Distribution 101: How to Sell Your Products

✍ EP🕐 2026-07-29📦 42.4 KB 🟢 已读 𝕏 文章列表

文章探讨了在产品构建日益简单的当下，分销与营销成为了关键技能。作者重点分享了在 X 平台进行有效产品销售的实战经验，包括如何设置个人资料、选择高价值利基市场、确立独特角度以及持续提供有价值的内容。

# distribution 101: how to sell your products **作者**: EP **日期**: 2026-07-28T12:38:41.000Z **来源**: [https://x.com/eptwts/status/2082083321383997553](https://x.com/eptwts/status/2082083321383997553) --- ![Image](/doc/3523/image/img_6793430b3773.jpg) the consensus on AI twitter right now is that BUILDING is the skill worth learning... and while that's true, building has gotten so easy that the product is rarely the bottleneck anymore. everyone is hyperfocused on building & almost no one is genuinely good at marketing. that gap is exactly why distribution is the #1 skill you can possibly learn right now. over the years i've tried a ton of distribution channels for pretty much every type of product - software, info products, agency services, etc. and i've really honed down two of them. there's a third (short-form content) that isn't my specialty, but i've still used it to get direct sales & farm leads, so i'll cover what i know about it. what i won't go over is SEO & paid ads. they're not my specialty whatsoever and i don't wanna lead you guys down the wrong road. this does not mean they're not worth doing, i just don't know enough to guide you in that department. in this article i'll cover what has actually worked for me: how each platform works & what type of content to make to actually get sales. # channel #1: X ## setting up your profile most X growth guides are absolute BS & tell you to build these weird overoptimized profiles that look like they belong on linkedin. the truth is you need to do the opposite, i've noticed all the best accounts keep it simple with straight to the point personas. people crave authentic interactions, someone who can vibe with them, not a stiff corporate linkedin user. profile picture keep it simple and not try-hardy. don't overthink this. name simple and easily memorable. your real name, a nickname, etc. avoid cringy names like "John - Landing Pages For Business Owners" or any other role descriptor in your name - don't make it obvious that you're just here to sell something. bio a few words that explain what you do in a sleek, non-tryhard way. a one-liner. here are a few good examples: - "running paid ads for DTC brands" - "web dev wizard" - "vibe coding & sharing my learnings" - "founder of {startup}" link in bio either your product's landing page or an alternate channel like your newsletter. i click into some people's pages and it's 3 lines in their bio with a CTA, their name is "John / Landing Pages For Business Owners"... it's like, can you make it any more obvious that you're trying to sell me something lol? the name, pfp & bio don't matter too much as long as you don't make it look try-hardy. simplicity & authenticity win here. ## niches = networks on X there are niche "buckets" - groups of people w/ shared interests that form networks across the platform. when a post starts getting pushed, it goes to the immediate network first for example an AI post gets pushed to the AI network. once a post goes stupidly viral it breaks out of its network, and you start to see the difference in the reactions you get. this is why choosing the right niche from day one is so critical, you're choosing which network you want to tap into and dominate. ## the 1:1 niche-product rule the niche you pick has to be 1:1 with the product you're selling, this is non-negotiable. selling a SaaS for graphic designers? you need to get on design twitter. selling a coaching program for marketers? marketing twitter. your audience must be your potential customers. too many people mess this up... they build an audience in one niche because it's "easier" or more interesting, then try to pivot & sell to a completely different one. it doesn't work - you end up with thousands of followers who don't care about what you're selling. your audience is your distribution. 10,000 fitness enthusiasts are worthless to your B2B SaaS while 1,000 followers in the exact niche you sell to can generate serious revenue since they're pre-qualified buyers who already trust you. ## pick a niche you actually know ideally you already have above average knowledge in the thing you're selling. content creation gets infinitely easier sharing from experience instead of constantly researching - you can speak with authority & give real examples beginners in your space genuinely can't offer. if you don't have that knowledge yet you can build it up & grow while you do it turn your learning process into content positioned in an authoritative way. learn something -> format it as an educational tweet. choosing a niche is choosing a problem you know (or are learning) how to solve. the more painful & obvious the problem, the better your content performs - people are actively looking for things that'll help them. ## not all niches print money you can work just as hard growing an account in a low-value niche as a high-value one, and the financial outcomes will be drastically different. high-value business, marketing, AI, tech, crypto, trading, sports betting, gambling. the people in these niches either have money or are actively trying to make money. the pain points are massive & the solutions result in premium pricing. mid-value fitness, health, design. still profitable, but the audience is more price-sensitive & ticket prices are generally lower. you'll need a bigger audience to hit the same revenue numbers. low-value entertainment, sports, pop culture, facts, conspiracies. you can build massive followings here and still struggle to monetize. the audience came to be entertained, not to solve problems or buy products. the best niches solve painful, obvious problems that people are desperate to fix, money seems to be the peak of them. ## your unique angle you can't be a pointblank copy of someone who's already crushing it. there are successful accounts in every niche - if you just repeat what they're saying in the same way, why would anyone follow you instead of them? your angle is the overlap of what you know, what you've done & what you can teach. maybe you took an unconventional path to a desired outcome, or figured out how to do something faster, cheaper or simpler than everyone else. that's your angle. a good niche + a clear angle makes tweet ideas endless. if you're constantly struggling for content ideas then you chose the wrong niche, or an angle that isn't defined clearly enough. ## if you're a beginner you don't need to be the world's leading authority, you just need to be a few steps ahead of your target audience. watch youtube videos, read the articles, take courses, note down