Doubts and red flags
Meta ads
Unverified, outdated, risky or contested claims per reel. Check here before acting on any tactic.
@socialmasla: For lead-gen campaigns, remove Audience Network, Threads and...
- No data shows Audience Network leads are worse for this account; CPM is not lead quality. Meta's own UI text in the recording says Advantage+ placements lowers cost per result (about 12%). Excluding placements can raise CPM (the @evanseech creator concedes "it could potentially").
- A commenter notes Advantage+ campaigns do not let you edit these settings; the tip applies to manual placement ad sets only.
- Cheap leads often come from targeting, lead-form friction (instant forms) and creative, not just placement; commenters note tracking and form design matter more.
- Meta's UI changes often (a "Value rule creation is changing" banner is visible); steps may vary.
- Client campaign names visible on screen ("Radha - Aug - Offline") may be real client data.
@evanseech: A checklist of Meta ad settings the creator turns off by default...
- No data (no before/after CPA, no spend numbers); opinion from one account.
- Some advice is account- and objective-specific (e.g. keep multi-advertiser on for static; keep reveal-details for video); turning everything off can raise CPM and lower reach; Advantage+ enhancements sometimes improve results for some advertisers (Meta's UI claims uplift).
- Meta settings keep changing; toggles named may be renamed or reset by Meta.
- Caveat on "run individual ads": Meta then picks winners unevenly, which can starve testing.
@ritesh_shridhar: A 38-second checklist of 16 Meta ads rules for ecommerce from an...
- No data, case study or account screenshots; "pure claim" authority (8 years). Rules are heuristics and some conflict with Meta's current learning-phase guidance (scaling 20% every 2-3 days and killing at 2x CPA can restart learning or kill creatives before significance).
- "2-3k spend" has no currency or threshold relation to AOV/CPA; judging by spend alone ignores purchase count.
- 70% UGC is brand/category dependent.
@socialmasla: Meta Ads now has a Reply to keywords option inside the ad-level...
- Feature rollout is partial; no evidence of performance (no cost per DM, conversion, or case data).
- Engagement-objective optimisation may deliver comment-prone audiences, not buyers (commenter's point). Meta policy risk for health/finance verticals.
- "Sent you a message!" replies suggest the creator's own funnel is a separate automation, so what Meta's native feature can actually do in the DM (flows, links, qualification) is not shown on screen.
@iamsoorajhanda: Feed Meta's algorithm four custom Conversions API events (Lead, MQL,...
- No evidence or results shown; "very easy" understates the engineering (needs CRM, event_id deduping, match quality).
- Meta requires enough event volume (about 50/week per optimisation event) to learn; MQL/SQL events may be too sparse for low-volume real estate.
- Transcript truncated; details inferred from caption.
@nicoaguirre.td: The Meta pixel works but reports incomplete data because of iOS...
- "40 to 60% data missing" is stated with no source; ATT opt-in rates and actual loss vary by account and are partly modelled back by Meta (Aggregated Event Measurement, modelled conversions). The caption "half the audience opts out" is a loose paraphrase.
- Only half the list is delivered; the real value (Part 2) is not in this reel.
@apurv_sngh: Meta's new AI agent Muse puts AI on both sides of advertising...
- Muse and in-Meta payments are described as rolled out/integrating; the creator mixes launched features with predictions (a commenter flagged this). Availability in India or for advertisers is not stated. Verify against Meta's own announcement before acting.
- "Free tokens" comment is likely a scam.
- Checklist content is unseen (behind DM).
@akshitraja: Meta's Andromeda retrieval engine shortlists a few thousand ads per...
- The claim "Meta shortlists a few thousand from tens of millions" is cited to Meta's engineering blog by the creator; I could not verify it (no network). Core mechanism (embedding-based retrieval) is plausibly accurate, but the video's simplification ("bid doesn't matter if not retrieved") is a model, not an official statement of how every placement works.
- The claim that the user vector uses "interests and demographics" is the creator's illustration; Meta has been reducing detailed targeting, so treat as schematic.
- Transcript between 30s and 60s is truncated in transcript.txt; vector-content details come from frames.
- "Meaningfully different" criterion is promised for a later video and is not in this one.
@performance.marketing.by.rohit: An EdTech brand was scaled from Rs 50 lakh/month to Rs 1.5...
- Spend figures only, no revenue/ROAS/MER or time frame; commenter correctly flags that. The brand is not named.
- EdTech with 30-40 courses and an existing brand and demand is not comparable to a small D2C account; budget at Rs 1.5Cr/month gives the algorithm far more data than a small account gets, so "one campaign" may not translate to a small budget.
- "Hygiene being met" is undefined; CBO budget skew across 50+ ad sets is unexplained (a commenter asked).
@socialmasla: Claims you can pull a competitor's Instagram followers (usernames...
- Strong red flags: scraping followers and enriching with emails/phone numbers is personal data processing without consent; breaches Instagram/Meta terms and likely privacy law (GDPR, India DPDP). Uploading such lists to Meta can lead to ad account restrictions.
- Match rate is likely low and the contact data is probably inaccurate (the screen shows mixed international numbers). Meta custom audiences require at least 100 matches and only match users whose email/phone are tied to accounts.
- No results or ROAS shown. Disclosure of an affiliate relationship is absent. Critics in comments confirm the claim "Meta can do this now" is misleading; it is a third-party tool, not a new Meta feature.
@daniel_iles: Reuse the same post/ad ID (use existing post) when duplicating a...
- 20% cost reduction is not backed with data. Commenters note that Meta now generates placement variations, and Advantage+ Creative enhancements may break pooling; also unaffected by multi-objective campaigns. A catalog/dynamic creative ad cannot use an existing post.
- Social proof pooling across ad sets is real, but whether it lowers CPA in the auction is not proven; A/B test before trusting it.
- Reusing the same post can concentrate comments (including negative ones) and prevents testing copy/thumbnail variants.
@mohdaadilofficiall: Claims a Gurugram luxury real-estate client got 2 bookings worth Rs...
- The sheet visibly says "DEMO DATA - NOT ACTUAL CLIENT DATA" (orange tag, readable around 14s and 22s) and the dates are "01 Aug - 30 Aug 2026"; the spoken claim is that this is a real client result. Treat the numbers as an illustrative template, not proof. Every stage cost reconciles exactly to Rs 2.5 L, which is typical of a mock-up.
- "800X ROI" is gross booking value divided by ad spend, not revenue or profit. A developer's sales value is not the agency's/ client's return; brokerage is typically ~1-2%.
- 2 bookings from 386 leads with 18 site visits, 0.52% lead-to-booking, is plausible for one month but a sample of 2 proves nothing.
- Strategy never revealed; the follow-up "part" is a sales call funnel. Claim of "Meta ads sell 10-20 Cr property" is a hook, not demonstrated.
@marketing.harshit: Meta has launched Ads Creative Studio inside Ads Manager
- Availability and rollout unclear (region, account eligibility, beta); the demo account has high spend (million-dollar figures), not typical.
- Last ~13s of speech not transcribed; claims about "what to do next" and any CTA are unverified.
- "Craziest update" hype; no performance results shown from generated creatives.
@daksh_manochaa: A 4-campaign Meta structure (interest CBO, broad creative testing,...
- Math: Rs 30L x 7.5 = Rs 2.25cr, roughly consistent with Rs 2.33cr, but Shopify total may be store-wide not product-only.
- Inconsistency: "2 months ago" (spoken) vs "single month" (caption); fourth campaign called "retargeting" in caption but "testing campaign" in speech; 17.40 ROAS on a test campaign is likely small-spend noise.
- Ad account blurred, no spend per campaign, no CPA/AOV/returns/RTO/COD share. Revenue is not profit; dropshipping promo, likely lead-gen for a course.
@rj_n_agency: An agency claims its first ChatGPT (OpenAI) ads test spent Rs 720...
- No dashboard or ad creative shown; lead numbers are "raw leads" by phone number, not sales; Rs 720 for 80 leads may be a one-night anomaly (small sample, fine-tuned audience, novelty).
- A commenter reports Rs ~3k spent for 2 unqualified leads, contradicting the claim.
- CPL comparison mixes different advertisers/offers/time periods; "Google is finished" is hype. The cost of the AmEx tip is a payment convenience, not a performance advantage, and gives no GST detail (a commenter claims no 18% GST, unverified).
- The vertical and exact targeting/format of the ChatGPT ads are not stated; replicability is unknown.
@sevenfiguresmp: A study of 578,000+ Meta ad creatives and $1.29B in spend found only...
