Outlier Scores & Intelligence Metrics
Understand how Mostly calculates Outlier Scores, engagement velocity, and virality benchmarks.
When analyzing millions of short-form videos and ads, total view count alone can be misleading (a creator with 10 million followers getting 100k views is underperforming, whereas a creator with 2k followers getting 500k views is a massive statistical outlier).
Mostly calculates Outlier Scores to isolate true creative breakouts.
How Outlier Score is Calculated
The Outlier Score measures how significantly a specific video outperforms the creator's median benchmark baseline:
$$\text{Outlier Score} = \frac{\text{Video Views (or Engagements)}}{\text{Creator's Median Baseline Views (Past 30 Days)}}$$
Outlier Score Tiers:
- 1.0x – 2.0x (Standard Performance): Expected baseline performance for the creator's audience size.
- 2.5x – 4.9x (Strong Outlier): Above-average resonance; hook and topic gained broader distribution.
- 5.0x – 9.9x (Breakout Hit): High algorithm velocity; indicates exceptional hook retention in the first 3 seconds.
- 10.0x+ (Viral Megahit): True market breakout; proven mass psychological angle that should be adapted into your own creative testing matrix.
Key Creative Metrics in Discover
- Hook Retention Score: AI analysis of the visual and verbal opening 3 seconds.
- Engagement Rate: Ratio of comments, shares, and saves relative to total view count. High save/share ratios signify high-intent buying consideration.
- Active Ad Run Days: On paid Meta/TikTok ads, the number of consecutive days an ad creative has stayed active. Ads running 30+ days indicate profitable ROAS and sustained ad spend.