what you learn & turn it into valuable tweets. you're acting as a curator & translator, making complex information digestible - most people don't have the patience to consume hours of content, they want the key insights delivered quickly. authority comes from clear, repeatable value. ## the "VALUE" rule every tweet must bring value to the reader. that's the only way to get engagement that turns into $ and truly become a thought leader. ask yourself before hitting post: "would i find this valuable if i saw it in my feed?" if the answer is no, rewrite it or don't post it. ## what NOT to do X guides usually tell you what to do. understanding what not to do is just as important, maybe even more. here's what i'd avoid if i was you... don't be a reply guy it's very low status & you just end up being annoying. if you have something meaningful to add, sure, reply. but the weird AI slop replies that add barely any value? if you try to systemize that, you're going to cook your account. and you don't need replies at all to get traction - i personally did zero replies and got my account to 7k followers within my first 7 days. "tweeting too much" doesn't exist as long as you're tweeting value, you can do 100 tweets a day and it doesn't matter. the algorithm doesn't punish volume but it rewards engagement. the fear of posting too much comes from people who don't have enough valuable things to say which is a content quality problem, not a volume problem. don't "document your journey" this is one of the most harmful pieces of advice floating around twitter. nobody cares about your journey unless you've already made it. people follow you for solutions & expertise, not to watch you figure things out in real time. documenting positions you as a beginner & kills your authority before you build it. share what you're learning like you're teaching it - principles, not personal anecdotes. once you've actually made it then you can look back and tell the story. format well, but don't overformat write like you talk, no walls of text - add line breaks and format your content in digestible chunks. but don't overformat into something that sounds AI generated either. X is one big group chat, behave like a real person - if someone reads your tweet and thinks "this sounds AI-generated", you're cooked. write simply, not smartly you go viral by dumbing down complex topics. beginners & intermediates are 90% of any audience - if you write at a level only experts can understand then you cut off 90% of your reach. write in short sentences, explain something assuming the reader has 80IQ. diversify your formats one content type posted repetitively saturates your audience & your engagement will plateau. mix educational threads with quick insights, throw in a contrarian take. test formats & double down on what works. followers does not = reach big accounts often have a ton of dead followers - i'd argue i get higher engagement at 117k followers than many 300k follower accounts. the algorithm doesn't care how many followers you have, only how many people engage. build an engaged audience!!! ## how the algorithm actually works quick caveat: the algo is constantly changing, what worked last month doesn't work as well today & that will continue - you must experiment & be adaptable... here are the main metrics that matter: engagement metrics likes, replies, bookmarks, retweets & quote tweets all signal your content is worth showing to more people, but they're not equal. retweets & quotes expose your content to entirely new audiences. likes & bookmarks are good, replies used to be good but kinda got nerfed - they still boost your post but not as much. profile clicks are extremely good - clicking through to your profile signals you're a brand people want to follow & its a big engagement signal. dwell time this is how long a user spends on your tweet - one of the most underrated factors. if you scroll past in half a second that's a negative signal. if you stop, read & think for 10-15 seconds that's a positive signal. this is why my prompt tweets used to do so well - people read them, copy them & go test them in ChatGPT, so dwell time gets very high. it's also part of why media outperforms plaintext. author reputation think of it as your credit score on twitter, determined by how people react to your content. consistent good metrics = the algorithm treats you as a trusted source, giving future posts more initial reach. low-quality content that gets skipped throttles you. big account engagement the reach of whoever engages with your content matters. when a big account in your niche engages with your tweet, it gets shown to their audience - and when their audience engages too, the algo reads it as "this post is resonating with this niche bucket, show it to more of them". this is not about begging for engagement, it's about deliberately curating it from bigger accounts (more on this in a second). engagement velocity how much engagement you get in the first 30 minutes. the algorithm uses it as a leading indicator of how well your tweet will perform - strong immediate engagement means it pushes the post harder. the relationship graph the algorithm maps how closely connected users are by what they interact with - not just follows, but likes, replies, retweets, profile clicks, even DMs. when you tweet, the algo first tests the post on your immediate graph, the people most likely to find it relevant. strong early engagement expands distribution to similar graphs, and it keeps expanding until engagement drops off. that's how a tweet goes from 100 views to 100k. it's also why engagement groups & follow4follow cook your account: fake engagement dilutes your graph's signal, confusing the algo into thinking your content is for everyone, so it shows it to no one. your core audience is your launchpad. build real relationships in your space & the graph gets dense enough to amplify everything you post. twitter premium gives you an increase in account ranking, but i blew up even before i had it lol. i'd get it anyway, but it's not a must-have. the algorithm changes constantly don't obsess over gaming it. if your content is genuinely valuable to the person you're targeting & formatted in an easy to consume, attention-span friendly way, it'll ALWAYS do well, because it'll get good engagement metrics no matter how the weights shift. your job as a creator is to clarify lessons and condense them into one tweet - the algo handles the rest. ## 0 to 1k: how to avoid posting into the abyss the hardest thing for someone who's never grown on X is taking an account from 0 to even 1k followers without posting into the void. X really penalizes new accounts - you can't just "post" and expect to blow up. the