- The "5%" is the speaker's rounding of a table showing about 3.7-8.8% depending on spend tier, so the "20 creatives per winner" figure ranges from about 11 to 27; he says "roughly".
- "Winner" is defined by spend share (10x the account's median creative spend and at least $500), which reflects Meta's delivery allocation, not profitability or ROAS; a creative can get spend without being profitable, and in low-spend accounts $500 threshold is demanding.
- The study is not named in speech (read as Motion from the screenshot) and not linked; the statement "creative winner rate falls with spend" partly reflects the $500 minimum.
- Andromeda blame framing is the speaker's opinion. He is selling agency services.
- Indian D2C context: study is global; costs per creative test differ.
@performance.marketing.by.rohit: Definition only
- The reel itself says nothing beyond definitions. All substantive claims (horizontal scaling is "no more valid", four levers) are in a comment and are unverified. Meta's own guidance on consolidation exists, but "no longer valid" is a strong claim.
- "Rs 5K to Rs 50K/day" has no method, thresholds (for example, 20% budget increase steps) or numbers shown. Scaling rules of thumb were not covered.
- Very low engagement, so low social proof.
@palakmehhta: Relatable founder rant
- No evidence for what her targeting, creative or placements were, so the diagnosis ("Meta is broken") is unverified. A Rs 15k test is small.
- Commenter claims (Andromeda scans copy and visuals, broad is best, ROAS 20.1x with 5 million in sales) are unverified. Several commenters are agencies pitching.
- The notification overlay ("liked your reel") suggests she may be showing engagement on organic content, not necessarily ads reach. The reel is ambiguous about whether the "audience" is ad impressions or reel likers.
@ecomtushar: A three-Business-Manager firewall setup to survive Meta's Q4 mass bans
- "Meta sued for 1100 crores" is not sourced; check before repeating.
- The "Q4 mass ban every year" pattern is anecdotal. Multiple commenters report that multiple BMs administered by the same profiles, sharing the same pixel across ad accounts, or creating new BMs can trigger restrictions. The method may go against Meta policies on duplicate or circumventing accounts; use only genuine, verified businesses and do not create fake profiles for admins.
- The method protects assets only if the ban is at the ad-account or BM level. Bans at the personal-profile, domain, pixel or payment-method level (as commenters said) would hit all BMs.
- Sharing one pixel across several ad accounts can pollute data (commenter claim); not verified here.
- "10 years" and "most foolproof" are self-reported.
@saadkhanads: Scale Meta ads by warming the account, training the pixel, then using...
- One screenshot with no date range, no ROAS, no revenue; GBP 14-17 CPA means nothing without AOV and margin.
- "Warm up the account" and "train the pixel" are not explained; the reel withholds the actual method.
- Bid caps are risky: too-low caps throttle delivery, and the one commenter says spend ignored the cap (could be a cost-cap vs bid-cap confusion). Bid cap also can cause volatile spend and is not the same as cost cap.
- 6.7k plays is a small sample; the claim that this is "the" scaling method is a pure claim backed by a single screenshot.
@hormozi: 50-5-3 ad method
- Claim of "from this to this" results: the figures shown are not labelled (metric/spend/time) and are unverified.
- 150-750 ads per week is a theoretical combinatorial count, not a tested budget: Meta Andromeda diversity benefit is plausible but each combo needs spend; small accounts cannot feed hundreds of ads. Creative fatigue vs learning-phase fragmentation not addressed.
- Dollar figure mismatch ("$100" spoken vs "$100M" on screen). Promotional funnel (the roadmap tool) is self-serving.
@apurv_sngh: Meta is rolling out agentic AI inside Ads Manager (you talk to Meta...
- Rollout details (who has it, regions, Muse naming/enterprise) are unverified here; no screenshots of the product. "Rolled out aggressively" vs caption "testing" is inconsistent.
- The "official link" is gated behind a DM; the reel is also a lead funnel.
- Likes count is near zero in the data, likely hidden/odd metric.
@praxads.ig: When you double ad budget and results worsen, nothing broke
- Numbers (2/100 and 1/100) are made up; real auctions are more nuanced (creative quality, learning phase, audience expansion). Reel offers no solution (what to do), only the diagnosis.
- Meta's help-page quote cannot be checked from the frames.
- Creator is a crypto-ads agency; relevance to D2C is generic.
@hellodeeep: In the last 10 days his Meta ad account overspent at night (e.g
- Only one anecdote; no screenshots of the hourly breakdown or numbers before/after.
- Turning campaigns off/on can reset learning and hurt delivery; better options are scheduling (lifetime budget), bid/cost caps or removing placements.
- Meta's daily budget can overspend up to ~25% on a day (and 7x weekly average rule) but not a pattern like this; cause not proven.
- "Meta ads not working anymore" is clickbait.
@ravivermamarketing: For a ~Rs 300/day Meta ads budget on a leads campaign, narrow the...
- Unverified: no result data. Commenters report poor results with this exact setup.
- Disabling the "limited spend to excluded placements" box and Audience Network is common advice but not proven for lead quality; the "6.7% lower cost per result" text is Meta's generic suggestion for radius reach, not his result.
- Transcript is incomplete; some steps inferred from frames.
@gentlerainman: Affluent buyers do respond to advertising, but not generic ads with...
- All generic assertions; no data, no case studies from creator. "Wealthy don't use referrals" is overstated in both directions.
- Funnel is for his own book/consulting business; advice is basic.
@dakota.hermes: The angle matrix
- "Angle drives at least 80% of results" and the "30 ads in one ad set means only 3 or 4 spend" are unsupported claims; Meta delivery is tied to Advantage+/Andromeda creative similarity, not literally to this taxonomy.
- Many angles will be concepts rather than distinct promises (commenter pushback valid); India CPA thresholds differ.
- Free community and "Claude skill" are lead-gen for his paid education, not independently verified.
AI tools and automation
@jens.heitmann: Opus 5.5 in Claude Code builds studio-quality motion-design videos in...
- "Opus 5.5" capability claims and "guaranteed" results are untested; quality and token cost vary, and critics call the output slop.
- Install command pulls third-party code; inspect the repo before running it.
- The prompt text is visible only briefly; details above come from frames and may be partial.
@therealbrianmark: A copy-paste Claude prompt that produces a 7-day Instagram Stories...
- Output example is a weight-loss niche with specific claims ("lose 10-15 pounds in 30 to 60 days"); copying such claims would be risky for supplements and unrealistic. Use only the structure.
- "Money printing machine" is a pure claim; no evidence that stories from this prompt convert.
- Model label "Opus 5.5" is shown on screen; irrelevant to prompt validity.
@apurv_sngh: Use GA4's Cohort (retention) report with Claude to find patterns...
- "Pattern recognition" over a handful of weekly cohorts with low volume risks false correlations (post-hoc attribution); no mention of statistical significance.
- GA4's cohort exploration is only for users with user IDs/device data and subject to data thresholds; the video does not cover this.
- Example numbers (100/20/10/5) are illustrative, not real.
- The actual prompts are not shown; value claimed but unverified.
@itstundeanthony: Agent Reach, an open-source GitHub repo, lets Claude (an AI agent)...
- Security: commenters report suspicious code in the repo; the creator did not address it. Third-party agent "skills" run with your credentials and can exfiltrate data.
- "Without any API costs" probably means free scraping backends which may break, be rate-limited, or violate ToS.
- 1,655,336 plays with 9,892 comments suggests heavy keyword farming; comment counts may not reflect real interest.
- Claim "OpenAI uses this in their official harness" is an unverified commenter statement.
@angus.sewell: Hype reel about a new LLM called Jev (written jev / model id...
- Product ("Jev", TypeSafe, "stealads") is unverified to me; claims of "724 ads in 40 s for 9 cents" come from a tweet, not tested here.
- Core claim "making so many things go extinct" is hype. Commenters note the same flows work with ordinary LLMs.
- Data privacy: sending CRM/IP/email data to third-party model.
- The reel gives no results, only ideas; prompts are given by DM so I could not check them.
- Transcript spelling "Jev" may be an auto-transcript of a different name; frames consistently show "jev-1.13.0".
@aadityabhardwwaj: Promo for Blynk Ads, an AI platform where you type a plain-language...
- Excluding neighbourhoods labelled "slum areas" is socioeconomic profiling by location. Meta location targeting is imprecise (radius pins, IP/phone estimates), so the exclusion is crude; the ethics/brand-safety and (for special categories) policy implications are not discussed. Meta's Advantage+ delivery may also override constraints.
- Quality of leads is better handled by qualifying questions on the lead form and CRM feedback (conversion API) than by area exclusion.
- No results, no pricing, no proof the tool actually changes campaigns; promo for a third party; the audience is asked to hand it ad-account access.
