Video Title Patterns: Correction to a 260-Result Search Sample

Correction, 2026-11-16: We withdraw the treatment of the up-to-three-minute group as verified Shorts and the general rules about short titles, hashtag counts and removing hashtags from longer videos. Original numbers and charts remain as historical observations, not a success formula. They were not independently recomputed because the raw dataset was unavailable.

Shorts classification depends on aspect ratio and upload date as well as duration. See YouTube’s official Shorts classification guidance. YouTube’s hashtag guidance allows relevant hashtags in titles and descriptions for videos and Shorts. This sample did not test an optimal count or a benefit from removing them.

Summary

  • The original record gives median title lengths of about 26 characters for videos up to three minutes in the Korea-region searches and 66 in the US-region searches. These are observations from different topics and search samples.
  • Recorded title hashtag use was 39% and 59% in the two up-to-three-minute groups. No title hashtag was observed among the 62 results longer than three minutes.
  • These features appeared alongside high views; this does not show that the titles caused them.

Titles by the numbers

The October 23, 2026 record covers 260 search results: the top 20 by views from the previous seven days across 13 topics, grouped at three minutes. We did not verify aspect ratio, so the shorter group is not a verified Shorts sample.

GroupTitle length (median)Uses hashtagsUses emoji
Korea, 3 min or less (100)about 26 chars39% (1.3 on average)7%
Korea, over 3 min (40)33 chars0%30%
US, 3 min or less (98)66 chars59% (2.3 on average)45%
US, over 3 min (22)52 chars0%36%
Title length (median characters)
Korea ≤3 minabout 26
Korea >3 min33
US ≤3 min66
US >3 min52
Titles using hashtags
Korea ≤3 min39%
Korea >3 min0%
US ≤3 min59%
US >3 min0%

Recorded examples from Korea-region videos up to three minutes

  1. One short sentence plus one or two hashtags. The most viewed video in the budget-cooking search sample was a 38-second video whose title translates as ‘Food that's better because it's flawed’ (about 2.53 million views). It explains nothing and leaves one question.
  2. ‘Reason’ or ‘why’ titles. The record reports these words in 14% of Korea-region titles up to three minutes, such as ‘Two reasons your stir-fried anchovies fail’ (about 300,000 views). The problem goes in the title and the answer stays in the video.
  3. Numbered series. One channel placed three episodes of a daily challenge, ‘Living on a ₩350,000 food budget, day 18’, in the budget-cooking top 20. No other topic showed numbered titles like this, so treat it as a single example.

The US-region sample up to three minutes looked different

Videos up to three minutes in the US-region search sample often described the content at length and added two or three hashtags plus emoji, for example ‘Breakfast meal prep’ tagged #asmr #breakfast #mealprep. Many of these results came from Indian creators and covered different topics from the Korean searches, so they don't describe US viewers as a whole.

Titles of videos over three minutes in this sample had no hashtags

None of the 62 videos over three minutes used a hashtag in the title, in either group. Some Korean vlogs (4 of 17) opened with a format label in square brackets, such as ‘[VLOG]’.

Examples

The videos mentioned above. Titles and view counts on the cards are current YouTube data.

Optional experiments, not production rules

  • A short sentence and a more descriptive title can be candidates to compare. A median of about 26 characters does not establish an optimal length or a 30-character limit.
  • Use a ‘reason why’ or ‘mistake’ frame only when the video really answers that question; a forced frame makes the title and the video disagree.
  • Consider whether episode numbers help viewers understand a series. One channel’s example does not establish a view-count benefit.
  • Do not remove hashtags from all long videos because this sample recorded 0%. If comparing relevant hashtags or format labels, change one element at a time and record differences in period, impressions and traffic source. Before-and-after results alone cannot establish causation.

A caution when learning from popular videos

Borrowing a title format is different from reusing someone else's video. Clipped drama scenes also ranked in this data, but YouTube's monetization policies exclude ‘reused content’, meaning existing videos repurposed without meaningful commentary or changes. On July 15, 2025 the ‘repetitious content’ policy was renamed ‘inauthentic content’ (YouTube channel monetization policies). Borrow the format, but shoot and narrate it yourself.

Method and limits

  • The raw per-video title dataset and aggregation code were unavailable for this correction, so the original figures were not independently recomputed. Deduplication, the denominator for mean hashtag count and character/emoji counting conventions remain unverified. Figures are preserved as previously published observations.
  • On October 23, 2026 we counted the titles of the 20 most viewed videos from the last seven days for 7 Korean and 6 US topics, 260 titles in total. The Korean ‘AI’ search was left out because only 3 of its 20 titles were in Korean. View counts are as of that day.
  • The Korean and US searches covered different topics, and a region setting does not exclude videos from other countries.
  • These features appeared together with high views; this does not show that the titles caused them. Thumbnails and video content were not reviewed.
  • Length by topic is in Video length by topic and interpretation correction.

To check recent videos in your own topic, search a keyword in Rising videos.