algorithm doesn't trust you yet, your author reputation is non-existent & your tweets get shown to maybe 50 people if you're lucky. in my case, i went from 0 to 1k followers in one day. why? i have a system that works, and i've replicated it over a dozen times with my mentees. ## the outside push strategy the system is to get a bigger account in your niche to engage with your content from the jump. that's the entire strategy distilled to one sentence. the nuance is in the execution. first, what counts as a "big account" - not follower count, average impressions. i wouldn't work with anyone averaging less than 5k impressions per tweet. someone can have 100k followers & average 2k impressions because their audience is dead, while someone with 20k followers averages 15k because their author reputation is strong. the second account is the one you want. the affiliate play: find a big account in your niche with a product, offer to become an affiliate & ask them to push your early content in return. many will agree because it lands money in their pockets. ## the two variables growing an account from 0-10k comes down to two variables. variable 1: content that's ACTUALLY valuable to one specific niche non-negotiable. if your content is shit, no amount of outside push saves you - people click through because a big account engaged, but if the profile doesn't deliver, they don't follow. the reaction to your tweets you're after is "i need to save this and apply it". variable 2: an outside push from a big account in that same niche great content without an initial pool of eyes is useless. with zero followers the algorithm doesn't know who to show you to - one big-account engagement is a massive signal to show this to their audience. fulfil both and your chances of success are literally 100%. most people fail because they only solve one - great content nobody sees, or a big push on mediocre content. if you get both right the algo will love you. # channel #2: YouTube ## channel authority youtube decides how much it should push your channel based on its authority level. i don't know exactly how youtube determines authority, but one thing is very clear - high authority channels easily outrank low authority channels when ranking videos, even if they're getting less engagement. can you start fresh? yes - youtube is one of the most forgiving platforms for new accounts, your videos will still get pushed. would buying an aged account help? not necessarily, an aged account with no videos, or old botted ones, has no authority. the algorithm is too sophisticated for workarounds that simple. authority is more than age, views or subscribers. from experimentation old accounts that got views years ago & whose old videos still get some views today have significant authority by default. think about it this way... imagine MrBeast uploaded a video called "how to make money online". it would outrank every single biz-opp youtuber before it even gets a single view. youtube already knows how it should react to his videos based on previous experience. (an interesting quirk: super old channels tend to rank well on google search, while newer high-authority channels do better in youtube search. the two algorithms treat age differently.) apart from buying an old channel, authority comes down to channel age & a history of good engagement. the whitehat way to build it is simply to build a good channel. as far as i know, there's no way to game it organically. ## the five discovery types people find your video in five ways, and youtube treats each one as an isolated bucket. views that come from search won't impact how well your video does in browse & vice versa - every discovery type tracks its algorithm metrics separately. browse features the home page & subscriptions feed have super high discoverability - most people click videos straight from their home page. this has massive volume, but nit's ot insanely targeted, so intent to buy is lower people are there for entertainment, which is harder to sell through. for an educational video to do well in browse it still has to spike dopamine - the allure of "secret knowledge" + good structure + a rapid-fire release of information. recommended your video showing up next to another video. it means your video is related to that video & the user's browsing profile. search they manually searched for your topic. if the keyword has high buying intent, these people are 100x more likely to buy your product than someone who got recommended your video randomly. channel pages pretty irrelevant to discovery, these views won't do much for you. external they found your video through a link - google search, a share on X, a DM. these views won't do much directly, but people who watched & interacted start getting your future videos recommended. sharing to your big X following won't make a video go viral, but it puts your videos on their youtube home pages too - you're doubling your touchpoints. ## the metrics that matter CTR (click-through-rate) the % of people who see your thumbnail & title that actually click. people say it should be 10%+, but those are entertainment numbers - the algorithm doesn't work off benchmarks, it weighs who gets the best metrics within that niche's pool of creators. i've had a video do insanely well at 3% CTR, and 15% CTR videos that flopped. regardless, CTR is a real factor which is why packaging matters: good title + good thumbnail, clearly visualizing the benefit of clicking. retention youtube wants people on the platform as long as possible, so total minutes beat percentages... a 15-minute video at 40% AVD (6 minutes watched) beats a 5-minute video at 60% (3 minutes) in youtube's eyes. longer videos can perform better if you hold retention. the patterns youtube likes is high initial retention (fight the drop-off in the first 30 seconds) & a smooth curve session time this is how long someone stays on youtube after your video. if your 10-minute video starts a session & they watch 3 more videos for 30 minutes total, youtube loves you - you kept someone watching ads for half an hour. binge-session starters get pushed harder than videos people exit youtube after. optimize with end screens pointing to your other videos, playlists that trigger autoplay & shorts that tease the long-form via "related video". ## the testing waves this is the part most people completely misunderstand. youtube doesn't randomly blast your video out and see what sticks. when you upload, youtube doesn't know if the video is good. so it shows it to a seed audience first - your subscribers, people who've watched you before, people who watch similar content - and the seed's size depends on your channel authority. then it watches how they