- Transcript is translated and partly lost after 24 s; spoken content after the keyword is silence/outro.
@briarcochran: Tier list of where AI/agents help content creators
- All tier rankings are opinion with no data. "AI is smarter than you" and "can't tell the difference" for design are claims; commenters report scripts underperform. $2,000 editing offshore figure is unspecified (per month?).
@ankit_sr: Three resources for better presentations (Chronicle HD, Deck Gallery,...
- Spoken claims on Chronicle "Apple quality" are marketing; the transcript has a gap at 33-41s. Using another creator's reel as a hook may carry copyright or reach risks (he credits it). The resource list itself was not seen (behind DM).
@priyanksinghofficial: A Claude skill (run in Claude Code, with a long master prompt) plus...
- "Equivalent to a real photoshoot", "30 minutes", and "killed agencies" are unverified claims. The visuals show generated clips but not cost, retries or fidelity against the real pack.
- Using a Lay's (PepsiCo) pack and trademarks for a demo is a licensing risk if used for an actual ad. The skill file, the guide and "Seedance 2.5" version are not independently verifiable. I cannot confirm the model name or that it exists as stated.
- The creator admits the hook is designed to "trigger the algo", so the claims are engagement-driven.
- The "master prompt" is not shown legibly, so I cannot list its contents.
@angus.sewell: Use AI to analyse retention data from paired video tests (trial...
- "Hormozi spent $4M" is an on-screen headline pulled from search, not verified; the claim "you can vibe code his whole system" is rhetoric.
- Retention analysis on tiny samples is noisy; his own on-screen output warns that findings are hypotheses ("Early data", "a hypothesis").
- TikTok analytics may require a developer API application (commenter flagged it), and scraping or saved-login automation may violate platform terms.
- No outcome numbers (views or follower growth from using the system).
@millee.md: A 4-step recipe to build a free AI video editor with the Claude Code...
Plagiarism allegation from original creator; claim "free" ignores Claude subscription/usage; Whisper/FFmpeg install steps are via AI and may break; "no gatekeeping" but still a gated upsell. Cutting every pause >0.1 s sounds robotic. Version names (Opus 5.5) not verified.
@waitwhatdidyoudo: Cloudflare open-sourced a security-audit skill for coding agents (map...
- The skit's numbers (4 confirmed, 2 needs validation, 7 rejected; "15k this week" stars) are illustrative; verify the real repo, its name and its stars yourself. The caption says it is real, but the reel does not show a real run.
- "One run finds about half of what repeated runs find" is attributed to Cloudflare's notes; unverified here.
- An AI auditing AI-written code can share the same blind spots; it is not a substitute for a human or professional pentest.
- Tool distribution via a comment-to-DM bot: install lines should be checked against the official repo.
@kanishyamo: A Human + Claude reaction reel
- Facts are only roughly right and unsourced in the video (Gros Michel/Cavendish/Panama disease/Leuven gene bank are widely documented; "artificial banana flavour based on Gros Michel" is flagged as "said to be" in the script).
- It is unclear how much of the visuals/narration is Claude vs. manual; "Human + Claude" is a claim. The "Process" CTA promise is hypothetical ("if enough people").
- Caveat: if your brand rule is to never label posts as AI, this reel's open "Human + Claude" badge conflicts with it; do not copy the badge.
@opusjake: A free open Founder Skill pack for Claude (11 sub-skills
- 100 simulated buyers are LLM role-play, not real customers; "100 real buyer scenarios" is misleading. Buy/pass counts will reflect prompt priors, not demand.
- "Trained on" Hormozi/Thiel/Jobs is persona prompting, not training.
- "Free" and "no API key" is a claim about the skill; Claude usage still costs tokens/subscription.
- Installing a skill from a link delivered via DM is a supply-chain risk (prompt-injection / malicious skill text); read the SKILL.md files first.
- Posted date field is inconsistent with comment timestamps.
@himanshu.exploresss: Same AI coding model, different outcomes
- The "two developers" experiment is likely fabricated/illustrative; no data. Model-swap claim ("difference is the checklist") is directionally plausible but unproven.
- Hindi transcript has speech-to-text errors; use the on-screen titles/comment list.
- Engagement-gate funnel flagged as low-value/scammy by viewers; do not expect the DM.
- A checklist prompt does not guarantee security; needs actual pen tests/review.
@ai.with.etqad: A GitHub repo catalogues 500+ viral AI-generated motion-design...
- The repo is never named on screen, 500+ and "viral" are unverified; quality, licensing and the model used are unknown.
- Link from a commenter is untrusted; check before use.
- "Killed motion designers" is hyperbole.
- Transcript is garbled in places (auto-transcription); Hindi text translated by me.
@theericmichaud: The #1 trending GitHub repo of the week, Reverse Engineer Anything...
- "Number one repo" and 70k stars are momentary; on-screen the stars counter shows 74,726 for REA vs 51,745/70,000 text mixing.
- Legality/ToS of reverse engineering apps depends on jurisdiction/licence; "not stealing" is a simplification.
- The DM gate for a public link is low-value friction.
- Tool quality unproven beyond anecdotes.
@pranavjoshiii_: Claude's default output for a lyric video looks like slop, but with...
- Guide content unseen; the MCP harness details are private.
- Lyrics of copyrighted songs raise licensing risk for posting.
- "Not oneshot" means time cost; no hours or cost stated.
- Transcript translated from Hinglish; "Naam Sujal" likely a transcription mistake.
@realrileybrown: You can give Claude Code (or Codex / GrokBot) a real phone number in...
- Pasting an API key into a chat (security): creator only says limits are set; use env vars/secret storage instead.
- Telephony costs, spam labelling, India calling regulations unaddressed.
- Setup is marketed as 5 minutes; real reliability unknown; the Bland-specific skill is in the DM guide.
- Voice recognition and decision errors could cause wrong changes; no safeguards shown.
@ashokreddyai: Claude (spoken as Opus 5.5) plus a Motion skill inside invideo turns...
- "Opus 5.5" and "Motion" tool specifics are unverified; link not public. Likely sponsored/affiliate for invideo.
- "Days of motion designer work" is a generic claim; quality in the clip is shown but no timing or cost shown.
- Consistency across long videos is claimed on a 38 s clip only.
@riverisbuilding: OpenFilm is a free, MIT-licensed, open-source video-editing...
Early-stage software, quality and stability untested; agent usage costs not free; "never edit a video again" is hype; one-line npx skill install should be reviewed before running (supply chain).
@robonuggets: Three steps to have Claude edit your videos
"Every cut done by Claude" is unverifiable; model names (Opus 5.5, Fable 5.1) unconfirmed; gated skill; low engagement suggests weak distribution.
@gagantalksbiz: Harvey is an open-source Claude Code agent (Python, runs locally)...
- Repo not verified; claims from the creator. Token costs, deliverability, and cold-email compliance not addressed.
- "Free" refers only to the repo; needs Claude Pro/API.
- README text partially occluded; stage details above partly inferred.
@thevarunmayya: Three GitHub repos (150k stars combined) that extend coding agents
- Star counts as proxy for quality; claims such as "ahead of most people" are hype. Verify repo contents and install commands before running; third-party skills run with agent permissions.
@backendlore: A story-driven walk-through of how an internal company chatbot grows...
- Simplified; costs shown are illustrative, not real figures. Layers are standard practice, nothing proprietary. PDF is a lead magnet.
@rollins.io: Vibe coding becomes software engineering only when you build a CI/CD...
- Claims of "easy and free" tests ignore token and CI cost (creator's own reply shows over $1,000/month); coverage of every line is not sufficient for correctness.
- No evidence; "most teams never built E2E" is an unsupported generalisation. Comment "Frontier model does this on default" suggests partial overlap with built-in agent behaviour.
Organic content and IG growth
@ryandeiss: A four-book marketing degree for the price of a few books
- Four-book list is opinionated and ad/leads-centric; commenters rightly note it is not "marketing" in a full sense (product, pricing, distribution, analytics).
- Hormozi's book at #1 and his comment may bias; no evidence for "university-level" claim.
- Breakthrough Advertising is out of print and expensive; "cost of a few books" is optimistic.
@damianappice: Use the Edits app Versions feature to make 5 hook/music variants of...
- "Over a million followers and a billion views in a year" is unverified; screenshot is selective.
- Scheduling trials is not available to many users per comments; the Versions feature may be region/version-gated.
- Duplicate-content detection risk (comment: variants get flagged and deprioritised); 25 uploads/day of near-identical content may hurt account health.
- Date-bound ("end of 2026") tactic; Meta features change often.
