The formula, and the second version
The simple version is (likes + comments) / views x 100. It is what almost every tool reports, which makes it the version to use when comparing across tools.
The stricter version divides by estimated reach rather than views, but reach is not public. So the simple version wins on practicality, not on precision.
| Version | Formula | Use it for |
|---|---|---|
| Simple | (likes + comments) / views | Comparing channels with public data |
| Weighted | (likes + 2 x comments) / views | Weighting conversation over reactions |
| Completion-adjusted | Engagement / average view duration | Your own dashboard only |
Pick one and stay with it. Switching formulas mid-analysis produces a chart that looks like a trend and is really a unit change.
Benchmarks by channel size
Small channels look better on this metric for a structural reason. Early viewers are subscribers who chose to be there. Recommendation bursts bring in strangers who watch and leave.
| Subscribers | Strong | Healthy | Weak |
|---|---|---|---|
| Under 1,000 | 6 to 10% | 4 to 6% | 2 to 4% |
| 1,000 to 10,000 | 5 to 8% | 3 to 5% | 1.5 to 3% |
| 10,000 to 100,000 | 4 to 6% | 2 to 4% | 1 to 2% |
| 100,000 to 1M | 3 to 5% | 1.5 to 3% | 0.7 to 1.5% |
| Over 1M | 2 to 4% | 1 to 2% | 0.5 to 1% |
Compare within a band. A 400-subscriber channel at 4% is underperforming. A 400,000-subscriber channel at 4% is doing well. The number is identical and the meaning is opposite.
Long-form, Shorts and live behave differently
Shorts inflate engagement because the like button sits under the thumb and the viewer count is enormous relative to the watch time. Live inflates it for a different reason. Chat counts as a comment.
| Format | Typical range | Why |
|---|---|---|
| Long-form | 2 to 6% | Time cost filters casual viewers out |
| Shorts | 4 to 12% | Low friction, high volume, short attention |
| Live | 8 to 25% | Chat is a comment stream |
Keep separate benchmarks for each format. Mixing them creates the illusion that a channel is collapsing when it just changed formats.
Reading the four common patterns
- High views, low engagement. The packaging over-promised. People clicked and left without reacting.
- Low views, high engagement. The video is good and the doorway is weak. Fix the title and thumbnail first.
- Both low. Usually a topic problem. Editing cannot rescue a subject nobody wanted.
- Both high, few subscribers. The channel lacks a reason to subscribe. Add a next step inside the video.
Each pattern points at a different fix. Reading the pair of numbers together is what makes the metric useful.
Five levers that actually raise it
- Ask for the like once, at the point of value, not at the start.
- Answer the first comment on every video within an hour. It doubles as a signal to later commenters.
- End on a question the audience can answer in one line.
- Cut the first fifteen seconds. Retention drives everything downstream.
- Post the same topic as a Short and a long-form. The Short feeds the long-form's comment section.
None of these are tricks. All of them are work.
Comparing two channels fairly
Comparisons go wrong in predictable ways. Four rules fix most of it.
- Same size band. A 2,000-subscriber channel and a 200,000-subscriber channel are not comparable on this metric.
- Same format. Shorts against long-form produces a gap that means nothing.
- Same age. A video in its first week is still accumulating reactions. Measure after thirty days.
- Same formula. If one number includes shares and the other does not, you are comparing two different metrics.
Follow the four rules and the comparison becomes useful. Skip them and you will conclude that a healthy channel is dying, or the reverse.
What a sudden drop means
Engagement rarely drifts. It falls, and the fall has a shape. Three shapes cover most cases.
- A cliff on one video. The packaging promised something the content did not deliver. Fix the title first, then look at the opening.
- A gradual slide over weeks. The audience mix is changing as recommendations widen. Usually not a problem, sometimes a warning about topic drift.
- A drop across every video at once. Something structural changed. A format shift, a posting schedule change, or a run of videos on a weaker topic.
Look at the shape before the number. A cliff and a slide call for opposite responses, and the percentage looks identical in both.
Why comments count for more than likes
A like is one tap and costs the viewer nothing. A comment costs a sentence, a decision to be visible, and a small social risk. That is why a video with 3% likes and 400 comments tends to outperform one with 6% likes and 20 comments.
The weighted formula reflects that by counting a comment twice. It is not a scientific adjustment, but it ranks videos closer to how a human would rank them, which is the point of any metric.
Practical use: when a video over-performs on comments, read them. The topic that generated the conversation is the topic to make next.
Matching the fix to the pattern
The number only helps if it points at an action. This table is the short version.
| Pattern | Most likely cause | First fix | Second fix |
|---|---|---|---|
| High views, low engagement | Packaging over-promised | Rewrite the title to match the content | Cut the first fifteen seconds |
| Low views, high engagement | Good video, weak doorway | New thumbnail, same title | Then test a new title |
| Both low | Topic nobody wanted | Change the topic, not the edit | Ask the audience what they want |
| High views, high engagement, low subs | No reason to subscribe | Add a next-step line mid-video | Publish on a predictable schedule |
Change one thing at a time. Two changes at once and you learn nothing about either, which is how channels spend a year making the same mistake in four different ways.
Reading your number honestly
Large channels carry a crowd of passive viewers who arrived from a recommendation and never meant to stay. Small channels usually have a higher share of people who chose to be there. Comparing a 2,000 subscriber channel against a 2 million one is comparing two different rooms, not two different skills.
Use your own history as the benchmark. A drop from eight percent to five percent on the same channel means something changed in the packaging. Five percent flat for a year on a channel that never grew means the format is stable and the reach is the problem, not the content.
Frequently asked questions
Should I include dislikes?
They are no longer public, so no.
Does the calculator work for Shorts?
Yes, but use the Shorts benchmark. A 10% rate on Shorts is ordinary and a 10% rate on long-form is exceptional.
Is a high rate always good?
Not on tiny numbers. Twenty percent on fifty views tells you nothing.
Can I compare with a competitor?
Yes. Views, likes and comments are public on both.
Why did my rate drop after a viral video?
Because the recommendation brought in strangers. It usually recovers as the audience mix settles.
Does engagement affect the algorithm?
It is a signal, not a switch. Watch time and click-through carry more weight.
More free tools for YouTube creators
AI Title Generator
48 worked title examples by niche.
Money Calculator
What a video earns before you film it.
Tag Extractor
Read the tags behind any video.
Engagement Calculator
Compare your rate with your size band.
Thumbnail Downloader
Study the covers that earned the clicks.
Channel ID Finder
Resolve a handle to the permanent UC ID.