react, e.g. did they click, did they watch, what did they do after? if you perform well with the seed relative to your niche the seed size expands. but if you perform poorly youtube says "this ain't it" - the video is essentially dead. if your video doesn't do well with your seed audience it's cooked. if you pass the test the expansion comes in waves. wave 1: people who watch similar creators, with browsing patterns & demographics matching your seed audience. wave 2: related but not identical content, broader demographics, more discovery surfaces. wave 3: people outside your typical niche entirely, hard pushes on browse features, aggressive recommendations on other videos. wave 3 is when view counts explode - and most videos never make it there, especially not educational/promotional ones. the audience matching system youtube has a profile on every user - what they watch & skip, what they rewatch, share and comment on, down to the time of day & device. your video is shown to people youtube thinks will enjoy it. this is why packaging matters so much: show the perfect person a thumbnail they don't click and the algo will think to not show this video to this type of profile... how youtube categorizes your video there are three inputs. your metadata: title, description, thumbnail. the actual content: the algorithm literally watches your video, transcribes everything you say & identifies the visuals. initial viewer behavior: "people who like X also like this video, so it must be about X." your first few views are teaching youtube who your audience is. the pool system think of youtube as millions of "pools" of viewers - the fitness pool, the make-money-online pool, i'd imagine even sub-pools like "course sellers with interests in VSL creation". your video gets placed into pools & youtube asks: how does it perform against the other videos competing in the same pool? 5% CTR & 45% retention in a pool averaging 3% & 35%, you'll get pushed harder. below your pool's average will make your video tank though. this is why benchmarks are misleading - a 5% CTR might be incredible in one niche and terrible in another. it shows your video to people it deems pre-qualified to be interested, in phases, tests engagement at every discovery touchpoint, and decides whether to push harder based on the metrics that come back. ## youtube SEO is massively misunderstood lots of people say you need to optimize your description, your tags, etc. but this is massively overstated. before i explain how ranking actually works, here's why you'd want to rank in the first place. search converts differently search is usually a small share of total views, so most people underweight it. but someone searching "best AI automation tools" is actively hunting for a solution right now - this means better conversion & more qualified leads, because they're already solution-aware and further down the funnel. seach beats recommended any day when it comes too intent to purchase. low-intent vs high-intent keywords most people chase massive search volumes without understanding searcher intent. take "AI" for example: 7M+ searches a month, but who's searching it? students writing papers, people reading news, people with zero clear intent to buy anything. if you rank rank #1 for it you'll get massive traffic with terrible conversions. now take "ai tool for email marketing", this gets 1.45k searches a month. it's significantly less traffic, but significantly higher view:conversion ratio - everyone searching it is actively looking for exactly what you're selling. high-intent keywords follow patterns: "best [solution]", "[solution] for [use case]", "[solution] review", "[solution A] vs [solution B]", "alternatives to [solution]", "how to [achieve outcome]", "fix [problem]". keyword research w/ vidiq open the keyword research tool at vidiq.com, search your high-intent patterns & look at monthly youtube volume plus the related variations it suggests. pass criteria = 1k+ monthly searches, high buying intent & clusters of related high-intent terms (a broader topic ecosystem) fail criteria = only broad low-intent keywords, under 1k searches total disregard vidiq's competition score, it's usually off - gauge competition by searching the keyword on youtube & seeing which channels actually rank. how ranking actually works the title and description don't really matter as long as the keyword is in both once and the video is actually about that topic. you don't even need tags - i rank videos with no tags all the time. youtube's algorithm analyzes the actual content and understands what your video is about. what it actually cares about, e.g. does the video content match the keyword, and do people who search that keyword click and watch. that's literally it. put the keyword in your title once, in your description once, and focus on delivering the search intent. per-keyword metrics metrics get tracked separately per traffic source - and it goes deeper, the keyword gets split too. say your video targets "best AI automation tools", "AI automation for small business" & "small business automation": youtube tracks CTR/AVD from each keyword separately. 12% CTR & 60% AVD from the first but 4% & 30% from the third, and you rank up for the first & down for the third. channel authority is the multiplier some channels get good CTR and watch time off the bat and still don't show up in rankings, because youtube deems them low authority. established channels can publish mediocre metrics and still rank because youtube trusts them. the simplified formula: channel authority x (keyword-specific CTR + keyword-specific AVD) = your ranking optimizing for one keyword package the video for the exact person searching it. take "how to automate my business with AI": that's a business owner, not a developer - he's overwhelmed & wants practical steps, not theory. the thumbnail should make them think "this is exactly what i was looking for". the title should match their mental model: "AI Is Changing Everything" promises nothing but on the other hand "How to Automate Your Small Business with AI (Step-by-Step)" delivers the exact promise. then deliver on it in the first 30 seconds - no intro + no story - and structure the rest to answer their intent as efficiently as possible. ## the production workflow i picked this workflow up while working with a youtube automation agency 6 years ago. it was efficient enough that i've stuck with it ever since anytime i need to systemize content creation. most people approach video creation backwards: they make the entire video first, then slap on a title and thumbnail at the end. i'd advise the opposite. stage 1 - ideation & packaging get the idea, title & thumbnail made first, before production also