@robertmckeaveney: A clothing-brand co-owner explains that a pause the video at the...
- The 7M views / 15k followers / "crazy sales" are self-reported with no revenue, conversion or margin numbers; follower graph screenshot shows 16,038 (total or period) ambiguous.
- Manual DM discount flow is not scalable (he admits); a 15% site-wide code sacrifices margin; code leakage/stacking risk.
- Concept is not original (pause-frame trend), so repeat performance for other brands is uncertain.
- Instagram may treat "screenshot DM for discount" as engagement bait; keep within Meta promotion rules.
@itshardikway: Claims that boosting posts trains Meta to throttle your organic...
- No evidence that boosting suppresses organic reach; Meta has not stated this. Correlation is likely (boosted audiences are less engaged, lowering engagement rates, which can depress later organic distribution). Treat as unverified.
- Mixes follower count with sales; boosting for conversions is different from boosting for followers.
@brock11johnson: Brock reacts to another creator who near-verbatim recreated his reel...
- The 40K vs 1.3M numbers are from a reply, not on screen; the counters on screen show near-equal early growth, so the comparison is partly staged.
- "Welcome to copy" is the creator's own policy, not legal advice; credit norms differ and brand work should not copy others' IP.
- Commenters dispute his own originality.
@tanmaythinks_: If starting an Instagram page from zero (hypothetical HairWolf...
- Clipped; only two of "3 or 4" metrics appear. The claimed priorities (AVD, shares) are plausible but not evidenced here.
- "Hypothetical page" example, no results shown.
@brock11johnson: Stop posting 3-second read the caption reels with no value in the...
- "Worked really well in 2024" and "most viewers find it annoying" are unmeasured opinion; no data. The commenter anecdote (3,300 -> 9,800 followers) is not attributable to this single change.
- Follower counter overlay (889,643 +128,927) is unclear (meaning not explained); do not treat it as verified.
- Repost + comment CTA is an engagement-bait mechanic; Instagram can deprioritise engagement bait in some contexts.
@jimruitang: Max-follower formula
- 27-second video shown as source of 75k followers; the exact number differs between speech (70,000) and screenshot (75,144). One outlier is generalised into a "mechanism"; survivorship bias, plus reported conversion (3.7% views to follows) is atypical.
- Case-study view counts are cherry-picked; no failure rate. Comments include creators for whom it failed.
- Partial contradiction: "credible vs impressive" vs underdog; the quest advice may force manufactured narratives; he himself says it feels fake.
@thethinkschool: Teaser clip from a Think School podcast
- The method is never shown; the whole clip is a teaser. Numbers (41 posts, 6.8M views, 5.6M, 1M followers in six videos) are unverified, no handles shown, view count of one niche (interior fit-out) video.
- Auto-transcript says "first deal" vs on-screen "first reel"; "Kumar method" is an unclear reference.
- Sponsored (Odoo) and the guest sells courses/services; survivorship bias (six videos that worked are shown, not the failures).
- "Six videos" claims can mislead: a page with 6.8M views in six posts may reflect one viral outlier, not a repeatable system. Commenters note guest has ~140K followers himself.
@brock11johnson: An 8-step hidden emoji Instagram Story recipe (reset the Stories...
- The headline results (7k to 104k views, 180+ leads) are from a friend and a mastermind member, with screenshots shown but not independently verifiable. The "you only get more views" claim is stats on follower reach (99% followers), so it does not grow non-follower reach.
- The "reset your Stories algorithm by not posting for 24 hours" step is an unverified folk theory. Instagram has not documented anything like it.
- Paid ManyChat sponsorship means the claim "this only works with ManyChat" is a sales claim. Other DM tools exist.
- Hiding emojis plus prize language may need IG promotion guidelines (giveaway rules) compliance.
- The caption says "free tutorial", but the funnel is a ManyChat signup, with pricing concerns raised by a commenter.
@thehayleyjennings: Find proven viral template reels in your niche, remake them with your...
- "9/10 go viral" and "90% chance" are unverified, survivorship, and the author sells the guide. Her first reels at 44k-76k are probably also helped by being new and by trend audio, not just the method.
- Copying another creator's hook and audio is a content-originality and (for music) a licensing risk.
- Claim that Edits connects a "digital footprint" that boosts reach is speculation; Instagram has not stated this.
- Trial-reel tricks tend to decay as everyone copies them.
@brock11johnson: Post every reel five times
- Unverified: no data given that 5x posting increases net reach; may reduce trust with followers/ Instagram originality signals.
- Commenters flag shadowban/duplicate-content risk; creator answers are assertions, not Instagram documentation.
- Mirrored video can distort on-screen text/logos; "everything" is overbroad (low performers waste effort).
@iamchrischung: Winner-iteration loop
- Tiny sample; no data beyond the screenshot, and the screenshot numbers are not clearly tied to the iteration chain (may be unrelated reels).
- "You get another winner" is stated as a certainty (survivorship bias).
- The video is just 14 s with no CTA; low engagement may indicate low resonance or a rough cut.
@brock11johnson: A 17-second here, not here gag that packs four reel-production rules
- Title/subtitle placement is a style opinion, not tested; the creator provides no retention data.
- The "ideas from AI, not here" line is a pure claim.
- 439k plays largely driven by the "Harry Potter" novelty and the format, not necessarily the advice.
@rapvilleuk: Use Instagram Trial Reels (shown to non-followers) with three rules
- "30k in 7 days" is one case; the follower numbers are screenshots, unaudited. Audience pushback shows reach can drop.
- Claim that IG "doesn't detect duplicate content when the hooks are different" is unverified; IG's originality system uses visual/audio fingerprints, and re-edits may still be flagged.
- Trial Reel availability and thresholds changed through 2026 (commenter under 300 followers cannot use).
- The transcript line "pull two interviews" is a mis-transcription.
@ai.noobstoninjas: Her 0-to-100k Instagram method
- Tool pricing, name/URL and data source (IG scraping legality) not given; likely a promo.
- "0-100k" relies on one account; no timeline.
- Recreating competitor topics risks copycat content; "own take" is the safeguard.
- Transcript translated; "OpenLive Studio" in the transcript is a mis-transcription of "Open in Studio".
@august_carlsson: A creative director applies for a job in public
- "No strategy" verdict on a 10M+ audience account is opinion with no data; plan mostly adds two borrowed formats.
- Blunt tone ("stuck in 2016", "horrible") drew the heaviest pushback; commenters say it repels decision makers. One reply is survivorship.
- Claim that the target already changed his bio is unverified.
Lead gen, sales and client acquisition
@joybaik.global: Good cop / bad cop negotiation
- Entirely self-reported with no named deal or numbers; she is selling negotiation coaching.
- The caption looks like pasted LLM template output ("Title/Description/Hashtags"), and a commenter says the script looks AI-written; treat the story as a polished illustration rather than a verified anecdote.
- Tactic assumes she holds leverage (as commenters say) and works poorly against sophisticated or powerful buyers; it can damage trust if the planted "tension" is detected.
- Mild ethical risk: manufactured conflict and the "customer believes they won" framing read as manipulation to many viewers.
@calebchae: Live speed to lead test
- Possibly staged or sponsored; one instance proves nothing about conversion; the 60-second rule has research behind it elsewhere but this reel gives no data. The 5-second callback suggests automated dialling or an AI/auto-dialler, which commenters suspected; the creator denies it.
- "$1.4 billion" is the creator's claim.
- Contradictory advice in comments on immediate calls reducing trust; no controlled evidence.
- Publishing a recording of the company agent ("Alex") without apparent consent may raise privacy/recording-consent issues in some jurisdictions. Also the lead used was a fake test enquiry that consumed the company's sales time.
@arshdeep_sohal: A wedding/events videographer's Rule #1 of freelancing
- The transcript is too garbled to confirm the exact claim or any number; the speaker may be saying the opposite in parts (that he never refused work). The headline rule comes only from the on-screen text.
- Blanket "never work with small clients" ignores portfolio-building and referral value for early freelancers.
- Conflict of interest: creator is advertising wedding services; no evidence offered.
@theguy: A 3-step Instagram Story template (frame it as a rare one-off,...
- Only a views screenshot is shown; no leads, opt-in rate or revenue numbers. "$100,000 offers" and 118 hours are self-claimed.
- Inconsistency: story says 90+ hours, speech says 118 hours.
- Fulfilment glitch visible in comments (email not received). A $1 charge adds payment friction/compliance (and in India would need a gateway) and may attract low-intent buyers.
- Claim "works for all businesses" is unproven; frame "craziest thing I've ever done" is manufactured scarcity and can read as inauthentic for a luxury brand.
