How to tell whether your YouTube views are real

Someone sells you ten thousand views. The counter goes up. Did you get what you paid for?

The view counter is the least informative number on your channel. It tells you a playback was registered. It tells you nothing about who registered it, whether they watched, or whether the number will still be there next month.

Everything useful is one click deeper, in analytics you already have. Here are four checks, in order of how much they tell you.

#Check one: where the views came from

This is the decisive one, and most people never open it.

In YouTube Studio, go to Analytics → Advanced mode → Traffic source. You'll get a list of every route people took to your video.

YouTube names these routes explicitly. Browse features. Suggested videos. YouTube search. Playlists. Notifications. End screens. External sources, which Google defines as "traffic from websites and apps that have your YouTube video embedded or linked to."

And one more, which matters here: YouTube advertising. Google's own description is blunt — "If your video is used as an ad on YouTube, you'll see 'YouTube advertising' as a traffic source."

That single line is why this check settles the question. Paid promotion through YouTube's own ad system appears in your analytics as its own labelled row. You can see it. You can screenshot it. Nobody has to tell you it happened.

Views from a panel or reseller do not get that row. They typically land under Direct or unknown, which Google defines as "traffic from direct URL entry, bookmarks, signed out viewers, and unidentified apps."

Now the honest part, because this is where a lot of online advice oversells itself: Direct or unknown is not proof of anything. It is a genuine category with genuine uses. Someone pasting your link into a group chat lands there. So does a signed-out viewer. So does an app YouTube doesn't recognise.

What makes it a signal is proportion. A small slice of direct traffic is ordinary. A video where direct or unknown is the overwhelming majority, on a channel whose other videos look nothing like that, is a video whose traffic did not come from YouTube.

#Check two: whether anyone did anything

Real viewers leave traces. They like, they comment, occasionally they subscribe. Purchased viewers almost never do, because the economics don't allow it — engagement costs more to fake than playback does.

So compare the ratios across your own videos rather than against some internet benchmark:

• Likes as a share of views. Healthy channels usually sit somewhere around two to five percent, though this varies enormously by niche and audience.
• Comments as a share of views. Much lower, but it should not be zero on a video with meaningful reach.

What you're looking for is not an absolute number. It's a break in the pattern. If eleven of your videos sit at three percent likes and one sits at a tenth of that while carrying forty times the views, the outlier is telling you something.

Look up your video's real numbers → — free, no account, pulled live from YouTube's public data.

#Check three: whether they watched

Open Average view duration and Average percentage viewed for the video in question, then compare against your channel's norm.

Bought views collapse here. A view registered by something that opened the page and left does not produce watch time. On a video with ten thousand such views, average duration falls off a cliff relative to everything else you've published.

This check is useful precisely because it's hard to fake cheaply. Holding a playback open long enough to generate watch time costs real resources, which is why most services don't bother.

#Check four: who and where

Open the Audience tab and look at geography.

Your real audience has a shape. It's concentrated in particular countries, skewed toward particular ages, arriving at particular times. That shape is stable across your catalogue.

Purchased traffic rarely matches it, because it comes from wherever the supply is cheapest. A channel whose audience is reliably British and American, with one video whose views are overwhelmingly from somewhere that appears nowhere else in its history, has a video that was delivered rather than discovered.

#The number can leave again

This is the part people find out too late.

YouTube validates engagement after the fact. In its own words: "To verify that metrics are accurate, YouTube may temporarily slow down, freeze, or change your metric count, and discard low-quality playbacks."

So the count you see today is provisional. Validation runs over hours and days, and views judged invalid are removed — after you've paid for them.

This got sharper in August 2026, when YouTube changed how views are counted so that a view registers the moment a video starts to play, across Shorts, long-form and live. The first number you see now means less than it used to, and the number that survives validation means more. If you're judging a purchase on the counter an hour after delivery, you're reading the least reliable version of it.

#What a genuinely viral video looks like

Worth saying plainly, because every check above can produce a false alarm.

A video that escapes your subscriber base and reaches strangers will have high views, a lower like rate than your usual, unfamiliar geography, and lower average duration. That is not fraud. That is reach.

The difference is in the traffic source. A video that took off organically shows it: Browse features and Suggested videos climb, because YouTube's recommendation system is what carried it. A video that was delivered shows nothing of the kind.

Check that row before you conclude anything. One screenshot separates the two cases, and getting it wrong in either direction is expensive.

#If the checks come back badly

Three things are true at once, and it's worth being clear about all three.

You probably can't get the money back. Services selling this are not generally in the business of refunds.

The views may be removed anyway, on YouTube's schedule rather than yours.

The damage that outlasts the views is the signal damage. YouTube's recommendation system reads how people behave around your video. Ten thousand playbacks that produced no watch time and no engagement tell that system your video isn't worth showing to anyone. That verdict doesn't reverse when the fake views disappear.

The practical response is to stop adding to it, let the video settle, and judge your next upload on what real viewers do with it.

#The principle underneath all four checks

A view is a claim about a person. The counter records the claim; your analytics record the person — where they came from, how long they stayed, what they did, where they live.

Any promotion worth paying for should survive that inspection. If a service can't show you where its views appear inside your own Studio, that isn't a detail it forgot to mention.

---

Sources: YouTube fake engagement policy, Understand your YouTube video reach, How engagement metrics are counted.