get a mockup of the packaging to judge the video's potential before you spend anything making it. use thumbnailpreview.com to see how it'll look on the home page. ask does this clearly present the value? is it clickable? does it match the keyword's search intent? would i click this in my search results? address any no. use claude to ideate concepts & title options for thumbnails you can literally AI-generate very good ones nowadays, just check your competitors & replicate one of their thumbnails but not 1:1, everyone on YT does this, it's part of the game. get this stage wrong and nothing else matters, low CTR will disqualify retention from even mattering. stage 2 - scriptwriting (optional if you're recording raw) the goal is to get the point across in the simplest language w/ as little fluff as possible. length matters less than you think - people would rather get their solution in 3 minutes than 15. and if the script is unclear no amount of charisma or editing will save it. have AI write the first pass: give it the title, topic, video type & a bullet-point summary of the process w/ all the specific quirks from your knowledge - you're the expert, you have to give the LLM the facts, you can't expect it to "know" things. it outputs the hook, the content structured around your bullets, visual cues & the conversion CTA. then humanize it, replace AI-esque phrasing with how you'd actually talk, add personal examples. ask yourself does it deliver on the title's promise in the first 30 seconds? is every sentence necessary? will my audience understand it? stage 3 - recording there are two approaches here: faceless or faced. faceless can work extremely well if done right. i've personally made 6 figures with a faceless channel in the trading niche using the elevenlabs Adam voice & screen recordings showcasing the thing i was selling. it works, but it isn't optimal - i'm sure conversions would've been way better w/ a real trustable personality presenting the offer. what faceless buys you is speed & scalability. as for the voiceover i'd recommend recording it yourself or hiring someone if you can. you can also use elevenlabs, it takes minutes to generate, costs like $5-$30/month & it's easy to regenerate sections, but people can often tell & it's harder to build connection. AI voice is the move when you're testing ideas fast at high volume. for visuals screen recordings are the easy option - show the process step by step in the actual software. slideshows, miro boards & whiteboard drawing while speaking all do well too. keep elements flowing, nobody wants to stare at one static slide. the alternative is visualizers, motion graphics, b-roll & AI-generated footage illustrating the point. for faced videos you can record yourself, or hire a face. people buy from people they trust & the founder on camera builds that fastest - but it takes practice and it's harder to scale. the best part of faced videos is you don't need special editing to do well. face +good script + engaging delivery wins even with basic editing. stage 4 - editing you can either edit yourself (free, full control, time-intensive) or hire an editor, unfortunately there's still no good AI video editor afaik find editors through twitter portfolio searches, discord youtube automation communities & upwork. as for editing philosophy: don't overdo it, make it clean. every edit has three jobs. visualize the concepts (when you say "this tool automates email follow-ups", the screen shows the tool doing it). keep attention (cut empty space). and make your offer appealing. ## make every video a hidden VSL when youtubers discuss strategy, they're usually optimizing for views. when it comes to business & revenue, views are a vanity metric. what you should do instead is target your specific ICP with high-intent videos & position your solution as better than anything they've seen. fewer views than entertainment-optimized content & a smaller TAM, but it converts at a 100x rate. educate & pitch simultaneously don't make a great educational video and bolt a CTA on the end - "if you want to learn more, check out my course in the description" doesn't work. the whole purpose of the video should be tied in with what your offer solves. your videos must solve the same problem (or a sub-problem) as your offer say the offer is an SMMA course & the video is "how to do cold outreach for SMMA (step-by-step guide)". cold outreach is a sub-problem of running an SMMA - anyone searching it is trying to start or scale one. you teach the outreach step by step, real value, and as you teach you naturally position the course as the complete solution... you're educating them while simultaneously showing that your offer is the shortcut to results. you're reminding them of the problem they have & bringing up the solution you have. why this converts 100x better the traditional youtube strategy is to make broad entertaining content, get millions of views, & hope 0.001% convert the high-intent strategy is to make specific educational content for your exact ICP, get thousands of views & convert at 0.1%-1%. don't gatekeep the biggest mistake is holding back value because you're afraid of "giving away too much". the more you give, the more they trust you & the more they buy. your video should be so valuable that someone could implement it & get results without ever buying. they should see your offer as the shortcut to getting X outcome. they buy because you proved you know what you're talking about in the video. the formula summarized: target your ICP with high-intent topics, teach a sub-problem of what your offer solves, position the offer as the complete solution throughout, one clear CTA at the end & don't gatekeep. # channel #3: shortform (instagram reels + tiktok) quick reminder from the intro: this isn't my main channel, but i've used it for direct sales & lead farming, and this is the system. reels & tiktok run on nearly the same playbook, so i'm covering them as one channel & calling out the discrepancies where they actually matter. most of this info i got from my good friend @casualhyped who is a goat when it comes to shortform, i recommend giving him a follow. ## where the two platforms differ before the shared playbook, here's the discrepancies between the two. US IP optional but recommended on instagram - US accounts get better reach, features first & better monetization. on tiktok it's damn near required if you are trying to focus a US audience. warming instagram takes 1-2 days and you can engage actively on day 2 (follows, comments, saves). tiktok is stricter, you should spend 2 full days of pure consumption, then a follow test that decides whether the account lives. follower count matters heavily on instagram