227k YouTube Comments: What Audiences Actually Ask For
Every YouTube strategy deck starts with views, CTR, and retention. None of those metrics tell you what to make next. The one dataset that does is sitting ignored under every video: the comment section. We analyzed 227,445 AI-classified comments across 24 public channels and 11 niches (259,406 collected, 2,634 videos) and published the full benchmark with CSVs. Three numbers changed how we think about creator analytics.
Key Takeaways14% of all comments are explicit content demand — questions (8.1%) and suggestions (5.9%): 31,879 verbatim requests for videos that mostly don't exist yet.86% of audience questions get zero replies. The demand signal is not just unmeasured; it's unanswered.Demand structure differs by niche: Technology audiences ask at 24.9% of the reactive layer, Entertainment at 14.1%, Spirituality at 4.6%. You cannot extrapolate "what to post" across niches.13.5% of emotions are disappointment/skepticism — an early-warning layer that appears months before subscriber churn.1.5% of comments carry commercial intent — purchase and sponsorship signals that function as invisible lead gen.
The dataset, honestly
This is not a random sample of YouTube. The channels are tracked on OneTube, which skews toward creators and agencies who take analytics seriously — Technology accounts for 63% of comments. Classification is model-labeled (intent and emotion per comment), and reply_count = 0 is a lower bound on "unanswered," not proof of neglect. Full methodology, limitations, and CSV downloads are at the end. We'll re-run this benchmark quarterly; the next slice lands 2026-11-23.
With those caveats, the distributions are the first open look at what YouTube audiences actually write at scale.
What audiences actually write
| Intent | Share of analyzed comments | Count |
|---|---|---|
| Discussion | 52.9% | 120,399 |
| Praise | 24.5% | 55,637 |
| Criticism | 8.4% | 19,128 |
| Question | 8.1% | 18,485 |
| Suggestion | 5.9% | 13,394 |
| Spam / other | 0.2% | 402 |
AI audit of any YouTube channel
Drop a competitor's URL. In 5–15 minutes, get the full breakdown of what's working, what's broken, and exactly what to film next.
- 🎯Their content ideasVideos their audience keeps asking for that they never made
- ⚠️Their weak spotsExact topics and formats where viewers tune out or push back
- 💬Audience questionsStraight from their comment section — your next 10 scripts
- 📋A ready content planRanked backlog of what to film next, pulled from real demand signal
- 🔥Their superfansWho's emotionally invested in the channel and what gets them to talk
Just a URL and an email. Report lands in your inbox.
Half of every comment section is people talking to each other, not to the creator. Praise is a quarter. And the two categories that map directly to decisions — questions and suggestions — are 14%: tens of thousands of explicit "make a video about…" requests per quarter across a mid-sized channel portfolio.
Thoughtful, multi-sentence comments (40,648) outnumber quick reactions (20,821) roughly two to one in the analyzed set — the actionable layer of YouTube is text, and text is analyzable.
The unanswered question gap
Of 18,485 question-comments, 15,853 (85.8%) have zero replies. Whatever the reason — volume, workflow, or simply never seeing them — the most valuable comments in the section are the least likely to get a response.
For creators, that gap is a content calendar nobody is reading. For competitors' channels, it's even more interesting: the questions their audience asks and never gets answered are videos they will probably never make — and you can.
Demand differs by niche
Question share within the reactive layer (praise + question + criticism + suggestion):
| Niche | Question share | Channels | Comments |
|---|---|---|---|
| Technology | 24.9% | largest cluster | 163,668 |
| Entertainment | 14.1% | 4 | 23,519 |
| Spirituality | 4.6% | 2 | 41,225 |
A Technology audience behaves like a support forum plus a roadmap committee: questions (15,715) and criticism (14,964) nearly match praise (21,176). A Spirituality audience is devotional: praise dominates 15-to-1 over questions. Entertainment sits between, with the highest criticism-to-praise ratio after Technology. "Best practices" for content ideation that ignore this are averaging together different species of audience.
The emotional early-warning layer
| Emotion | Share | Count |
|---|---|---|
| Neutral | 43.6% | 99,216 |
| Grateful | 20.0% | 45,458 |
| Eager | 14.4% | 32,672 |
| Curious | 8.5% | 19,441 |
| Disappointed | 6.9% | 15,615 |
| Skeptical | 6.6% | 15,009 |
Disappointment, skepticism, and related negative emotions total 13.5%. Separately they're noise; as a trend line per channel they're an early-warning signal — format fatigue and unmet expectations show up in comments months before they show up in subscriber counts.
The commercial layer
3,493 comments (1.5%) are actionable commercial signals: "where can I buy," "do you have a course," "sponsor me." Small share, outsized value — this is lead generation happening inside comment sections that most creators never systematically read. For agency-managed channels, surfacing these is a deliverable clients can price.
Multilingual by default
In the weighted language sample, English is 65.4%; the remaining 34.6% spans 50+ languages (second largest: Uzbek, then Spanish and French). Audiences are multilingual by default. Any comment analysis that assumes English is silently discarding a third of the signal — and the demand inside non-English comments is exactly the content localization roadmap.
Methodology & limitations
Dataset: OneTube platform slice of 2026-08-23 — 259,406 collected comments, 227,445 analyzed (intent + emotion per comment), 2,634 videos, 24 unique public channels, 11 niches. Intent taxonomy: discussion / praise / criticism / question / suggestion / spam / personal story. Demand share = (question + suggestion) / analyzed. Unanswered gap = question-comments with zero replies. Negative emotion share = disappointed + skeptical + related / emotion-marked. Niche question rates exclude categories under 500 comments. Limitations: non-random channel pool (Technology-heavy), model-labeled classification without manual verification in this slice, reply_count = 0 is a lower bound. Aggregates only — no channel or commenter is identifiable. CSVs and SQL outputs ship with this article; quarterly refresh cadence.
Frequently asked questions
How many YouTube comments are actual content requests?
In this benchmark, 14% — questions (8.1%) plus suggestions (5.9%), 31,879 comments out of 227,445 analyzed. It varies by niche: Technology audiences ask far more than Spirituality audiences.
Do creators answer audience questions?
Rarely. 85.8% of question-comments in the dataset have zero replies. The demand signal is mostly invisible to the creator and completely invisible to view-based analytics.
Can comment analysis predict what video to make next?
It's the strongest direct signal available: explicit requests and recurring questions are stated demand, not inferred demand. Treat it as a ranked backlog — validated by repetition across videos — not as a crystal ball.
Is YouTube comment analysis only useful for big channels?
No — small channels benefit most from the competitor side: larger channels in your niche generate the question volume your own audience hasn't produced yet.
What to do next
Paste any public channel — yours or a competitor's — into onetube.io/audit: no card, no account, and you'll get a Pulse report showing what that audience is asking for, praising, and complaining about. If their viewers keep requesting videos that don't exist, that's your calendar. The 7-day trial continues with ongoing tracking.