Why Your Comment Count Doesn't Match the Number of Entries

It is the most common message creators send about comment pickers, and it always arrives right before a giveaway goes live: the video header says 4,812 comments, the picker finished loading and reported 2,930 entries, and something is obviously broken.

Almost always, nothing is broken. YouTube's displayed count and the entry pool a picker builds are measuring different things, and the gap between them comes from five separate causes. Understanding them takes ten minutes and prevents both a panicked support email and, more importantly, an announcement you have to correct later.

1. Replies are counted, but are not entries

This is the single biggest cause of the gap. The count YouTube shows under a video includes both top-level comments and every reply nested beneath them. A comment thread with one comment and forty replies contributes forty-one to that number.

Comment pickers draw from top-level comments only, and that is deliberate. If replies were eligible, a single person could reply forty times inside someone else's thread and flood the pool. Reply threads on giveaway videos are also mostly conversation — people tagging friends, arguing, or asking when the draw is — rather than entries.

On a video with lively discussion, replies routinely account for 30 to 50 percent of the total count. If your gap is roughly a third, this is almost certainly the whole explanation.

2. Held-for-review comments are invisible

YouTube holds comments for review under several conditions: your channel settings, a word on your blocked list, a link in the comment, a new account with no history, or the automatic likely-spam filter. Held comments appear in your Creator Studio moderation queue but are not published to the public API.

A comment picker can only read what YouTube publishes publicly. If a comment is sitting in your moderation queue, it does not exist as far as any third-party tool is concerned.

Giveaway videos are especially affected, because giveaway comments trip spam heuristics constantly — short, repetitive, keyword-heavy comments from accounts with no posting history are exactly what those filters are built to catch.

If you run giveaways, check Content, then Comments, then Held for review in Creator Studio before you draw. Approving legitimate held entries beforehand is the single most useful thing you can do for fairness, and it must happen before the draw, not after.

3. Deleted comments and closed accounts still count for a while

The number displayed on the video is cached and updates on a delay. Comments removed by their author, deleted by you, or lost when an account is closed can remain in the displayed total for hours. Spam sweeps also remove batches of comments without the header count updating immediately.

The API, by contrast, returns what exists right now. So the picker's number is frequently the more current of the two — the video header is the one lagging behind.

4. Your filters removed entries, as instructed

This sounds obvious, but it is easy to forget which filters were left on from a previous draw. Each one shrinks the pool:

• Unique users only — keeps one comment per person, usually their most-liked. On a video where enthusiastic viewers commented five times each, this alone can halve the pool.
• Keyword filter — removes every comment that does not contain the required word, including people who misspelled it, translated it, or used different capitalisation if exact matching is on.
• Minimum likes — most comments on most videos have zero likes, so even a threshold of one can cut the pool dramatically.

Before you draw, set the filters deliberately and read the resulting entry count out loud on camera. Announcing the qualifying number is what turns this from a confusing discrepancy into a transparent, explainable one.

5. The retrieval limit

Fetching comments through the YouTube Data API is paginated — a tool requests them in pages of a hundred and keeps going until it has what it needs. Every page costs API quota and time, so tools cap how far they go. CommentDraw reads up to the first 3,000 comments a video returns.

For the overwhelming majority of giveaways this changes nothing, because most videos never reach 3,000 top-level comments. But if yours does, the entry pool is a subset rather than the complete set, and you should say so when you announce.

Working out which cause applies to you

The gap can usually be attributed in a couple of minutes:

1. Turn every filter off and reload. If the number jumps substantially, filters were the main cause and the tool is behaving correctly.
2. Check the held-for-review queue in Creator Studio. A large queue on a giveaway video is common and explains a big chunk of the difference.
3. Estimate your reply volume by scrolling the comment section. Heavy reply activity explains most remaining gaps.
4. Compare the picker's number against 3,000. If it stopped near that figure, you hit the retrieval limit.
5. If the numbers still look impossible after all four, that is worth reporting — include the video URL, since nothing can be reproduced without it.

What this means for a fair draw

None of these effects favour any particular entrant. Replies are excluded for everyone, the spam filter does not know who you want to win, and the retrieval limit cuts by position in the returned order rather than by anything about the commenter. The draw within the pool remains uniformly random.

What they do affect is expectations. Someone whose comment was held for review will not be in the pool, will not know why, and may reasonably assume they were skipped. The fix is not technical, it is communicative:

• Approve held comments before drawing, not after.
• Announce the qualifying entry count on camera, alongside the video's total, so the gap is visible and explained rather than discovered later.
• Put your filters in the published rules so nobody is surprised by an exclusion.
• If your video is large, state the retrieval limit up front.

A gap between two numbers is only suspicious when it is unexplained. Explained out loud before the draw, it becomes evidence that you understand your own process — which is exactly the impression you want when you announce a winner.