but tiktok tests every video independently - each upload gets a fair shot at virality regardless of your following. length 60-90 seconds is the reels sweet spot. tiktok is barbell-shaped, either 5-9 seconds or 60+, and avoid the 15-45 second dead zone in between. the top signal saves are instagram's most powerful signal. on tiktok it's shares (one share is worth 10+ likes) and rewatch rate (20%+ gets you a massive push). captions captions are searchable on tiktok, they are an SEO surface there, so front-load your keywords. everything else is shared. here's the system. ## account setup sidenote: this following stuff only applies if you are trying to maximize US audience & multi-account, if not you can skip this section you can target the US by getting a residential US proxy from anyip.io (residential, NOT datacenter - pick a major city), the V2Box app on ios to run it & a fresh email per account. connect the proxy, search "what is my IP" in safari & confirm it shows US before you create anything. then sign up w/ a niche-relevant username, a clear one-line bio based on your product & your funnel link in the bio. instagram specifics for a few accounts your home IP works - separate IPs only become good practice once you're scaling. tiktok specifics the proxy is mandatory & so is the device setup. - fully reset your iphone (settings -> general -> transfer or reset -> erase all content and settings) - set it up with no SIM, region united states, english (US), new york or LA time zone - create a US apple ID - address from google maps if asked, a temporary US number from textverified if asked. - contrary to popular belief a US SIM is NOT required, but everything else must look like a brand-new american device. - birth date 25+ when signing up. verify the IP & location both show US before creating the account - if they don't, fix the proxy first. don't create it and hope. ## warming new accounts get flagged as potential spam on both platforms. warming establishes a level of trust. instagram (1-2 days) day 1 is passive, scroll reels 20-30 minutes, watch niche content 5-10 seconds each, like 15-20 posts, no follows & no posting. by the end of day 1 your feed should be full of your niche day 2 is another 20-30 minutes, follow 10-15 niche accounts, 2-3 genuine comments & save 5-10 reels - then you're ready to post. tiktok (2 days + the follow test) on tiktok do two full days of pure consumption, 20-30 minutes each... scroll the FYP, watch niche videos to completion (a full watch is a strong signal), like 15-20. no follows & no comment & definitely no posting on either days. then run the follow test: try to follow 1-2 niche accounts. if the follow registers (your following count is above 0) then the account is trusted & you can start posting. if it doesn't register the account is flagged - delete it, get a NEW proxy & start over. don't try to salvage a flagged account. ## the viral formula: 4 pillars every viral short-form video on either platform follows the same structure: hook, retention, reward, engineered engagement. ## pillar 1: the hook (first 3 seconds) the hook formulas that work: curiosity gap: "the one thing nobody tells you about [topic]", "this feels illegal to know" shocking statement: "everything you know about [topic] is wrong", "stop doing [common action] - here's why" relatable pain: "you know that feeling when [struggle]?", "this is why your [thing] isn't working" results-based: "how i gained 10k followers in 30 days", "watch me turn [X] into [Y] in 60 seconds" pattern interrupt: "wait for it...", "POV: you just discovered..." the hook is visual & audio too for example a first frame with contrast or a face showing real emotion, a bold text overlay (5-7 words max, readable in one second) & the trending sound starting immediately. test 3 versions of the same video w/ different hooks. the hook alone can be the difference between 1,000 views and 1,000,000. ## pillar 2: retention both algorithms prioritize watch time above everything. 70%+ completion gets massive reach, and on tiktok a 20%+ rewatch rate on top of that reads as addictive content. get to the point in 1-2 seconds. e.g. a good opening is "this AI tool made me . here's how." cut every 2-3 seconds (every 2 on tiktok), the viewer's brain reads each cut as new content. visual monotony kills retention. use open loops to hold them ("the best part is coming...") & make it loopable - the ending flows back to the beginning ("and it all started when...") so viewers rewatch without noticing. ## pillar 3: the reward deliver on the hook's promise within 5-15 seconds. viewers are impatient, and if they feel baited they won't engage. give them a quick win they can use immediately, entertainment (a laugh is valuable even when it's not "useful"), an aha moment ("wait, THAT's why it works?"), a before/after transformation, or validation ("if you've ever felt this, you're not alone"). stack them where you can - a video that teaches + entertains + validates outperforms one that only does one thing. ## pillar 4: engineered engagement both algorithms weigh engagement velocity heavily. engineer it on purpose. comment bait: end with a question ("which one are you: A or B?") or a hot take ("unpopular opinion: ..."). comments tell the algorithm this is worth distributing. share bait: "send this to someone who needs it", "tag your friend/boss/partner". on tiktok, shares are the single most powerful signal there is. save bait: "save this so you don't lose it". reference content like checklists, templates, anything step-by-step get saved naturally. on instagram, saves are the strongest signal. watch-again bait (tiktok): hide details that only show on a rewatch, or end with a twist that requires one. captions: these are searchable on tiktok, so front-load keywords - "AI money making trick: how i made $10k in 30 days" is a good example. use the exact phrases people would search & end with an engagement hook ("comment 'AI' and i'll send you the tool"). ## the video structure template seconds 0-3: the hook. pattern interrupt or curiosity loop, bold text overlay, trending audio already playing. seconds 3-5: context - one or two sentences on the problem we're solving. seconds 5-10: value delivery, so the actual reward, fast cuts, text reinforcing each point. seconds 10-15: payoff. complete the promise, add the twist, loop back to the start. in the caption add a keyword-rich hook + engagement question + hashtags. for audio, a trending sound (on tiktok, ideally less than 48 hours old) or original audio w/ a strong opening line. ## posting schedules instagram phase 1 (0-1k followers) is 2x per day w/ a 4-hour gap (say 9am & 1pm) - this ensures consistency without overwhelming a new audience, and