@amirsyed: When following up with wealthy/high-status buyers, never sound like...
- Pure assertion, no data; "high status" framing is generic. A valid pushback is that style alone does not hide need; a full pipeline does. Follow-up with no ask can also lose deals when a clear CTA is needed.
@rishabh.biswas_: Run Meta ads for your own agency straight to a Calendly booking page...
- 5.5 lakh in 3 months is revenue, not profit; ad spend unknown (most-asked question, unanswered).
- Only one GBP payment shown; guarantee language in the example ad ("revenue guaranteed in 60 days") is risky/compliance-sensitive.
- Calendly pixel can only track booking if the page allows script/Pixel (paid Calendly tier); not stated.
@apurv_sngh: Export your LinkedIn connections CSV, have ChatGPT/Claude score fit...
- Uploading personal contact CSVs to AI/third-party enrichers raises privacy/GDPR and LinkedIn ToS concerns.
- "20,000 to 5,000" is illustrative; LinkedIn exports include only name, URL, company, position, connected date, and emails only if the contact shared one (so "company size" in speech is wrong without enrichment).
- Promotes his own product (Jev); enrichment costs (Clay/Apollo credits) not given. LLM scoring accuracy on thousands of rows is unvalidated.
@fateh_and_nimrat: Five client-acquisition systems (cold email, LinkedIn outreach, Meta...
- Numbers are not outcomes: no leads, reply rates or CPL. "15,000 volume" for $200 ignores deliverability and spam-law risk (cold email).
- 500 LinkedIn outreach per account per month exceeds typical safe limits and risks restrictions; Meta "$500 start" is arbitrary.
- Transcript garbled; Quo spelling inferred from frames.
@amirsyed: Insinuation
- "$100M+" is an unverified claim; the example is generic and not tested. Planting loaded questions can feel manipulative and backfires if the competitor can answer them; use only where truthfully differentiated.
@highonsetting: Claims a three-part framework generated over $500,000 for a client...
- $500K "for one client" with no niche, timeframe, offer price, ad spend or screenshot beyond a friendly DM message is unverifiable; the DM screenshot shows only a client thanking them.
- Low reach (1.6K plays) and a coaching/setter-course positioning in the bio; pillars are generic ("sales psychology" undefined).
- "No phone closers" claim ignores that DM closing is still sales labour; not specified how payment/links/handoff work.
@sobusells: A website that generates leads is not enough for service businesses
- Demo is clearly mock/sample data (names, "Miranda"); no real client results or numbers on revenue lift. Tiny reach and no social proof. "Retakes" old leads claim is unspecified.
@stfu.nittish: Case-study reel
- Ranking in 2 days is plausible only for low-competition long-tail terms; no keyword volumes or before/after data shown. Auto-generated replies, "bot" engagement. Transcript here is poor quality; details taken from on-screen captions.
@promptedbyaadil: Build your own local-business lead scraper with Codex (Google Maps,...
- No dashboard numbers are explained; reply or positive-reply rate may be misleading without offer, list size and deliverability context.
- "Free" ignores Google Places API cost (if used), proxies, verification fees and sending infrastructure.
- Google Maps scraping may breach Google's terms; cold email to scraped lists carries legal and deliverability risk (GDPR/PECR in UK/EU).
- Not shown: the actual prompt used or the code, so "build it with Codex" is not reproducible from this reel.
@imheadrick: Harvey is an open-source Claude Code agent (GitHub...
- Unverified: I have not seen the repo; the reel shows README only. Autonomous cold emailing carries spam/deliverability and CAN-SPAM/GDPR risk; Gmail/Workspace sending limits could get the mailbox flagged.
- "Uses your Claude subscription" for automation may conflict with Anthropic usage terms; check before relying on it.
- Lead source quality (US local businesses via free source) is untested; India coverage likely weaker.
- Treat repo contents as untrusted code before running.
@theryankeip: When a prospect says I can get it cheaper somewhere else, do not...
- Anecdote-based; no conversion data. Works in consultative, discovery-led sales; weak for impulse D2C or when buyer is price-only. The silent "you probably can" can lose low-intent leads (see top comment).
@the_contentstrategist: Agencies should steal three consulting habits to win pitches
- The "CEO leaves after 15 minutes" and delivery-vs-discount stats are explicitly illustrative; no data.
- Generalisation of "consultants better than agencies" rests on one person's experience; the creator is an agency-side strategist selling the idea, not a consulting benchmark.
@kailodebiz: Alex Hormozi's Acquisition.com $100M Scaling Roadmap quiz is a...
- "Deemed illegal by the government" is clickbait and unverified: the cited item is a class action about Hims & Hers (securities), and the FTC claim is not shown; a commenter says it's not illegal. The quiz is not itself the subject.
- "25% of 600 ads" and funnel logic are based on an observed screenshot and may change; collecting phone/address has privacy implications (India DPDP act/consent).
- Creator sells quiz funnels, so the framing is promotional.
Business strategy and founder lessons
@gunpowder.doshai: Ambush marketing (her definition) means a brand or its agency hires...
- Her definition of "ambush marketing" differs from the standard marketing definition; the term is used loosely.
- Single second-hand anecdote (a friend's agency was approached); no evidence, no brand named. "You can tell they've done it" test is explicitly unreliable by her own admission.
- Accusation-by-implication against an unnamed competitor; a commenter claims the competitor is reposting the reel, unverified.
- Planted-review services do exist, but the reel gives no scale or data.
@build.sagar: The viral pitch buy the same product on IndiaMART for Rs 40-45 and...
- Evidence is mostly screenshots and analogy; no actual P&L from either side. Numbers are illustrative.
- The Google/news analogy is about editorial SEO, not product listings; the logic is weaker for marketplace resale (resellers do profit on some categories).
- The accusation against a named person is unverified.
- Transcript is incomplete (gap near 20-40s and cut off at the end), so parts of the argument and the exact closing line are not captured.
- Creator also sells guidance (follow ask), so there is an incentive to discourage rivals' methods.
@extraordinarywithchirag: A tough, perfectionist luxury-marketing boss (Sulakshna Pathak, BMW...
- Anecdotal; one-sided gratitude story; some viewers see "toxic leadership" being glorified. Survivorship bias.
- Tag of a senior executive and company logos (BMW/MINI) is used for reach; no corporate approval implied.
- The "Teachers Day" tribute is partly personal-brand marketing for a career-coaching business.
@professorpanache: Say no without a long apology or alibi
- Opinion piece; "never give reasons" can backfire with customers/clients where context builds trust; works better for personal than commercial relationships.
- Gendered framing ("a gentleman"), limited relevance for some contexts.
@theaustinschneider: A 10-item checklist of the operating systems an agency needs to go...
- Income claims ($200K+ months, $10.3M tracked results, 650+ owners) are self-reported and promo.
- Generic checklist, no how-to; the real product is the paid program.
@buildwithjigar: A customer needs to hear about a brand about 8 times, consistently,...
- The "8 times" number is a loose marketing rule of thumb (cf. the "7 to 8 touches" myth), not a measured constant; it varies by category and consideration.
- Impulse-purchase categories and word-of-mouth brands contradict the claim, as commenters note.
- Transcript is truncated; the example is incomplete.
@vicktipnes: Mindset and beliefs are the real bottleneck (not strategy, product or...
- Spoken content and the caption are different pieces of content: the caption's "60 customers a day" math is not in the video. Treat the math as the creator's caption only.
- Self-reported growth ("40% average growth over 8/9 years") is unverified and he states eight and then nine years; the comparison to "most companies grow 6-7%" is unsourced. No business or industry is named.
- "Every specific goal we've set, we've hit" is a survivorship claim.
@thesoloentrepreneur.yt: Idea pitch
Rs 800 margin ignores shipping, returns, payment fees, framing, ad CAC (Rs 25K for 120 orders is about Rs 208 CAC, optimistic). 120 to 1,500 orders growth and Rs 18 lakh claim are pure projection ("Rs 96,000 tools and ads" is mis-transcribed). The "Rs 400 to print" is unsourced. Competing Indian map-poster sellers likely already exist. Creator is selling the plan, not running it.
@the_shikhshit_talks_: Spiritual law of money monologue
- "Money likes peace, so be calm and money comes" is unverified and survivorship-biased. The commenter with a rude boss who owns a Dubai villa is a direct counterexample.
- The line "money won't come however hard you work" if you fight is a blame-the-victim framing that can mislead people in financial difficulty.
- Do not treat this as financial or business advice. The comments about manifesting an NYU place are anecdotes only.
- The transcript was auto-generated in Urdu script from Hindi speech. The translation here is my reading, and exact wording may differ slightly.
