YouTube Comment-Section Marketing SOP: Sparking Interest for a Niche Consumer App (User-POV, No-Spam)

0. Positioning (read first — it governs everything)

Don't frame the goal as "comment the product everywhere so people download," and don't over-engineer it into "get them to ask what it is" (that rarely happens and reads contrived). Frame it as:

Find the specific pain that your audience is talking about right now, help genuinely from a real user's POV, and — when the product honestly fits that exact moment — recommend it the way a user naturally would (short, specific, no hype). Being consistently helpful is what earns the occasional recommendation its weight.

The best outcome: a reader thinks "this person actually gets the hard part, and the thing they used sounds like it'd help me too." A direct, low-key recommendation from someone who clearly understands the struggle beats any clever open-loop.

1. Compliance & honesty (never break)

• Speak as a genuine user. If you actually use the product, a user-voice recommendation ("what helped me was...") is truthful. Never fabricate an experience you didn't have. Noting a deeper connection (e.g. "ngl i also help build it") is optional if it comes up naturally — not required. Don't pretend to be an unrelated stranger who accidentally stumbled on it either; be a real person who uses it.
• YouTube comment-spam policy: no identical/similar messages across many videos; no high-volume repetitive comments to drive traffic; no automation/AI to mass-produce near-identical comments. → Every comment is hand-written, context-specific, posted after actually watching.
• Product-claim honesty: never claim a false price (don't say "free" if it isn't) · never "cure / guaranteed / clinically proven / best app" · no fake ratings/user-counts/success-rates. For health-adjacent products, claims must match evidence — never extrapolate to specific unproven outcomes.
• 🚨🚨 Safety red lines — no product mention, usually skip entirely (no comment, no like): ① self-harm / suicidal ② underage or age-ambiguous (creator/commenter reads as a minor; kids/school channels) ③ eating-disorder / body-image vulnerability ④ serious medical context ⑤ anyone asking for medical advice — don't give it.

2. Identity: a real person who has the problem, and uses the product

You're not a marketing bot and not a fake stranger. You're someone who genuinely gets the struggle and uses the product for the hard moments — so you help first and recommend it the way you'd tell a friend. That honesty keeps the account and brand safe long-term. Ask of every comment: would a real person who actually gets this write it, and is everything in it true?

3. Find the right videos (don't just keyword-search — filter by pain density)

Not every on-topic video is worth commenting on. You want comment sections with real, specific pain, not generic explainers.

Priority tiers (quit-smoking example — adapt to your niche):

• A-tier (most worth it) — comments full of real "me too, day 1 / day 3 / relapse / cravings": quit smoking day 1 / quit smoking day 3 / quit vaping day 1 / i relapsed smoking / 30 days no nicotine / quitting vaping vlog / how i quit nicotine / struggling to quit smoking / nicotine withdrawal vlog.
• B-tier (comment, more restraint — no strong push) — high traffic but drifts into method debates: nicotine addiction explained / smoking timeline after quitting / cold turkey vs nicotine patch. Post a behavior-level observation, don't push the product.
• C-tier (watch/like only, or skip — NO product mention) — minors, serious illness, pregnancy, self-harm, body-image: teen vaping / pregnant and smoking / COPD smoking / cancer smoking story. Often just skip.

Discovery & timeliness (be timely, not "snipe"): speed is for joining while the discussion is live and real people are there, not grabbing the top slot with an ad.

• 🔔 Channel bell: subscribe to reliable creators in your niche, All notifications. The first ~1h after upload = fresh, active.
• 📡 Channel RSS (automation-friendly, no API): https://www.youtube.com/feeds/videos.xml?channel_id=<CHANNEL_ID> — poll on a timer, compare <published>, surface videos <2h old. (channel_id: channel page → view source → "channelId".)
• 🔍 Search by upload time: Filters → Upload date = Last hour / Today + Sort by upload date.
• Maintain a watchlist (handle + channel_id); add a few new channels weekly, drop dead ones.

4. Video Fit Score — is this video worth entering? (12 pts; <9 skip, <10 no product mention)

Rule: 0-7 = skip · 8-9 = watch/like/pure-value comment only, no product · 10-12 = may comment, and only if context is natural, may recommend the product (once).

5. Per-video action cap (don't "occupy" a comment section)

Leave one precise main comment + a few meaningful replies, never a carpet:

• ≤1 top-level comment · 0-2 replies · ≤3 total interactions
• ≤1 comment that recommends the product. If the top-level comment already mentions it → all replies are pure value. If it doesn't → a reply may mention it only when a user explicitly asks for a tool/app/method.

6. Before commenting, scan three things (don't reply to just anyone)

Don't write immediately. In order:

1. The video's detail anchor — watch ≥2-5 min, grab one concrete point (e.g. after-dinner craving / driving-home craving / day-3 irritability / relapse shame). Your comment must reference this anchor. No anchor = it reads like AI.
2. Top comments — learn the video's resonance point (what people agree on). Match that, not a generic "this is hard."
3. Newest comments — where the high-intent people are ("i'm on day 1 right now," "any app that helps?"). Best reply targets — they're expressing a need.

