RemoveSpam

Free spam check

How much spam is in your comments?

A free, private report of the spam and scam comments on your recent videos. No sign-in, nothing to connect.

Research · November 2026

State of YouTube Comment Spam 2026

What 1,594,353 YouTube comments, 980,615 of them spam, say about who is spamming creators’ comment sections, what they post, when, and how it gets past filters.

By Dan, founder of RemoveSpam · Data from October 2025 to November 2026 · Methodology

Key findings

of spam was fake-coin promotion
94%
of comments were spam on the median protected channel
60%
of spam arrived within 24 hours of an upload
90%
of spam accounts posted only once
65%
of spam accounts were under a week old (vs 1% of real commenters)
84%
sampled spam comments contained a link
0 of 200

What the spam is

Almost all of it is one thing: comments promoting fake or worthless crypto tokens, written to look like ordinary viewers talking. Impersonation, contact-number spam and fake advisors are far smaller by volume, though comment for comment they are often the most harmful.

  • Fake coin promotion
  • Matched known spam
  • WhatsApp / Telegram contact
  • Impersonation
  • Other
  • Fake advisor
Show as table
TypeShare of spam
Fake coin promotion: 895,489 comments93.7%
Matched known spam: 45,241 comments4.7%
WhatsApp / Telegram contact: 6,689 comments0.7%
Impersonation: 5,872 comments0.6%
Other: 1,061 comments0.1%
Fake advisor: 944 comments0.1%

“Matched known spam” was caught by similarity to confirmed spam without a more specific type. A comment with several types counts once, under the most specific. What each type looks like.

How much of a comment section is spam

On the 30 client channels in the data, the median channel’s comments were 60% spam; across all their comments together it was 69%. Spam removed per month, November 2025 to September 2026:

Show as table
MonthSpam comments removed
November 2025127,788
December 202556,889
January 202664,932
February 202655,311
March 202642,309
April 202655,164
May 202683,162
June 2026112,390
July 2026150,578
August 202693,531
September 202648,960

When it arrives

Spam targets a video while it is new. Across 7,271 videos, 90% of spam arrived within 24 hours of the upload. Real viewers comment first; from five minutes to 24 hours after publishing, spam outnumbers them. Share of new comments that were spam, by time since upload:

Show as table
Time since uploadShare of comments that were spam
0–5 min after upload (3,354 spam, 7,634 genuine)30.5%
5–30 min after upload (52,558 spam, 29,120 genuine)64.3%
30–60 min after upload (61,400 spam, 29,987 genuine)67.2%
1–6 h after upload (242,413 spam, 154,866 genuine)61%
6–24 h after upload (412,528 spam, 119,522 genuine)77.5%
1–7 days after upload (79,671 spam, 69,201 genuine)53.5%
7–30 days after upload (1,145 spam, 4,015 genuine)22.2%
30+ days after upload (124 spam, 937 genuine)11.7%

By time of day, spam peaks at 16:00 UTC and is lowest at 06:00 UTC, a 3.7× difference. Weekends are about 24% quieter than weekdays.

Show as table
Hour (UTC)Spam comments
00:00–00:59 UTC42,679
01:00–01:59 UTC43,466
02:00–02:59 UTC31,909
03:00–03:59 UTC28,158
04:00–04:59 UTC26,406
05:00–05:59 UTC18,715
06:00–06:59 UTC18,326
07:00–07:59 UTC24,643
08:00–08:59 UTC24,824
09:00–09:59 UTC30,928
10:00–10:59 UTC33,415
11:00–11:59 UTC37,386
12:00–12:59 UTC42,541
13:00–13:59 UTC45,043
14:00–14:59 UTC41,473
15:00–15:59 UTC46,267
16:00–16:59 UTC67,578
17:00–17:59 UTC59,385
18:00–18:59 UTC51,515
19:00–19:59 UTC53,879
20:00–20:59 UTC53,817
21:00–21:59 UTC57,614
22:00–22:59 UTC56,084
23:00–23:59 UTC44,564
Show as table
DaySpam comments
Monday153,120
Tuesday162,657
Wednesday149,669
Thursday138,912
Friday148,686
Saturday111,971
Sunday115,600

Where it hides

Most spam is posted as top-level comments, where every visitor sees it, not as replies. Replies are where the personal scams happen, impersonation and contact numbers, but by volume they are a small share.

of spam was posted as a top-level comment
95%
of all top-level comments were spam, against 19% of replies
69%

Who posts it

Spam accounts are disposable. Of 620,774 distinct spam accounts, 65% posted once and were never seen again, and almost all posted five times or fewer. Number of spam accounts by how many spam comments each posted:

  • Posted once
  • 2 to 5 times
  • 6 to 20 times
  • 21 to 100 times
  • More than 100
Show as table
Spam comments postedAccounts
Posted once402,026
2 to 5 times213,899
6 to 20 times4,481
21 to 100 times356
More than 10012
of spam accounts were less than a week old when they posted
84%
of genuine commenters were
1%

Account age is based on accounts with a known creation date: 2,696 spam and 5,567 genuine.

How it gets past filters

In a random sample of 200 recent spam comments, none contained a link, the thing filters are best at catching. Almost half (93) named a coin ticker for viewers to search instead, and contact-number spam hid “WhatsApp” behind lookalike letters such as 𝖶𝗁𝖺𝗍𝗌𝖠𝗉𝗉. Coin names mutate daily (SPX85K, SPX89K, $$SPX85K), so blocked-word lists fall behind.

More in why YouTube’s spam filter misses scam comments.

How fast it can be removed

On a channel’s newest video, which RemoveSpam checks every minute, half of spam was removed within about 100 seconds of being posted, 70% within two minutes and 99% within five. Part of that time is YouTube’s own delay in making new comments available.

Methodology and limitations

  • Source: comments on YouTube channels protected by RemoveSpam, collected through YouTube’s official API, from 26 October 2025 to 8 November 2026: 1,594,353 comments, of which 980,615 were detected as spam and 934,051 removed.
  • Channel mix: most protected channels cover crypto, finance and tech. Channels sign up because they have a spam problem, so spam shares here are higher than on a typical YouTube channel. Read them as what heavily targeted channels face.
  • What counts as spam: comments RemoveSpam’s detection classified as spam. Types come from the detection patterns each comment matched, grouped into categories by RemoveSpam.
  • Monthly chart: November 2025 to September 2026. October 2025 is partial, and October 2026 is excluded because the number of protected channels changed significantly that month.
  • Samples: link, ticker and contact figures come from a random sample of 200 spam comments from the last 90 days; account-age figures from accounts with a known creation date. Times are UTC.
  • Privacy: no channel names, commenter names or comment-level data are published; every figure is an aggregate.

Using this data

You are welcome to quote and chart these figures with a link to this page. Suggested citation: RemoveSpam, “State of YouTube Comment Spam 2026”, removespam.app/research/youtube-comment-spam-2026. Questions about the data: [email protected].

Try it on your own channel

Connect your channel and see what RemoveSpam finds in the first hour. No credit card, and nothing to install.

Start your free trial

14-day PRO trial • no credit card • nothing to install

Or get a free spam check of your channel