@ayoships: Decompiling the free game Block Blast, the creator claims an...
- Unverified: the code claims cannot be checked from the video; no full evidence yet (promised "soon"). Code could be out of context; names like "Death Algorithm" are the app's own trait names but the intent interpretation is the creator's.
- Revenue figures ($127M) are from a third-party estimate chart, source not named.
- Possible legal/ToS risk for decompiling; not relevant unless you are replicating it.
@hormozi: Ten one-line business rules (sell premium to the rich, do it faster,...
- Pure assertions; "sell to the rich" and "twice the price" are survivorship-biased and do not fit every category (rule 9 contradicts the common middle-market D2C approach).
- The video may be AI-generated or a clone (commenters suspect it); the "free roadmap" is a lead-gen quiz, so rules are marketing for the funnel.
Scripting, hooks and storytelling
@ksltos: A meta-reel that teaches the The sound of X vs the sound of Y / The...
- View counts of the example reels are shown on screen but the source accounts are not credited; cannot verify they are the creator's own or others'.
- "1M views caption" claim is about the hook text, not proof the hook causes the views (selection bias).
- The "X vs Y" format partly relies on trending audio/visual novelty that decays quickly.
@samkul773: Ten avoid -> say this swaps for marketing/sales copy (e.g
- "Never say" absolutism with no evidence; many pairs contradict each other in practice (e.g. "limited time offer" is the same as a discount).
- Buying-psychology gimmicks: "3 spots left" can be deceptive if untrue (consumer-protection risk in India under the Consumer Protection Act / dark patterns guidelines).
- Gated training content unknown; creator profile engagement looks low (3 likes).
@brock11johnson: Visual gimmicks (props, outfit changes, shouting) are not hooks
- The creator's own reel works partly because of the gimmicks; claim "gimmicks do not hook" is contradicted by his results and a top comment.
- The 25-hooks list content is unknown (not shown).
@ayeshargoyal: Role-play (if I were a snack company, here are three legal tricks to...
- Thresholds quoted in the caption (sugar free up to 0.5 g/100 g, low sugar up to 5 g, "50 names for sugar", nutrition claim rules) come from the creator; verify against current FSSAI regulations before repeating or using in marketing. The claim that jaggery's minerals "do not matter" is a simplification.
- The reel is a dramatization with a made-up cookie; no real brand or product tested.
- The "#psychology" framing and the word "manipulate" are intentionally provocative; the legal claims ("all legal") are not proven for every trick.
@viralrolloutlabs: A 13-second vocal-delivery demo
- The causal claim (voice rhythm makes videos go viral) is unproven; the success of this reel may owe more to the open loop, the very short length and the controversy than to delivery.
- The sentence "go viral... learn how to speak" is generic; many commenters call the advice generic or disagree with it.
- The "Flow" keyword mechanic is inferred from comments only; the freebie content is unknown. Creator is selling growth services (link in bio).
@briarcochran: Script template Hook, Explain, Illustrate, Apply
- "10x retention" claim has no data. "Edged through the video" is manipulation-flavoured; more is not better, overstuffing slows pace.
- Timing labels read from small text, treat as approximate.
@adswithfaizann: Five ad-hook frameworks with a handwritten example each
- The transcript is truncated (Whisper output has only two 30s blocks, cut mid-word at 30s and 60s), so hook 3, the end of hook 5 and the closing lines are missing. The frame text shows the details; parts of the handwriting are hard to read (for example the "I have 10 more" end slide).
- Claims like "insane results" have no data. The examples ("15 kg in 30 days", "1 Crore sales in a month") are exactly the claims that regulators and platforms flag as misleading if not backed.
- Wording of hook 5's validation step is my reading of a cut-off sentence.
@izhaan.os: A 19-second vocabulary quiz of 13 retention/hook/storytelling terms...
- The reel contains zero teaching; value claim is entirely in an unseen DM. "Goldmine theory" is not a standard named concept; likely his own label.
- "Zygmuntic effect" is a mis-transcription; the spoken word may be Zeigarnik. Unverified what he says.
- Engagement may be inflated by the giveaway mechanic (comments are not interest in the topic).
@adley: Use a physical prop to carry the rules of the game as an on-screen...
- "Billions of views per month", "18B/month" and "people will watch videos they HATE" are unverified self-claims.
- The mug is never shown paying off; the lesson is presented without any result data.
- Caption ends with "- ad": it is a promotion for the audit tool; the tool's quality is unknown.
- Wording/colour critique contradicted a bit by "red for 25 pounds" being a hindsight comment, not tested.
@howithooks: P
- Barnum anecdotes come from his own memoir (admitted by the creator); numbers are likely exaggerated, and the "Egress" sign story is a well-known tall tale.
- The tool mix is unnamed.
- Few conversions evident (no CTA), so value is reach/brand only.
Ad creative, UGC and hooks
@socialmasla: Same ad script, same performance-marketing course offer
- Single anecdote, no metrics, no control for audience/placement/creative differences; "performed good" is unquantified.
- Gender-as-hack framing can attract low-intent traffic, stereotyping and brand-trust risk; creator himself acknowledges the comments were off-topic.
- Transcript is garbled; claims above rely on the on-screen text, caption and comments.
@thejacobroach: Four native-looking Meta ad formats that convert (zero-star review,...
- Title promises 10 ads, only 4 given. No conversion data ("convert really well" is an unsupported claim).
- Fake-news and fake-thread ads may violate Meta's deceptive-practices/misleading-content policies; ad account risk.
- At least one user did not receive the list; delivery reliability unverified. The formats are heavily used already (commenter: expensive auction).
@architmehrotra.in: A comedy skit of a bootstrapped founder phoning an unqualified lead...
- Transcript is truncated (Whisper dropped the middle); translation of the "ghar girvi rakh ke" line is approximate.
- No data or solution given; the reel is entertainment, not a how-to.
@adsandcopy: The most viral Indian ads used only three product-integration patterns
- "Used only three ways" is a tidy generalisation drawn from three examples.
- Paid-media brand ads with big budgets and celebrities do not transfer directly to small D2C ads; the creator's actual process is behind the DM.
@adsandcopy: Teardown of supplement brand The Stack's Meta ads
- No performance data for The Stack's ads (spend, longevity, ROAS); "doing crazy" is inferred from their creatives alone, not from numbers.
- Supplement/health claims (doctor says poor sleep reduces lifespan; testing/study claims) are regulated; copying them could trigger ad rejections. The link delivered in DM is unverified.
@rudygoel_: A rapid list of 10 marketing concepts to use as creative angles for...
- Pure name-dropping; some labelled "concepts" are cognitive-bias terms rather than proven ad tactics. Claim "Meta treats look-alike ads as one ad" (Andromeda similarity grouping) is plausible but unverified here. Transcript errors ("Bevel Cow").
@omar.creo: A real thumbnail and title ideation session
- One winning case with a single test; watch time share across three variants proves little on its own (small deltas such as 37.8% vs 34.2% can be noise).
- The $1.2B Grüns exit is cited from an Axios Pro scoop (headline only).
- Staged dialogue, so "a window into our process" may be polished re-enactment.
@adswithsimon: Sell the identity/emotional outcome, not the product benefit list
- Examples are borderline/non-compliant ads (shaming, sexual performance, personal-attribute implications); likely to trigger rejections or account disables, as the one comment warns.
- No performance data shown to prove the "better way".
@omar.creo: A live YouTube packaging teardown
- The before/after chart is a single case with no baseline for seasonality; attribution to title plus thumbnail is unproven.
- "Solved" claims can be seen as clickbait; the guest himself ("too much sauce", "it's a lie" comment) hints at overclaiming. Not applicable verbatim to jewellery.
D2C and e-commerce conversion
@itsecomkid: A US dropshipping mentor reports that his India-based student has...
- Entire data point is from a single student; Indian operators in comments contradict it (RTO ~10%, delivery rate ~80% not 40%).
- Speaker is a US dropshipper not operating in India; numbers are second-hand.
- "5-10 ROAS is normal in India" is unclear (blended vs platform ROAS; dropshipping margins).
- COGS "less than $1" and shipping "20-30 cents" are unverified.
@saltandpepsam: For high-ticket Shopify orders, trigger a Zapier on every abandoned...
- "Made me millions" and "60% of business" are unverified self-reports; no conversion rate, call cost or cart-volume numbers.
- Needs phone numbers on abandoned checkouts, which Shopify only has if captured; calling without consent breaches TCPA/DND rules (US/India); the creator does not address this.
- Draft orders bypass some checkout fraud protections and may affect reporting/attribution (orders show as draft-order channel).