7. Top-level comment structure (name the pain, don't summarize the video)

Goal: a passer-by nods "yes, exactly this." Structure = video detail anchor + name a specific pain + a small observation/tactic + a light open-ended close + (optional) a natural user recommendation. One point per comment. Never "Great video! Stay strong" (AI + empty).

No-mention version (most videos):

the after-dinner part you mentioned is so real. people talk about "quitting nicotine" like it's one big decision, but half of it is just your brain looking for the old ending to a routine. food ends, hand reaches for a cigarette. breaking those tiny loops one at a time feels way less impossible than trying to become a whole new person overnight.

With product (only for 10-12 Fit videos, user voice):

the after-dinner part you mentioned is so real. a lot of it is surviving those tiny routine gaps where smoking used to sit. the thing that actually helped me there was an app called MoltAI — i open it the second a craving hits and talk it out for a few min before it turns into autopilot. not a magic fix, but it got me through the worst windows.

Why it works: honest user recommendation, specific to the pain, no cure/guaranteed/false-price, no link, no "download," product embedded in the moment (not tacked on).

8. Reply structure (catch them first, don't pitch a tool up front)

Shorter, more human. Structure = restate their situation + one small suggestion + optional question + recommend the product only if they clearly need a tool.

• "I'm on day 3 and losing my mind" → day 3 is nasty because it feels like your brain is bargaining every 10 minutes. don't make a forever decision today. shrink it down: shower, walk, water, timer, anything that breaks the loop for a few minutes. (no product)
• "I relapsed after 12 days" → 12 days still counts. the relapse doesn't erase the reps you built. the useful part is figuring out the exact moment it happened. that's the thing to build around next time. (no product)
• "Any app or tool that helps when cravings hit?" → i use one called MoltAI for exactly that — you open it the second the craving spikes and talk it out before autopilot wins. i'd still pair it with a physical reset tho: stand up, change rooms, water. (product OK, user voice)
• "Cold turkey or patches?" → i'd be careful taking medical advice from a comment section, but behavior-wise the overlooked part is the emergency plan — whatever method you use, there's still that moment the craving spikes and you need something to do. (no product, no medical advice)

9. How to recommend (natural user voice)

Recommend it the way you'd tell a friend what worked — specific to the moment, low-key, honest. Mentioning a deeper connection is optional; the default is simply a user recommendation.

OK: the thing that helped me was an app called MoltAI · i use one called MoltAI for that exact moment · it's for the "next few minutes," not a magic fix · (optional) ngl i also help work on it, grain of salt. Never: will help you quit for good · cures X · guaranteed · clinically proven · the best app · free (if untrue) · download now · link in bio · any experience that isn't true.

10. Recommendation frequency (conservative — makes each one land)

Entering others' comment sections is easily read as advertising, so keep it rare:

• Top-level comments: ~80% pure value / no product · ~20% natural product recommendation.
• Replies: ~90% pure value / no product · ~10% only when they ask.
• Per video: ≤1 comment recommends the product.

An account that names the same app in every comment gets read as marketing fast. Rarity keeps it credible.

11. Anti-AI: comments need "rough specificity"

AI smell = too complete, too positive, too summarizing, no concrete scene, no personal POV, every sentence too smooth, therapist voice, brand-copy voice.

Banned openers: X is a journey... · This is such an important reminder... · Everyone's path is different... · Stay strong, you've got this... · Every small step matters... Recommended openers (start with a concrete scene): the after-dinner craving is the sneaky one · day 3 is where your brain starts lying professionally · the car trigger is so underrated · the "i'll just have one" voice is evil lol Don't over-pack — ❌ Quitting requires patience, discipline, and self-compassion, and every craving you overcome proves you are stronger than your addiction. 🟢 the annoying part is the craving doesn't even need a good reason. it just shows up like "hey remember that thing we're not doing anymore?" lol One point per comment. Pick one resonance point; don't cram five.

12. Comment quality gate (10 pts; <8 don't post, <10 no product)

Rule: <8 don't post · 8-9 OK but no product · 10 may recommend. Four pre-publish questions (any bad answer → don't post):

1. Could this be copy-pasted under any on-topic video? If yes → don't post.
2. Does it reference a specific detail from the video/comment? If no → don't post.
3. Does it still stand up without the product? If no → don't post.
4. Is anything in it untrue? If yes → don't post.

13. Template directions by video type (rewrite every time)

• A. Early-days vlog (day 1/3) — value: name the "routines feel empty" pain; +product: "the thing that got me through those windows was an app called MoltAI, i open it the second a craving hits instead of white-knuckling it alone."
• B. Relapse — value: "relapse is usually information — what time, who, what excuse, what feeling"; +product: "i keep MoltAI on my phone for the moment a craving spikes, the no-shame restart part is what i needed."
• C. How-to / method — value: "people skip the 4-minute-before-i-cave plan"; +product: "for that window i use MoltAI, doesn't replace your method, just covers the moment."
• D. Quit vaping — value: "there's no clear smoke break anymore, it's a tiny button your brain keeps pressing"; +product: "MoltAI's the thing that interrupts that automatic reach for me."
• E. Health/science — usually no product: reframe why "just stop" is useless advice.
• F. Long-term success — usually no product (audience wants hope/identity, not a tool).

