Views are the first number everyone checks and the least comparable one. The same count can mean a video that people chose to watch, a video that was shown to everyone and ignored, or a creator whose audience always shows up. Reading virality means swapping totals for ratios.
Why a view count is a weak yardstick#
Platforms define a view differently, and they change the definition. YouTube states that since 31 March 2025, Shorts views count each time a Short starts to play or replay, with no minimum watch time. The previous measure still exists in analytics as “engaged views” (YouTube Help). A Shorts view count today and a TikTok view count are not the same thing, and a Shorts view count from 2024 is not the same as one from 2026.
Platforms also say they rank on actions, not on size. TikTok’s recommendation explainer lists user interactions, video information and device and account settings as inputs, gives a strong indicator like finishing a longer video more weight than a weak one, and says follower count and past high performers are not direct factors. Instagram’s ranking explainer says Reels ranking predicts whether a viewer will watch a reel all the way through, like it or reshare it.
If ranking rewards completion and sharing, those are the signals worth reading from outside. You cannot see completion, but you can see its shadows.
Five ratios that travel better than views#
| Metric | Formula | What it answers | Watch-out |
|---|---|---|---|
| Outlier ratio | Views ÷ creator’s median views on recent posts | Did this video beat its own account’s norm? | Needs 10+ recent posts to be stable |
| Velocity | Views gained in first 24 hours, or views per hour since posting | Is it spreading fast right now? | Needs timestamps or daily snapshots |
| Share rate | Shares ÷ views | Did people pass it on? | Some platforms hide or rename shares and sends |
| Save rate | Saves ÷ views | Is it useful enough to keep? | Strongest for tutorials, recipes, products |
| Comment intent | Share of comments by type (see below) | What did the audience want to do? | Needs manual or analyzed sampling |
A sixth, reach beyond followers (views ÷ follower count), is a rough guide to whether a video escaped the creator’s own audience. Use it as a screen, not a score.
Each ratio has a different job. The outlier ratio finds the interesting videos. Velocity tells you how current the interest is. Share and save rates tell you why people cared. Comment intent tells you what they wanted next.
A worked example (illustrative numbers)#
Three videos in the same niche. All numbers are invented to show the method.
| Video A | Video B | Video C | |
|---|---|---|---|
| Views | 1,200,000 | 180,000 | 640,000 |
| Creator’s median views | 400,000 | 12,000 | 500,000 |
| Shares | 1,200 | 5,400 | 640 |
| Saves | 2,400 | 7,200 | 1,000 |
| Outlier ratio | 3.0 | 15.0 | 1.3 |
| Share rate | 0.1% | 3.0% | 0.1% |
| Save rate | 0.2% | 4.0% | 0.16% |
Video A has the most views, and it is a good result for its creator. Video C has a decent total and barely moved against its own norm. Video B has a seventh of A’s views, but it ran 15 times its creator’s median, and one in 33 viewers shared it. If you could study one video, B is the one. It found an audience beyond the creator’s own, and people sent it on.
Cutoffs depend on the niche, so build your own baseline. Pull 100 videos in your niche, compute share rate for each, and use the top quartile’s range as your bar.
Reading comment intent#
Comments are the cheapest audience research you can run. Sort 100 comments per video into types:
| Type | Example | What it signals |
|---|---|---|
| Purchase intent | “Link?”, “where can I buy” | Demand with a missing path to buy |
| Question | “Does it work on thick hair?” | The next piece of content to make |
| Objection | “This would break in a week” | The proof beat you need |
| Share intent | “Sending this to my sister” | Strong shareability |
| Replay | “Watched this three times” | The hook or payoff holds attention |
| Joke or noise | Emoji strings | Little to read |
Two videos with equal views can have very different mixes. A product video with 25 purchase-intent comments per 1,000 comments tells you more than one with 25 jokes.
What public data cannot show#
You cannot see retention curves, average watch time, completion rate, audience demographics or sales. Those belong to the account owner. When you commission a creator, ask for a retention screenshot for the post you want to learn from. It is the one metric outsiders cannot infer.
ContHunt’s own scores, including hook and relevance scores, come from analyzing the video and judging its fit to your question. They are evidence about the content and are not measured viewer retention. Use them to find and sort videos, and confirm what matters in your own analytics.
Red flags before you copy a video#
- Views with nothing behind them. Big counts and almost no comments, shares or saves can mean paid amplification or low-quality traffic.
- Generic comment floods. Dozens of near-identical praise comments suggest an engagement pod or automation.
- A single spike. A video that shot up on one day and flatlined may have been pushed or reposted by a large account.
- Missing context. A video that looks like an outlier because the creator posted a rare format may not generalize.
Inflated numbers exist. The FTC’s final rule on fake reviews and testimonials covers selling or buying fake social media indicators such as bot-generated followers or views. These red flags are heuristics, not proof, so look at more than one before you decide.
A four-step triage#
- Filter a relevant set by outlier ratio, and keep videos at 3x or more.
- Rank the survivors by share rate plus save rate.
- Read 50 comments on each of the top ten and classify them.
- Tag the hook and format of the top five, using the hook taxonomy.
Doing this in ContHunt#
ContHunt surfaces velocity, meaning videos that are growing faster than their creator’s average, and shows metrics alongside each analyzed video. In a Deep Search, ask for the ratios directly, and ask it to say where a platform does not expose a metric:
For each video, compute views divided by the creator’s recent median, share rate and save rate where available. Classify the top comments by intent. List the ten videos with the highest outlier ratio and comment intent that suggests purchase or sharing, with links.
Track the winners over several days with account and video tracking (experimental) to see whether velocity holds or decays. The competitor tracking guide covers the daily metrics.
Get started and rerun your last “best performing videos” list with ratios instead of views.