each post gets its own algorithmic push. phase 2 (1k+) is up to 15 trial reels per day - trial reels test on non-followers before publishing to everyone, so you can post tons & publish only the winners to your main profile. tiktok phase 1 (testing) is 1x per day, same time every day, until ONE video breaks 1,000 views - tiktok gives each video 10-30 minutes to prove itself, posting more before traction isn't recommended. phase 2 (scaling) is 2x per day w/ a 4-hour gap. the first 30 minutes rule (tiktok) initial distribution is decided in the first 10-30 minutes - this is calculated via. completion rates, rewatches, engagement velocity & shares. don't be discouraged if your first videos flop. tiktok tests every video independently - your breakthrough could be video #5 or video #25. ## the shortform checklist - account warmed (1-2 days on IG, 2 days + follow test on tiktok) - feed showing your niche before the first post - hook in the first 3 seconds - cuts every 2-3 seconds (every 2 on tiktok) - text overlays & captions on every video - trending audio (under 48 hours old on tiktok) - length: 60-90s on reels, 5-9s or 60+ on tiktok - reward delivered within 5-15 seconds - engagement bait: comment, share, save or rewatch from there, all that's left is to post consistently & send the traffic down your funnel. # CONCLUSION this article took a while to write haha, if you liked it i'd appreciate a reply with your feedback everything in this article is the playbook i run hands-on in my X & youtube partnership program. for the past year i've been taking people who are great at what they do - developers, marketers, founders - and turning them into public figures with a proven, repeatable framework. it's taken many big accounts on here very far, some even past 100k followers & built me a massive network along the way. if you think you have what it takes and want me to help build up your brand, i'd love to read your application & see what you can bring to the platform. apply here: eptwts.com (p.s. i don't charge for this) hope you enjoyed the article. ## 相关链接 - [@eptwts](https://x.com/eptwts) - [103K](https://x.com/eptwts/status/2082083321383997553/analytics) - [vidiq.com](https://vidiq.com/) - [thumbnailpreview.com](https://thumbnailpreview.com/) - [@casualhyped](https://x.com/@casualhyped) - [anyip.io](https://anyip.io/) - [eptwts.com](https://eptwts.com/) - [Upgrade to Premium](https://x.com/i/premium_sign_up) - [8:38 PM · Jul 28, 2026](https://x.com/eptwts/status/2082083321383997553) - [103.9K Views](https://x.com/eptwts/status/2082083321383997553/analytics) - [View quotes](https://x.com/eptwts/status/2082083321383997553/quotes) --- *导出时间: 2026/7/29 09:35:36* --- ## 中文翻译 # 分销入门指南：如何销售你的产品 **作者**: EP **日期**: 2026-07-28T12:38:41.000Z **来源**: [https://x.com/eptwts/status/2082083321383997553](https://x.com/eptwts/status/2082083321383997553) --- ![Image](/doc/3523/image/img_6793430b3773.jpg) AI 圈目前的共识是，BUILDING（构建产品）是一项值得学习的技能…… 虽然这没错，但构建产品已变得如此容易，以至于产品本身很少再成为瓶颈。每个人都过度专注于构建，几乎没有人真正擅长营销。这种差距正是为什么分销是你目前能学到的最重要的技能。多年来，针对几乎所有类型的产品——软件、信息产品、机构服务等——我都尝试过大量的分销渠道。我真正精简出了其中的两个。还有第三个（短视频内容），虽然不是我的专长，但我用它来获得直接销售和挖掘线索，所以我也会涵盖我所了解的内容。我不会涉及 SEO 和付费广告。它们完全不是我的专长，我不想误导大家。这并不意味着它们不值得做，只是我在那个领域知道的不足以指导你们。在这篇文章中，我将涵盖那些对我真正有效的方法：每个平台是如何运作的，以及制作什么样的内容才能真正获得销售。 # 渠道 #1: X ## 设置个人资料大多数 X 增长指南都是废话，它们让你建立这些奇怪的、过度优化的个人资料，看起来像是属于 LinkedIn 的。真相是你需要反其道而行之。我注意到所有最好的账号都保持简单，人设直截了当。人们渴望真实的互动，一个能和他们产生共鸣的人，而不是一个死板的企业 LinkedIn 用户。头像保持简单，不要太用力。不要想太多。名字简单且容易记住。你的真名、昵称等。避免尴尬的名字，如 "John - Landing Pages For Business Owners"（约翰 - 为企业主提供落地页）或在名字中加入任何角色描述——不要让你只是来卖东西这件事变得太明显。简介用几句简洁、不造作的话解释你做什么。一句话。这里有一些很好的例子： - "为 DTC 品牌投放付费广告" - "网页开发向导" - "氛围编程 & 分享我的学习心得" - "{初创公司} 创始人" 个人简介链接要么是你产品的落地页，要么是另一个渠道，比如你的时事通讯。我点进某些人的主页，简介里有 3 行文字还加个行动号召（CTA），名字是 "John / Landing Pages For Business Owners"……这就像，你能再明显一点你是想向我推销东西吗 lol？只要你不显得用力过猛，名字、头像和简介并不太重要。简单和真实在这里胜出。 ## 细分领域 = 网络在 X 上有细分领域的“圈子”——拥有共同兴趣的人群在整个平台上形成网络。当一条推文开始被推送时，它首先会推送到直接的网络中。例如，一条关于 AI 的推文会被推送到 AI 网络。一旦一条推文变得极其火爆，它就会突破其网络圈层，你会开始看到你收到的反馈发生了变化。这就是为什么从第一天起选择正确的细分领域如此关键，你是在选择要接入并主宰哪个网络。 ## 1:1 细分领域-产品规则你选择的细分领域必须与你销售的产品保持 1:1 的对应关系，这是不可妥协的。在向平面设计师推销 SaaS？你需要上设计推特（Design Twitter）。在向营销人员推销辅导课程？上营销推特（Marketing Twitter）。你的受众必须是你潜在的客户。太多人搞砸了这一点……他们在某个细分领域建立受众是因为那个领域“更容易”或更有趣，然后试图转型并向一个完全不同的领域推销。这行不通——你最终会拥有成千上万个不在乎你卖什么的粉丝。你的受众就是你的分销渠道。对于你的 B2B SaaS 来说，10,000 个健身爱好者毫无价值，而 1,000 个处于你销售的确切细分领域的粉丝却能产生可观的收入，因为他们是已经信任你的预筛选买家。 ## 选择一个你真正了解的细分领域理想情况下，你对你销售的东西已经具备高于常人的知识。从经验分享内容创作会变得无比容易，而不是不断地做研究——你可以权威地发言，并给出你领域内的初学者真正无法提供的真实案例。如果你还没有这些知识，你可以在建立它的同时成长将你的学习过程转化为以权威方式定位的内容。学到一点 -> 把它格式化为一条教育性推文。选择细分领域就是选择一个你知道（或正在学习）如何解决的问题。问题越痛苦、越明显，你的内容表现就越好——人们正在积极寻找能帮助他们的东西。 ## 并非所有细分领域都印钞你在低价值细分领域和一个高价值细分领域发展账号可能付出同样多的努力，但财务回报却会截然不同。高价值商业、营销、AI、科技、加密货币、交易、体育博彩、赌博。这些细分领域的人要么有钱，要么正在积极尝试赚钱。痛点巨大，解决方案意味着溢价定价。中价值健身、健康、设计。仍然有利可图，但受众对价格更敏感，票价通常较低。你需要更大的受众才能达到相同的收入数字。低价值娱乐、体育、流行文化、冷知识、阴谋论。你可以在这里建立庞大的粉丝群，但仍然难以变现。受众来是为了娱乐，而不是为了解决问题或购买产品。最好的细分领域解决的是人们迫切想要修复的、痛苦且明显的问题，而金钱似乎是这些问题的巅峰。 ## 你的独特角度你不能完全照搬某个已经做得风生水起的人。每个细分领域都有成功的账号——如果你只是用同样的方式重复他们说的话，为什么别人要关注你而不是他们？你的角度是你所知道的、你所做过的和你能教会的这三者的交集。也许你采用了非传统的路径来达到预期的结果，或者你想出了如何比其他所有人更快、更便宜或更简单地做某件事。那就是你的角度。一个好的细分领域 + 一个清晰的角度意味着推文灵感源源不断。如果你经常为内容灵感发愁，那你选错了细分领域，或者角度定义得不够清晰。 ## 如果你是初学者你不需要成为世界领先的权威，你只需要比你的目标受众领先几步。观看 YouTube 视频，阅读文章，参加课程，记下你学到的东西，并把它们转化为有价值的推文。你扮演的是策展人和翻译官的角色，让复杂的信息变得易于消化——大多数人没有耐心去消费几个小时的内容，他们希望快速获得关键见解。权威来自于清晰、可重复的价值。 ## “价值”法则每条推文都必须为读者带来价值。这是获得能转化为美元的互动并真正成为思想领袖的唯一途径。在点击发布前问自己：“如果我在信息流中看到这个，我会觉得它有价值吗？”如果答案是否定的，重写它或不要发布它。 ## 不要做什么 X 指南通常会告诉你该做什么。了解不该做什么同样重要，甚至更重要。如果我是你，我会避免以下情况…… 不要做“回复哥” 这地位很低，而且你只会变得烦人。如果你有有意义的内容要补充，当然，回复。但是那些几乎没有任何价值的奇怪的 AI 废话回复？如果你试图系统化地做这件事，你会毁了你的账号。而且你根本不需要回复就能获得关注力——我个人在最初的 7 天里零回复，就把账号做到了 7k 粉丝。 “发推太多”并不存在只要你发的是有价值的内容，你可以一天发 100 条推文，这没关系。算法不会惩罚体量，但会奖励互动。对发太多的恐惧来自于那些没有足够有价值的话可说的人，这是内容质量问题，不是体量问题。不要“记录你的旅程” 这是推特上流传的最有害的建议之一。除非你已经成功，否则没人关心你的旅程。人们关注你是为了解决方案和专业知识，而不是为了看你实时弄清楚事情。记录过程将你定位为初学者，并在你建立权威之前就扼杀了它。像教学一样分享你学到的东西——分享原则，而不是个人轶事。一旦你真的成功了，你可以回首并讲述这个故事。格式要好，但不要过度格式化像你说话一样写，不要有文字墙——添加换行符，并将内容格式化为易于消化的块。但也不要过度格式化成听起来像 AI 生成的样子。X 就像一个巨大的群聊，表现得像真人一样——如果有人读到你的推文并认为“这听起来像 AI 生成的”，你就完了。写得简单，而不是写得聪明通过简化复杂的话题来疯传。初学者和中级水平的人占任何受众的 90%——如果你以只有专家才能理解的水平写作，你就切断了 90% 的触达。写短句，解释事情时假设读者的智商是 80。多样化你的格式重复发布一种内容类型会让你的受众饱和，你的互动会停滞。将教育性的长推文与快速见解混合，加入一个反向观点。测试格式并加倍投入有效的内容。粉丝数不等于触达大账号通常有大量的死粉——我认为我在 117k 粉丝时获得的互动比许多 300k 粉丝的账号更高。算法不在乎你有多少粉丝，只在乎有多少人互动。建立一个互动性强的受众！！！ ## 算法实际上是如何工作的快速说明：算法不断变化，上个月有效的方法今天效果没那么好，而且这种情况会持续——你必须进行实验并保持适应…… 以下是主要的重要指标：互动指标点赞、回复、书签、转推和引用转推都表明你的内容值得展示给更多人，但它们是不平等的。转推和引用将你的内容展示给全新的受众。点赞和书签很好，回复以前很好，但被削弱了——它们仍然会提升你的推文，但幅度没那么大。主页点击量非常好——点击进入你的主页表明人们想关注你，这是一个巨大的互动信号。停留时间这是用户在你的推文上花费的时间——这是最被低估的因素之一。如果你半秒钟滑过去，那是一个负面信号。如果你停下来，阅读并思考 10-15 秒，那是一个正面信号。这就是为什么我的提示推文以前表现这么好——人们阅读它们，复制它们，并在 ChatGPT 中测试，所以停留时间变得非常高。这也是媒体内容表现优于纯文本的部分原因。作者声誉把它想象成你在推特上的信用评分，由人们对你的内容反应决定。持续的良好指标 = 算法将你视为值得信赖的来源，给未来的帖子更多的初始触达。