14. Reply strategies by commenter type

• Just starting ("18 hours in, want to quit quitting") → shrink the timeframe, no product.
• Repeat-failer ("tried 6 times") → reframe as a trigger-pattern problem, no product.
• Explicit tool ask ("any app that helps?") → recommend, user voice.
• Scary symptoms / anxiety / can't sleep → no product; gently point to a clinician.
• Underage ("I'm 15 and can't stop vaping") → no reply, no like, no product. (red line)

15. Make it a natural recommendation, not a CTA

Don't write Download X / Try X now / Link in bio. Recommend it inside the specific moment, like telling a friend. If someone follows up ("what is it? does it work?"), answer plainly and honestly — what it does, and that it's not a magic button.

16. 30-second anti-AI rewrite (before every post)

1. Delete generic encouragement unless already very concrete.
2. Add one video detail.
3. Add one human micro-friction (lol / imo / weirdly / not even because...).
4. Cut brand voice → plain "the thing i open when i'm about to cave."
5. Read it once: if it reads like "task #7 of today's KPI," don't post.

17. Before / After

• 🔴 AI: Quitting is a difficult journey that requires patience. Every craving you overcome is progress. Stay strong, you are not alone.
• 🟢 Human: the after-dinner craving is the sneaky one. more like your brain wants the old ending to the meal than actual nicotine. changing that "ending" helped me more than trying to motivate myself harder.
• 🟢 Human + user rec: ...the app i use for that exact window is MoltAI — open it, talk the craving down, move on before it turns into autopilot.

18. Account health + warm-up (avoid bans, market long-term)

Use a real personal account of someone who genuinely uses the product (more authentic; spread across a few accounts beats one dedicated one). Each account still warms up and behaves like a real person.

Daily account-state check (<3 min)

• Shadow-removed? Log out / open incognito on a video you commented on — is your comment still there? (Logged in, you only see your own ghost comment.) Invisible = shadow-removed → stop recommending the product, pure value for a few days. This is YouTube's signature trap — check incognito regularly.
• Survival: are your last 5-10 comments still live?
• Engagement: any likes/replies? Long-term zero → change scripts/videos.
• Post smoothness: "held for review"? → pull back.
• Friction: verification / rate limit → pause 1-2 days, just watch.
• "spam/ad/bot" callouts → too promotional, reset to pure value.

Phases × weight (by account state, not calendar)

• P1 Warm-up (new / first 1-2 weeks): 80% warm-up / 20% outreach; almost no product mention; 3-5 comments/day. Watch a lot, like others' comments, reply to replies.
• P2 Grow (comments survive + get likes): 50/50; product per the §10 split; 8-12/day.
• P3 Mature (steady likes, healthy): 40/60; 15-20/day, still spread out.

Level up only after several clean days; drop a level the same day on any warning.

Ban-avoidance rules

1. Never copy a comment across videos — the most direct spam signal. Rewrite every one.
2. No links; recommend by name and let people look it up.
3. Never burst: ≤3 interactions per video; ≤1-2 comments per short session; spread timing across the day.
4. Consumption ≥ outreach.
5. No em-dash, colloquial, one concrete point.
6. Periodic incognito shadow-removal check.
7. Don't pile actions into a short window; slow down at low-activity hours.

19. Suggested cadence (small steps, no bursts)

Every ~15-20 min: quick account-state check → reply to replies you got → sweep for a fresh (<2h) video via bell/RSS/search → actually watch a video + like 1-2 good comments → pick one video (Fit ≥9), scan the 3 things, do one thing (top-level / 1-2 replies / like-skip) → quality gate → log. One video's comment section per session.

20. Comment CRM (track why it worked, not just likes)

Log each comment: video URL, pain anchor, commenter intent (day1 / relapse / asks-for-app / method-debate / emotional-support), comment type, product mentioned?, rec reason, stands-without-product?, AI-smell 1-5, likes, replies generated, alive-in-incognito?, negative signal, reusable insight.

Weekly review — not "which got most likes" but: which pains trigger the most replies? which scenes fit a natural recommendation? which phrasings read as ads? which video types aren't worth it? Turns comments into acquisition + user research.

21. Hard rules (non-negotiable)

1. Don't watch → don't comment.
2. No video detail → don't comment.
3. Per video: ≤1 top-level + ≤2 replies.
4. Per video: ≤1 comment recommends the product.
5. Recommend as a genuine user; never fabricate an experience.
6. No links, no "download now," no "link in bio."
7. No product mention around minors / self-harm / body-image / serious-medical.
8. <8/10 quality → don't post; <10/10 → no product.
9. Every comment must stand up without the product.
10. If one line could be pasted under 20 videos, it can't be posted.

22. Core philosophy

We're not advertising under other people's videos. We find the most real, most specific pain of people dealing with the problem right now, help as real users who get it, and — when the product honestly fits that exact moment — recommend it plainly and sparingly. The long game (a trusted, un-banned account) beats any single aggressive push; each weekly review should raise the hit-rate while keeping every action human, honest, and non-spammy.