Methodology & Architecture: How Feym Benchmarks 12,000+ Creators
Independent creator intelligence engine standardizing engagement benchmarks, percentile distributions, and audience anomaly detection across 180 market corridors.
1 Algorithmic ER Normalization across YouTube, Instagram, and TikTok
Cross-platform engagement rate comparisons are notoriously deceptive when performed without algorithmic normalization. Different social ecosystems operate under fundamentally contrasting distribution models and recommendation engines. On YouTube, genuine audience connection is best captured by measuring interactions against actual rolling 30-day video impressions, whereas Instagram historically benchmarks against registered follower counts. TikTok prioritizes algorithmic For You Page velocity decoupled from follower graphs. Feym implements a unified telemetry framework that strips away platform bias to reveal genuine social velocity and comparable commercial impact.
Mathematical Engagement Rate (ER) Formulas
Depending on platform architecture, Feym calculates standardized performance metrics scrubbed of inorganic bot traffic using the following equations:
(Likes + Comments) / Views × 100
Normalized against 30-day median video views.
(Likes + Comments) / Followers × 100
Scrubbed against verified active follower base.
(Likes + Comments + Shares) / Views × 100
Includes viral share velocity coefficients. 2 Size Cohorts and Statistical Percentile Distribution
An Engagement Rate of 1.8% for a mega-creator with 5 million followers on Instagram places them comfortably within the 90th percentile (top 10%) of their tier globally. Conversely, that identical 1.8% score for a micro-creator with 15,000 followers represents severe underperformance against category benchmarks and signals audience stagnation. Feym stratifies creators into five discrete cohorts: Nano (<20k), Micro (20k-100k), Mid-Tier (100k-500k), Macro (500k-2M), and Mega (>2M), calculating empirical percentiles (P25, P50, P75, P90) across 180 geographic and niche verticals.
3 Audience Quality Score (AQS) and Bot Farm Detection
The Audience Quality Score (AQS) evaluates account integrity on a rigorous 1-to-100 scale. Our automated algorithms audit lexical comment entropy, flag repetitive emoji clusters indicative of automated engagement pods, evaluate like-to-view ratios, and map geographical dispersion. Suspicious clusters originating from known click-farm IP blocks trigger automated anomaly indicators directly in the terminal.
4 Cohort Benchmark Reference Matrix (2026)
The reference table below presents aggregated engagement thresholds and expected authentic reach (Real Reach) percentages across all five creator size tiers indexed in the Feym database:
| Creator Cohort | Follower Range | Median ER (P50) | Top 10% ER (P90) | Avg Real Reach |
|---|---|---|---|---|
| Nano-Creators | 5,000 – 20,000 | 4.32% | 7.65% | 38% – 52% |
| Micro-Creators | 20,000 – 100,000 | 2.71% | 4.89% | 26% – 35% |
| Mid-Tier Creators | 100,000 – 500,000 | 1.84% | 3.25% | 18% – 25% |
| Macro-Creators | 500,000 – 2,000,000 | 1.31% | 2.40% | 12% – 18% |
| Mega-Stars | > 2,000,000 | 0.92% | 1.78% | 8% – 14% |
Frequently Asked Questions about Feym Screener
How frequently is creator telemetry updated in the screener?
How does Feym differ from conventional influencer marketing databases?
Can creators pay to artificially improve their ranking or AQS score?
How is Estimated Sponsorship Pricing calculated in the directory?
Audited & verified by Zordon Intelligence sp. z o.o. research team.