This Metric Can Predict a Video’s Success
How you can use data to see whether or not a video will stick.
Dear Creator Nation,
In all professions and works of art, there’s nothing worse than a lousy outlier. A bug in code. An inexplicable loss in sports. A flopped movie from an established actor. Lousy outliers are, logically, insignificant (see ‘outlier’), yet their influence precedes likelihood. Lousy outliers can sway public opinion and damage reputations. This is why the world’s best (Steph Curry, Mr. Beast, Tom Hanks) rarely fall prey to bad performances.
The same concept is applicable for YouTubers—one bad video can cause the algorithm to turn on you and stop recommending your videos to new audiences. What if there was a way to know how a video might perform without anxiously monitoring it over the course of three months? A single metric, for instance, equivalent to ‘take this video to the back of the shed and kill it’?
Common knowledge proclaims that analyzing traffic sources early in a video’s lifespan can help forecast a video’s long-term performance. To verify the validity of this statement, we ran a multi-linear regression—basically analyzing a shit-ton of variables—across 150 videos to determine the significance of dozens of metrics like click-through-rates, 72hr view count, and suggested impressions.
For any data-nerds out there, we focused on the p-value of independent variables in relation to the dependent variable, views from Day 4—Day 365. And for those of you whose math careers ended in high school (trust me, I feel you), here’s what this means in plain english: Any variable with a p-value of under 0.05 signifies that this variable is statistically significant and is likely correlated to the # of views after one year. Here’s what we found:
Takeaway: The single greatest indicator, by far, of a video’s long-term success is the # of views that come from suggested impressions after 72 hours. This is likely due to the fact that suggested impressions are indicative of the algorithm's willingness to recommend your video to non-core viewers.
So, the next time you upload a video on YouTube, make sure you monitor its suggested impressions within the first 72 hours to see if the video will be a long-term success or not. To view a video’s suggested impressions:
Sign into YouTube Studio
↳ From the left Menu, select Analytics.
↳ From the top Menu, select Reach.
The key metrics card at the top of the Reach tab gives you an overview of your impressions.
😴 TL;DR
The best YouTubers in the world produce consistent bangers.
Poorly performing videos aren’t just bad because they are bad. They can trigger the algorithm to ignore future videos.
Suggested impressions within 72 hours of upload is the most accurate indicator of a video’s long-term performance.
YouTubers should monitor suggested impressions to help guage if a video will be successful or not.
🎥 Video of the Week — Why Ryan's World is the Richest kid on YouTube by Colin&Samir
🍿 Youtuber: Colin&Samir
👀 Views: 150k
👍 Likes: 7.6k
💬 Comments: 721
🔔 Subscribers: 618k
📋 Content for Creators:
🎬 Action Steps:
Follow us https://twitter.com/amagrowth
Think this would help a YouTuber or creator you know? Forward it to them!
Have a great rest of the week creators!
Sent with 💜,
Great article! for AB tests, do you recommend prioritising the CTR for suggested over other traffic sources in that case if you're a small channel and trying to grow?