- Sample may be niche-specific (high-ticket).
@seemaa_choudharyy1: Users drop between Add-to-Cart and Initiate Checkout because impulse...
- No data or benchmarks; the veto framework is the creator's model, not a validated study. Veto 1 could not be recovered (transcript cut at the segment boundary; frames show no text for it).
- The claim that the cause "is not" persuasion is absolute; drop-off also comes from shipping-cost surprise, forced account creation, payment failures, none of which are mentioned.
@shovel.studio: Small catalogs (under 50 products) win with less choice (Yeezy
- All statistics (10x sales under 50 products, 26% more emails with 6-10s delay, 76% filter use, 20-30% model-shot lift, up to 40% AOV without express checkout) are stated without source; a commenter asked for sources.
- Yeezy hiding shipping times is both a conversion and a trust/legal risk (long ship times; commenters report months). Removing Apple Pay/Shop Pay generally reduces conversion on mobile; "up to 40% AOV" does not mean more revenue.
- Survivorship: Yeezy and Skims are celebrity brands with huge direct traffic; tactics may not transfer to unknown brands.
@ecomtushar: A brand-new product with no pixel data hit 44% prepaid on day one in...
n=9 orders, 4 prepaid: statistically meaningless, one day only. Launch day of ads may be warm audience. Discount cost (10%) vs RTO saving is not computed. Product (IPL hair remover) is a high-AOV, trust-heavy category and may not generalise. Template is a lead magnet, not shown.
@sobusells: There is no single best website layout
- No data behind the claims; the chart shown is unsourced. The binary framing ignores testing. Audit offer is a pitch, not evidence.
@amboras.ai: A top X% of stores ladder of product-page conversion features
- The "top 80/50/30/20/10%" figures are not based on any shown data; they are a rhetorical device. Do not cite them.
- Tactics like countdown timers, fake urgency and "price discounted" can be misleading if the timers or strike-through prices are not real (consumer protection rules in India and the EU).
- The creator is a product vendor (Amboras), so the "tutorial" is a sales funnel; one testimonial-style comment may be seeded.
- Free gifts "worth more than the order itself" can destroy margin. No margin or AOV data is shown.
@edward.builds: Seven before/after landing-page A/B tests show the same pattern
- Case-study numbers are vendor/marketer anecdotes from public CRO write-ups; single-test results, selection bias, mostly 2010s landing pages. Sample sizes not shown.
- Speech and caption disagree: Imaginary Landscape "doubled" vs +160%; Career Point "more than quadrupled" (3.12 -> 13.64 is 4.4x ok); Humana "over 15x" (+1,456% is ~15.6x).
- Removing navigation does not always help (ecommerce category browsing); results apply to single-goal landing pages.
- Video funnels to the creator's own site; the 10 examples are not visible in the reel (only 6-7).
@thinkwithsid: Meta's Muse AI shopping assistant plus the Meta-Shopify deal (agentic...
- "90 days to save your business" is hyperbole; no source for a deadline. Meta Muse dates and Shopify-Meta deal details are only shown via screenshots; verify rollout region (India uncertain; "Shopify Now Enrolls Your Store ... default" opt-out vs opt-in matter).
- "Not many images" contradicts general PDP best practice for human buyers.
- Gated guide behind comment keyword.
Video production and editing
@1cho.mejia: A sound-effects breakdown of a travel edit, labelled shot by shot...
Selling an SFX pack; one buyer complaint. SFX packs are widely free. No numbers on conversion or price.
@softgirlnocode: Three ways to make AI-built (Claude) videos look different
- Open-source licences vary; check licences before using in client or commercial work.
- Suno music licensing for commercial or ad use depends on plan; commenter notes download limits.
- Transcript spells "Claw/Claude"; no proof of time saved or performance gains from the "different" look.
@vibha.in.progress: One talking-head clip is re-edited into 7 different styles (glass...
- Claims "no editing app" but edit quality likely depends on extra tools (face tracking, rendering) that are not shown running; actual setup time, reliability and cost unverified.
- Counts inconsistent (7 / 36 / 42 effects). "Free starter versions" implies paywall for the good ones.
- Running
git cloneinto ~/.claude/skills of an unknown third-party repo is a supply-chain risk; I did not run it. - Spoken audio is a song, so all claims come from caption, frames and creator replies.
@chandlerintelligence: A reaction-style walkthrough of a Claude-edited video (by another...
- Much is a reaction to someone else's work, so the "I didn't edit" claim refers to that video; the actual code and prompts are behind a DM.
- "Free for $97 thing" is a competitor-baiting hook, and credit disputes appear.
- No timings or cost per edit; the workflow needs coding familiarity (Remotion, Node, FFmpeg).
- yt-dlp downloading others' reels for a wall carries copyright / terms risks.
@patchdstudio: Radio-style mic processing is six effects in this order
- Radio chain order contested (de-esser typically after compressor or before depending on style). Sample demo flawed (added fan noise).
- Compressor "brings voice to the front of the mix" is oversimplified.
- This is marketing for an app in beta; pricing unknown.
- Social-video dialogue often should sound more natural than radio-style (commenter).
@justyn.ai: Claude edits the whole video from raw footage via one-sentence...
- "Didn't edit a single second" is only partly verifiable; the demo video itself was filmed and prompts shown are scripted.
- No details of the actual tool/skill, the cost, or render time are given; the walkthrough sits behind the gate.
- Output is recognisably "Claude look" per comments; scale of manual fixes unknown.
- Waitlist suggests a product/lead-gen funnel, not free guidance.
@aevyvideoschool: Claude can generate full motion-design and map animations in about 30...
No proof of the 30-minute claim beyond clips; the course/webinar is the actual product. Comment count inflated by keyword mechanic.
@corbinsylk: Slash-command-per-effect system in Claude (e.g
Prompts are gated; effects quality unverifiable; no workflow shown for importing into an actual edit.
SEO, AI search and websites
@codvyn: Four sites to feed an AI coding agent with templates and references...
- Several sites are template galleries; title and 3D showpiece (Active Theory or similar studio work) are not products of the creator or these tools, a commenter points out.
- Performance, SEO and accessibility of 3D sites are not addressed; commenters are right that these hurt conversion for e-commerce.
- Transcript cut off; details of tasteskill.dev and threeui.com claims are from captions and may be incomplete. I did not visit the sites.
@casarthakahuja: The creator says Trump is replacing AI with Super Intelligence (SI)...
- Central claim unverified: I could not find any official US government statement; no source shown. Treat as likely false or exaggerated; do not act on it.
- Commenter saw a .si ad "today only", suggesting possibly sponsored/undisclosed promotion.
- OpenAI.si / OpenClaw.si purchases are not shown. .si does not become valuable because of an "SI" rename; domain speculation is risky.
@sidgnificant: A 1-month-old agency site gets cited in ChatGPT (and reportedly 16...
- "ChatGPT reads only raw HTML" is overstated: text inside an accordion that is present in the HTML source is readable (a commenter says this and the creator sort of agrees); the real failure case is content injected by JavaScript. "90% chance" is invented.
- 16 organic ChatGPT leads in one month for one site is a single anecdote; the query is a narrow niche with few competitors, so results may not generalise. ChatGPT citations vary by user/session/location and over time.
- "Almost Rs 2L a month for free" on-screen claim is never supported in speech.
- Auto-transcript is garbled; the hook probably means "watch and delete before competition rises".
- The Highprime homepage line "From sex toys to skincare, we've sold it all" appears on the site, not a client-safe line to copy. The "4.3 average ROAS" is self-reported.
@bodhigallo: Tier list of local-SEO website mistakes for service businesses...
- "H1 is the first thing Google crawls / most important" is an oversimplification; title tag and overall content matter, as a commenter notes. No data or case study.
- Tier placements are inconsistent with the title ("mistakes"): listing too many services is A and 500 pages is B; "reviews S" is more a GBP factor than a website factor.
- "500 service area pages" can look like doorway/thin pages; the risk is not mentioned.
- "Blogs not important in 2026" is a claim without evidence; also he later says blogs work if local and linked.
@edward.builds: AI Overviews and LLMs can be steered within hours by a cheap press...
- Works for an invented, zero-volume query; no evidence it transfers to commercial keywords (commenters note this). Results are volatile and may be patched. Press-release spam and fake-event seeding is ethically and possibly policy risky (Google spam policies). The reel mixes warning and tutorial; the link is a funnel.
@edward.builds: Google's Search Central page Google Search's guidance on using...
- Posted 3 Oct but says "yesterday on October 1st"; minor date inconsistency in the caption/script (possibly filmed 2 Oct).
