YouTube engagement rate formula
Engagement rate = (likes + comments) ÷ views × 100. A video with 48,000 views, 1,100 likes and 95 comments has an engagement rate of 2.49%. We divide by views, not subscribers: engagement by subscribers mostly measures how stale a subscriber count is, while engagement by views measures the people who actually watched, which is what a sponsor pays for. The same formula works as an influencer engagement rate check on any platform that shows views, but the benchmarks below are for YouTube long-form only.
How to read the result for a sponsorship
Engagement rate is a check on whether the views are real attention. It is not a price. Projected views tell you how many views a new video is likely to reach; engagement tells you whether those views look like a live audience or a number.
- Compare within size and format. Bigger channels usually get fewer comments per view. A 1M+ channel at 1 comment per 1,000 views is typical; a 30K channel at the same rate is well below its band. Shorts behave differently again, so don't compare a Short with a long-form video.
- Read comments per view first. Likes are one tap. A comment takes intent, so it is the stronger signal and the one that moves most between channels.
- Watch the like-to-comment ratio. In our sample, the typical channel gets roughly 9 to 14 likes for every comment on a long-form video, more on bigger channels. A video with more comments than likes is unusual: giveaways, comment-bait or bots. Read the comments before you count them.
- Very low is a flag, not a verdict. We treat fewer than about 1 comment per 2,000 views as an engagement mismatch worth investigating, because it can mean the views were inflated. It can also mean a topic people watch quietly. Check before you conclude.
- Use several videos. One video tells you little. Run five or ten recent long-form uploads and look at the middle one, not the best one the creator sent you.
Why engagement moves the price
The ThoughtLeaders fair-value estimate for a channel starts from projected views at day 30 and a cost per view that falls as channels get bigger. Then it adjusts up or down for a handful of signals: whether brands came back to the channel, the view trend, how long videos keep earning views, US audience share, whether the creator is on camera, and engagement, measured as views per comment. A channel with unusually few comments per view gets pushed down; one with an engaged audience gets pushed up. We don't publish the size of each adjustment, and engagement is not the biggest lever: brand renewals are.
You can see this on a real channel in how to check a YouTube sponsorship quote with your AI assistant, where a channel at about one comment per 667 views landed in a band that lowered its estimate. Those engagement bands are the estimate's own, not the medians on this page, so a channel can sit near its size band's median here and still be adjusted there. Use this page to compare a channel with its peers, and the estimate for price. For what brands actually paid by channel size, see what YouTube sponsorships cost in 2026.
Benchmarks: the typical channel by size
Medians from a live ThoughtLeaders query run on October 15, 2026. Long-form videos (over three minutes) published August 24 to September 15, 2026, with views, likes and comments all read at day 30. Every video in the sample was published after YouTube's view-count change, so all of it is on the new public count. For each channel we took its median video, then the median across channels in the band, so prolific channels don't dominate. Channels were a sample of active YouTube channels with at least four uploads in the last 90 days and at least three qualifying videos.
| Channel size | Engagement rate | Like rate | Comments | Sample |
|---|---|---|---|---|
| 10K to 100K subscribers | 2.1% | 1.6% | 1 per 370 views | 23 channels, 163 videos |
| 100K to 1M subscribers | 1.8% | 1.6% | 1 per 590 views | 43 channels, 484 videos |
| 1M+ subscribers | 1.9% | 1.8% | 1 per 1030 views | 41 channels, 791 videos |
Comment figures are rounded to the nearest ten views. The middle column of each row is a median of its own, so the like rate and comment rate don't add up exactly to the engagement rate. These are small samples; read them as a rough guide, not a standard.
Limits
- Comments and likes can be bought. So can views. A rate that looks great can be manufactured, and a rate that looks bad can come from bought views diluting a real audience. For how to tell, see the guide to checking a channel for fake views, which looks for late view spikes, engagement mismatches and bot comments.
- YouTube changed how views count. On August 24, 2026, public views started counting from the first frame. In our measurement, long-form public counts now run about 1.2 to 1.5 times the old basis in a video's first week. More views with the same likes and comments means lower rates, so a long-form video measured on the new count can look less engaged than one measured before. Our benchmarks are all on the new count, and they came out lower than the same query on videos published in the weeks just before it. Shorts didn't change.
- Video age matters. Views keep coming after comments slow down, so rates drift down as a video ages. Compare videos at the same age; day 30 is what we use.
- Engagement isn't sales. A loud comment section doesn't guarantee clicks or conversions. Brands renewing a channel is stronger evidence than any engagement rate.
Run it on a whole channel with your agent
Typing numbers in one video at a time is fine for a quick check. To see a channel's last ten long-form videos at day 30 next to this page's medians, connect the ThoughtLeaders CLI to your AI agent and ask in plain language:
Pull the last ten long-form videos for this channel with views, likes and comments at day 30. Give me the median engagement rate and comments per view, and tell me whether anything looks off.
Under the hood the agent starts with the channel record, then reads every video's day-30 snapshot and keeps the ten most recent long-form videos (over three minutes) that have one:
Or skip the typing: the ThoughtLeaders Chrome extension shows a channel's projected views, sponsor history and estimated price in a side panel right on YouTube.
FAQ
How do you calculate YouTube engagement rate?
Add a video's likes and comments, divide by its views, and multiply by 100. A video with 50,000 views, 1,200 likes and 150 comments has an engagement rate of 2.7%. This calculator uses views, not subscribers, as the denominator, because a sponsor pays for the people who watched.
What is a good engagement rate on YouTube?
In our October 15, 2026 sample of long-form videos read at day 30, the typical channel (the median) sat between about 1.8% and 2.1% depending on channel size. Above your size band's median is good; well below it is a reason to look closer. There is no universal cutoff, so compare a channel with channels of its own size and format.
Why measure comments per view instead of a percentage?
Comment rates are small numbers that are hard to compare as percentages. "1 comment per 500 views" is easier to hold in your head, and it is the same framing we use when we explain what moved a channel's fair-value estimate.
Should engagement rate use views or subscribers?
Views. Subscriber counts include years of people who no longer watch, so engagement per subscriber mostly measures how stale the subscriber count is. If you enter subscribers, the calculator uses them only to pick the right size band for comparison and to show views as a share of subscribers.
Does engagement affect what a YouTube sponsorship should cost?
Yes. In the ThoughtLeaders fair-value estimate, engagement is one of several adjustments that can move the estimate up or down from the projected-views baseline. It matters less than projected views and brand renewals, and it is easy to fake, so treat it as a check rather than the headline.
Can YouTube comments be faked?
Yes. Comments can be bought or generated by bots, and so can views. A high comment rate full of generic, repeated or off-topic comments is a warning, not a win. Read the comments, and run an authenticity check before you book.