idea at a time against a prediction.
• Keep what already earns and expand its traits.

Make a shortlist this month

Pick five videos with highest revenue per thousand views in the last 28 days. Write one line for each trait they share. From that list, write one video script idea that uses only those shared traits. That is your highest odds test for growing premium revenue without guessing.