- Wording like "critical to manually fact-check" appears only as quoted from on-screen; not independently verified against the live Google page. Treat as unverified until the doc is opened.
- "Gives Google one more reason to take out websites" is the creator's inference, not Google's statement; commenters correctly note the guidance mostly restates prior policy (scaled content abuse).
@stfu.nittish: Case-study clip
- Unverified claim: top ranking in 3 days for an unnamed site with no keyword/position shown; 49 -> 97 performance is a lab score, not rankings.
- "Agentic browsing" score tool is not named; AEO/GEO "completed in two days" is likely a checklist, not durable results.
- Inconsistent date math ("25 days" vs 5 days); auto-replies from the account are generic and sometimes off-topic; low social proof (3 likes).
@apurv_sngh: A Claude Project SEO system
- The PDF is behind the DM; the file contents shown are partially illegible, so file names/descriptions here are reconstructed from blurry frames (flagged).
- Transcript says "cloud"; interpreted as Claude from visuals.
- Claim that AI quoted a free blog over a brand page is anecdotal; "Claude project capacity" shows files are small; no results/traffic numbers given.
- Contradiction risk: promoting an AI-based SEO system but stating no AI content generation for clients (research/monitoring only?). Unclear what the scheduled agent publishes automatically.
Personal brand and creator business
@therealbrianmark: A top X% ladder for an Instagram profile built to sell coaching
- The "top 90%...1%" percentiles are invented; no data.
- Coaching-sales template; many steps do not transfer to D2C products.
- Social-proof line like "$200 million in revenue" must be true; example only.
@vyre.less: A page started from zero followers on 1 July 2026 and made Rs...
- Rs 5,74,000 unverified: no invoices shown; could include barter/product value.
- Playbook content not seen; quality unknown. Pricing "500 or 5000" framed as unknown may be a staged hook.
- Comment engagement is mostly keyword-gated, so reach/interest is inflated.
@accentcoachrahtish: For Indians selling to Western/international clients, a...
- Unsupported claim, no evidence; creator sells accent coaching so there is a conflict of interest.
- Trust is better built by proof (portfolio, testimonials, video calls); accent over-correction can feel inauthentic. Generalises scam stereotype.
@therealbrianmark: Seven-point Instagram profile checklist for coaches selling...
- $1.5M/month claim is unverified (the revenue card is a screenshot; gross, not profit). Blue check cost and benefits vary by country and are unproven for conversion. Advice is for high-ticket coaching, not D2C. "Take out Linktree" ignores shop/product traffic.
@samstoffel: Followers are not money
- Unverified: $50M, eight-figure exit, "most influencers are broke, I know from DMs" (anecdote).
- Advice is generic; the actual hard part (finding a differentiated solved problem) is skipped, as commenters note.
- Possible AI/staged visuals; the setting may not be his real garage. Wealth-flex funnel selling "get rich" coaching is a standard risk-claim pattern.
Pricing, offers and positioning
@dhananjay.scale: Market sophistication has 5 levels (claim, unique mechanism, proof,...
- Framework is simplified and not attributed (Schwartz's levels are 5 stages but differ in detail: stage 4 is "mechanism", stage 5 is "identification", and guarantees are not a stage). Presented as fact without evidence.
- No data or case study; the claim "sales explode without any marketing" at Level 1 is illustrative only.
- Transcript segments end mid-sentence at 30s/60s/90s/120s boundaries (tool truncation), so a few words are lost, including the final instruction sentence.
@louischinasourcing8: When negotiating with Chinese suppliers, reframe three weak...
- Anecdotal advice, no evidence of price reductions. Framing intent "we will place big orders later" can be a bluff that harms trust if untrue (the creator says "don't lie" but this nudges it). Creator is a sourcing agent with a commercial interest.
@decoyreport: Add one extreme, rarely bought option (a decoy/anchor) next to your...
- Decoy effect is real in behavioural economics but the effect on a flying Roadster is speculation by the creator, and Tesla's motives (tech demo, hype) are not proven. Details on Tesla pricing tiers are as stated; unverified. Results from small brands are not shown.
@gentlerainman: To move from $5,000 to $50,000 clients, do not pitch the big...
- No evidence of conversion rates; generic advice repeated across his reels (funnel for his own podcast/programs).
- A paid pilot credited or not toward the main fee is not specified.
Branding, design and packaging
@barnardco: Clients break their own logos in three ways (ignoring safe space,...
- "Don't rotate" is stated without reasoning (a commenter asks why); rules are conventions, not laws.
- Pure authority/personal-brand piece with no data; advice is standard practice among logo designers.
@allstarsteven: A product-discovery reel about the Japanese colour-combination book...
- Affiliate relationship not disclosed in the reel; the tool-link funnel is not verified in the data.
- "Every designer is obsessed" is framed from one clip. Book's palettes are 1930s; accuracy of "Wada No." cards (AI-made visuals) not verified. Link hand-out depends on automation reliability for ~15k comments.
@mixesintheattic: Ranks nine brand ingredients for building a new brand in 2026, most...
- Pure opinion, no data. Conflicts with marketing science (distinctive assets/consistency/mental availability) and with the creator's own placing of consistency at 5 while "stick with it" for logo. Commenters in branding point out categories differ (luxury/consumer packaged goods vs services).
- Brand guidelines "not important for small brands" is risky when more than one person produces content (agency, freelancers, UGC).
- Creator's own authority is uncertain: replies include "an absolute nobody" to a critic; no stated credentials in the clip.
- "Purpose is last" is true for ads, not necessarily for gifting/luxury where story and values support price.
@itsdesignertom: Hiring managers read three Claude-default tells in design portfolios...
- The creator admits the hover example was weak; the generalisation is trend-based, not data. Eyebrows are legitimate pre-AI.
Retention, email, WhatsApp, CRM
@i_mails_: Cold email deliverability checklist from an agency operator sending...
- No numbers for what safe volume per inbox is, how to warm up, or which verification tool; "30,000 per client" is unverified. Advice to re-run sequences on the same contacts after 3 months may annoy prospects and harm sender reputation.
- No mention of SPF/DKIM/DMARC, separate domains or inbox rotation, which are the basics.
@wellcopymax: Post-purchase upsell that converts one-time buyers to subscribers by...
- "Refund" wording misleading; likely violates platform and consumer-protection norms and risks chargebacks (comments flag it). No data shown. Subscription cancellation/retention unaddressed. Promotional tie-in to Aftersell.
@suraj.adsriddle: Super Acquire is a Shopify app that identifies anonymous site...
- Strong legal/compliance risk: India's DPDP Act and WhatsApp Business policy require consent for promotional messages; sending to identity-resolved visitors who never gave a number is likely non-compliant and could get the WhatsApp number banned.
- 5% conversion and "guaranteed increase" are unverified claims; test window is 3 days, the dashboard screenshot date format appears to be from 2024 (possible reused footage).
- The product may be a reskin of existing tools (a commenter suggests GoKwik Engage); availability disputed ("isn't available").
- Spoken numbers are garbled in the transcript ("800 se 200" vs caption 150-200).
Finance, ops, legal and business tools
@prathamjindal14: Tier list of international payment platforms for Indian freelancers...
- Rankings are one person's opinion and possibly a sponsored placement (comment says "promoting"; he says "we use Skydo", no disclosure shown).
- Claims about Stripe being invite-only in India and Razorpay's 4-5% effective cost may be outdated or imprecise; no fee tables are shown.
- The "eight year"/"proper test tier" lines in the auto transcript are mis-transcriptions of "A tier" and "S tier". I used the final on-screen board to confirm.
- Crypto suggestions in comments carry legal and tax risk in India and are not recommended here.
@heysansser: Nine open-source GitHub projects that replace paid SaaS (Higgsfield,...
- "Free" is misleading for several: Open Higgsfield needs paid model APIs; VoiceStudio is heavy; self-hosting has server and time cost; AppFlowy/Cal.com/Chatwoot have paid cloud tiers.
- None of the tools are verified as feature-equivalent replacements; no comparison given.
- Links reach you only via DM from a bot; verify each repo's name and licence yourself before installing (impersonator repos exist).
@madetoscale2026: Oracle Cloud's Always Free tier (Arm A1 Flex
- "Always free forever" not guaranteed; Oracle can reclaim idle instances and terminate accounts; verification charge claim contradicted by commenters (small hold). Capacity often unavailable.
- Account has 3 likes in metadata and 898 comments, likely engagement farming via keyword.
The best reels I save, in one email
When I find a reel worth acting on, you get the short version: what it says and what to do. Free, rare, one click to leave.
Rather have it done for you? Work with Tanish