YouTube engagement rate calculator

As of September 21, 2026, the median YouTube engagement rate across the overall sample of 31,749 videos from 521 curated faceless channels is 3.12%, measured as (likes + comments) / views. The 10th percentile is 1.10%, the 25th is 1.87%, the top quarter starts at 4.77%, and the top 10% starts at 6.79%. We lead with the median because a few viral videos skew the average, and this sample is curated faceless channels, not all of YouTube. Other published averages are often much lower because they divide by followers instead of by views. Quid's 2026 Social Media Industry Benchmark Report (the Rival IQ benchmark, now published by Quid) reports a YouTube median of 0.21% across industries, calculated by dividing total interactions by follower count. A per-follower rate and a per-view rate are different measurements. Source: https://www.quid.com/knowledge-hub/resource-library/blog/2026-social-media-industry-benchmark-report

The common definition, and the one sponsors use, is (likes + comments) ÷ views × 100. Some tools divide by subscribers instead; that punishes channels whose videos travel beyond their subscribers, which on YouTube is most of them. This calculator shows the per-view rate and also the like rate and comment rate on their own.

Across 31,749 recent videos from curated faceless channels the median is 3.12%, the top quarter starts at 4.77% and the top 10% at 6.79%. It falls as views rise, which is why the calculator compares you with videos of similar reach rather than with one global number.

From the Niche Finder snapshot on this page. Sample: curated faceless channels, mostly English. We included: “Elevate Niche Finder analytics: latest stats per video for the curated faceless channels (521 with a qualifying upload), videos published in the last 365 days, over 3 minutes long, with at least 1,000 views; hidden like counts excluded.” The formula is (likes + comments) / views, per view, shown as a percentage. We report the 10th, 25th, 50th, 75th and 90th percentiles. Snapshot date: September 21, 2026. View bands are stored as their own counts: 21,296 (1k–10k), 10,318 (10k–100k), 1,540 (100k–1M), 63 (1M+). Those add up to 33,217, which is 1,468 more than the 31,749 overall sample. Each band's maximum is the next band's minimum (10,000, 100,000, 1,000,000), so these rows are not a split of the 31,749. The number on a row is that row's stored count. The 31,749 is the sample for the overall percentiles. The niche menu lists 19 labels stored on this snapshot. Their video counts add up to 16,944, which is less than the 31,749 overall sample, so the menu is not a split of that sample. Each label keeps the percentiles stored on its own row.

Early viewers are subscribers and people who searched for the topic; they like and comment. Later viewers arrive from browse and suggested, watch more passively, and many leave before the end. So the denominator grows faster than the numerator. Judge a viral video against other viral videos.

Yes, as a filter. A healthy rate says the audience is real and paying attention; a very low one on a high-view channel suggests bought or passive traffic. Beyond the filter they care about your niche, your audience's country and age, and average view duration, none of which this number shows.

No. The three numbers you type stay in your browser. We keep nothing.