低质量的内容被跳过会限制你。大账号互动与你内容互动的人的触达范围很重要。当你细分领域中的大账号与你的推文互动时，它会展示给他们的受众——而当他们的受众也互动时，算法将其解读为“这条帖子正在与这个细分领域圈子产生共鸣，展示给更多这样的人”。这不是关于乞求互动，而是关于故意从大账号那里策展它（稍后再详细说明）。互动速度你在前 30 分钟内获得多少互动。算法用它作为你的推文表现如何的领先指标——强大的即时互动意味着它会更用力地推送帖子。关系图谱算法通过用户的互动来映射用户之间的联系有多紧密——不仅仅是关注，还包括点赞、回复、转推、主页点击，甚至私信。当你发推时，算法首先在你的直接图谱上测试帖子，即那些最可能觉得它相关的人。强大的早期互动会将分发扩展到相似的图谱，并不断扩展直到互动下降。这就是一条推文从 100 次浏览变成 10 万次浏览的方式。这也是为什么互动群组和互粉会毁了你的账号：虚假互动稀释了你图谱的信号，混淆算法，让它以为你的内容是给所有人看的，所以它展示给没人看。你的核心受众是你的发射台。在你的领域建立真实的关系，图谱就会变得足够密集，以放大你发布的所有内容。推特 Premium 会增加账号排名，但我即使在没有它之前也爆火 lol。无论如何我都会弄一个，但这并不是必须的。算法不断变化不要痴迷于破解它。如果你的内容对你针对的人真正有价值，并且以易于消费、注意力友好地方式格式化，它总是会表现良好，因为无论权重如何转移，它都会获得良好的互动指标。作为创作者，你的工作是阐明课程并将其浓缩成一条推文——剩下的交给算法。 ## 从 0 到 1k：如何避免发布到虚无中对于从未在 X 上增长过的人来说，最难的事情是将账号从 0 做到甚至 1k 粉丝，而不只是对着虚空发布。 X 真的会惩罚新账号——你不能只是“发布”并期望爆火。算法还不信任你，你的作者声誉不存在，如果你的推文被展示，可能只有 50 个人看到。就我而言，我在一天内从 0 涨到了 1k 粉丝。为什么？因为我有一个有效的系统，并且我已经在我的学员身上复制了十几次。 ## 外部推动策略这个系统是让你细分领域中的一个大账号从一开始就与你的内容互动。这就是浓缩成一句话的整个策略。细微之处在于执行。首先，什么算作“大账号”——不是粉丝数，是平均展示量。我不会与平均每条推文展示量少于 5k 的人合作。一个人可能有 10 万粉丝，平均展示量 2k，因为他们的受众是死的，而一个人有 2 万粉丝，平均展示量 15k，因为他的作者声誉很强。第二个账号才是你想要的。联盟玩法：在你细分领域找一个有产品的大账号，提议成为联盟成员，并要求他们推送你的早期内容作为回报。许多人会同意，因为这能把钱放进他们口袋。 ## 两个变量将账号从 0-10k 增长归结为两个变量。变量 1：对一个特定细分领域真正有价值的内容不可妥协。如果你的内容很烂，多少外部推动都救不了你——人们点击是因为大账号互动了，但如果主页没有兑现，他们就不会关注。你对推文期望的反应是“我需要保存这个并应用它”。变量 2：来自同一细分领域大账号的外部推动没有初始眼球池的优质内容是无用的。零粉丝意味着算法不知道该把你展示给谁——一个大账号的互动是将其展示给其受众的巨大信号。同时满足两者，你的成功机会字面上是 100%。大多数人失败是因为他们只解决了一个——没人看到的优质内容，或者是对平庸内容的巨大推动。如果你两个都做对了，算法会爱你。 # 渠道 #2: YouTube ## 频道权威 YouTube 决定根据频道权威级别推多少流量给你。我不确切知道 YouTube 如何确定权威，但有一点很清楚——高权威频道在视频排名时很容易超越低权威频道，即使他们获得的互动较少。你可以从零开始吗？可以——YouTube 是对新账号最宽容的平台之一，你的视频仍然会被推送。购买老账号会有帮助吗？不一定，没有视频的老账号，或旧的机器人刷的账号，没有权威。算法太复杂了，没那么简单的变通方法。权威不仅仅是年龄、浏览量或订阅者。从实验来看，几年前获得浏览量且旧视频今天仍有一些浏览量的老账号默认具有显著的权威。这样想……假设 MrBeast 上传了一个名为“如何在线赚钱”的视频。它甚至在获得一次浏览之前就会在排名上超越每一个商业机会类 YouTuber。YouTube 已经根据以前的经验知道应该如何反应他的视频。（一个有趣的怪癖：超老频道往往在谷歌搜索中排名很好，而新的高权威频道在 YouTube 搜索中表现更好。这两种算法对年龄的处理不同。）除了购买老频道，权威归结为频道年龄和良好的互动历史。建立它的白帽方法只是建立一个好频道。据我所知，没有有机地破解它的方法。 ## 五种发现类型人们通过五种方式找到你的视频，YouTube 将每一种视为一个独立的桶。来自搜索的浏览量不会影响你的视频在浏览中的表现，反之亦然——每种发现类型都单独跟踪其算法指标。浏览功能主页和订阅信息流具有极高的可发现性——大多数人直接从主页点击视频。这有巨大的体量，但并不是极具针对性的，所以购买意图较低人们在那里是为了娱乐，这更难通过它销售。为了让教育视频在浏览中表现良好，它仍然必须刺激多巴胺——“秘密知识”的诱惑 + 良好的结构 + 信息的快速释放。推荐你的视频显示在另一个视频旁边。这意味着你的视频与那个视频以及用户的浏览档案有关。搜索他们手动搜索了你的主题。如果关键词具有高购买意图，这些人购买你产品的可能性是随机推荐你的人的 100 倍。频道页面与发现几乎无关，这些浏览量对你没什么大用。外部他们通过链接找到你的视频——谷歌搜索、在 X 上的分享、私信。这些浏览量不会直接起太大作用，但观看并互动的人会开始收到你未来视频的推荐。分享到你庞大的 X 粉丝群不会让视频疯传，但这也把你的视频放在了他们的 YouTube 主页上——你正在加倍你的触点。 ## 重要的指标 CTR（点击率）看到你的缩略图和标题并实际点击的人的百分比。人们说它应该是 10%+，但那些是娱乐数字——算法不基于基准运作，它权衡的是在该细分领域创作者池中获得最佳指标的人。我有一个 CTR 为 3% 的视频表现极佳，也有 CTR 为 15% 的视频惨败。无论如何，CTR 是一个真正的因素，这就是为什么包装很重要：好的标题 + 好的缩略图，清晰地可视化点击的好处。留存率 YouTube 希望人们在平台上停留尽可能长的时间，所以总分钟数胜过百分比…… 在 YouTube 眼里，15 分钟视频的 40% AVD（观看 6 分钟）胜过 5 分钟视频的 60%（3 分钟）。如果你能保持留存率，较长的视频可以表现更好。 YouTube 喜欢的模式是高初始留存（对抗前 30 秒的下降）和平滑的曲线会话时间这是在你的视频之后有人在 YouTube 上停留多久。如果你 10 分钟的视频开启了一个会话，他们又看了 3 个视频总共 30 分钟，YouTube 爱你——你让人看了半小时广告。开启狂欢会话的视频比人们看完就退出 YouTube 的视频被更用力地推送。用指向你其他视频的结束画面、触发自动播放的播放列表以及通过“相关视频”预告长视频的 Shorts 来优化。 ## 测试波这是大多数人完全误解的部分。YouTube 不会随机大量推送你的视频看看什么能行。当你上传时，YouTube 不知道视频好不好。所以它首先展示给种子受众——你的订阅者，以前看过你的人，观看类似内容的人——种子大小取决于你的频道权威。然后它观察他们的反应，例如，他们点击了吗，他们看了吗，看完后他们做了什么？如果你相对于你的细分领域对种子表现良好，种子大小就会扩大。但如果你表现糟糕，YouTube 会说“这不行”——视频基本就死了。如果你的视频对种子受众表现不好，它就完了。如果你通过了测试，扩展就会分波进行。第一波：观看类似创作者的人，浏览模式和人口统计与你的种子受众匹配。第二波：相关但不完全相同的内容，更广泛的人口统计，更多的发现表面。第三波：完全在你典型细分领域之外的人，在浏览功能上的大力推送，在其他视频上的激进推荐。第三波是浏览量爆炸的时候——大多数视频从未到达那里，尤其是教育/推广类视频。受众匹配系统 YouTube 对每个用户都有档案——他们看什么和跳过什么，他们重看、分享和评论什么，甚至到一天中的时间和设备。你的视频展示给 YouTube 认为会喜欢它的人。这就是为什么包装如此重要：向完美的人展示他们不点击的缩略图，算法会认为不要向这种类型的人展示这个视频…… YouTube 如何分类你的视频有三个输入。你的元数据：标题、描述、缩略图。实际内容：算法字面上观看你的视频，转录你说的所有内容并识别视觉效果。初始观看者行为：“喜欢 X 的人也喜欢这个视频，所以它一定是关于 X 的。”你最初的几次浏览正在教 YouTube 你的受众是谁。池系统把 YouTube 想象成数百万个观众“池”——健身池，在线赚钱池，我想象甚至有子池，比如“对 VSL 创作感兴趣的课程卖家”。你的视频被放入池中，YouTube 会问：它与同一池中竞争的其他视频相比表现如何？5% CTR 和 45% 留存率，而池平均是 3% 和 35%，你会被更用力地推送。低于你池的平均值会让你的视频一落千丈。这就是为什么基准是误导性的——5% CTR 在一个细分领域可能不可思议，而在另一个细分领域可能很糟糕。它向它认为预先有资格感兴趣的人展示你的视频，分阶段测试每个发现触点的互动，并根据返回的指标决定是否更用力地推送。 ## YouTube SEO 被严重误解很多人说你需要优化你的描述、标签等。但这被严重夸大了。在我解释排名实际上如何运作之前，这里是你首先想要排名的原因。搜索转化方式不同搜索通常占总浏览量的一小部分，所以大多数人低估了它。但搜索“最佳 AI 自动化工具”的人现在正在积极寻找解决方案——这意味着更好的转化和更合格的线索，因为他们已经意识到解决方案，并且处于漏斗的更下游。在购买意图方面，搜索每天都能击败推荐。低意图 vs 高意图关键词大多数人在不理解搜索者意图的情况下追逐巨大的搜索量。以“AI”为例：每月 700 万+ 次搜索，但谁在搜索它？写论文的学生，看新闻的人，没有任何明确购买意图的人。如果你为它排名第一，你会获得巨大的流量但转化极差。现在看“ai tool for email marketing”（邮件营销的 ai 工具），这每月获得 1.45k 次搜索。流量明显更少，但观看量与转化率的比例明显更高——每个搜索的人都在积极寻找你正在卖的确切东西。高意图关键词遵循模式：“最佳 [解决方案]”、“[用途] 的 [解决方案]”、“[解决方案] 评测”、“[解决方案 A] vs [解决方案 B]”、“[解决方案] 的替代方案”、“如何 [达成结果]”、“修复 [问题]”。使用 vidiq 进行关键词研究在 vidiq.com 打开关键词研究工具，搜索你的高意图模式，并查看每月 YouTube 量以及它建议的相关变体。通过标准 = 1k+ 每月搜索，高购买意图和相关高意图术语集群（一个更广泛的主题生态系统）失败标准 = 只有广泛的低意图关键词，总共少于 1k 次搜索忽略 vidiq 的竞争分数，它通常是不准的——通过在 YouTube 上搜索关键词并查看哪些频道实际排名来衡量竞争。排名实际上如何运作标题和描述并不太重要，只要关键词在两者中各出现一次，并且视频确实是关于该主题的。你甚至不需要标签——我一直用没有标签的视频排名。 YouTube 的算法分析实际内容并理解你的视频是关于什么的。它真正关心的，例如视频内容是否与关键词匹配，搜索该关键词的人是否点击并观看。就这样。把关键词放在标题中一次，放在描述中一次，并专注于传达搜索意图。每个关键词的指标指标按流量源分别跟踪——而且更深入，关键词也被拆分。假设你的视频针对“best AI automation tools”、“AI automation for small business”和“small business automation”：YouTube 分别跟踪来自每个关键词的 CTR/AVD。第一个 12% CTR 和 60% AVD，但第三个 4% 和 30%，你对第一个排名上升，对第三个排名下降。频道权威是乘数一些频道一开始就有好的 CTR 和观看时间，但仍然没有出现在排名中，因为 YouTube 认为它们是低权威。成熟的频道可以发布平庸的指标并仍然排名，因为 YouTube 信任它们。简化的公式：频道权威 x (特定关键词的 CTR + 特定关键词的 AVD) = 你的排名针对一个关键词进行优化为搜索它的确切人包装视频。以“how to automate my business with AI”为例：那是企业主，不是开发者——他不知所措，想要实用的步骤，而不是理论。缩略图应该让他们觉得“这正是我在找的”。标题应该匹配他们的心理模型：“AI Is Changing Everything”（AI 正在改变一切）什么也没承诺但另一方面“How to Automate Your Small Business with AI (Step-by-Step)”（如何用 AI 自动化你的小企业（分步指南））传达了确切的承诺。然后在前 30 秒兑现它——无介绍 + 无故事——并构建其余部分以尽可能高效地回答他们的意图。 ## 制作工作流程我在 6 年前与一家 YouTube 自动化机构合作时采用了这个工作流程。它非常高效，从那以后每当我需要系统化内容创建时，我都会坚持使用它。大多数人倒着处理视频创作：他们先制作整个视频，然后在最后贴上标题和缩略图。我建议相反。第一阶段 - 构思与包装首先获得想法、标题和缩略图，在制作之前还要获得包装的样机，以便在你花钱制作之前判断视频的潜力。使用 thumbnailpreview.com 查看它在主页上的样子。问自己这清楚地展示了价值吗？它值得点击吗？它符合关键词的搜索意图吗？我会在搜索结果中点击它吗？解决任何否定的答案。使用 Claude 构思概念和标题选项对于缩略图，你现在真的可以 AI 生成非常好的，只要检查你的竞争对手并复制他们的缩略图之一但不要 1:1，YT 上的每个人都这样做，这是游戏的一部分。这一阶段做错了，其他都不重要，低 CTR 会让留存率变得无关紧要。第二阶段 - 脚本编写（如果你录制原始素材则是可选的）目标是用最简单的语言传达要点，废话越少越好。长度比你想象的没那么重要——人们宁愿在 3 分钟内得到解决方案也不愿 15 分钟。如果脚本不清楚，多少魅力或编辑都救不了它。让 AI 写初稿：给它标题、主题、视频类型以及流程的要点总结，加上你知识中的所有具体怪癖——你是专家，你必须给 LLM 事实，你不能指望它“知道”事情。它输出钩子、围绕要点构建的内容、视觉提示和转化 CTA。然后人性化它，用你实际说话的方式替换 AI 短语，添加个人例子。问自己它在前 30 秒兑现了标题的承诺吗？每句话都是必要的吗？我的受众能理解吗？第三阶段 - 录制这里有两种方法：不露脸或露脸。如果做得对，不露脸可以非常有效。我个人在交易细分领域用一个不露脸频道通过使用 elevenlabs Adam 声音和屏幕录制展示我卖的东西赚了 6 位数。它有效，但不是最优的——我确信如果有真实可信的人性展示优惠，转化率会好得多。不露脸给你的是速度和可扩展性。至于旁白，我建议你自己录制或雇人。你也可以使用 elevenlabs，生成只需几分钟，每月费用约 5-30 美元，并且很容易重新生成部分，但人们经常能分辨出来，并且更难建立联系。AI 语音是在高体量快速测试想法时的做法。对于视觉，屏幕录制是简单的选择——在实际软件中逐步展示过程。幻灯片、miro 板和在说话时的白板绘图也都做得很好。保持元素流动，没人想盯着一张静态幻灯片看。另一种是可视化工具、动态图形、B-roll 和 AI 生成的素材来说明观点。对于露脸视频，你可以录制自己，或雇一张脸。人们从他们信任的人那里购买，镜头上的创始人建立这种信任最快——但这需要练习，并且更难扩展。露脸视频最好的部分是你不需要特殊的编辑就能做得好。脸 + 好脚本 + 引人入胜的传达即使有基本编辑也能赢。第四阶段 - 编辑你可以自己编辑（免费、完全控制、耗时）或雇一个编辑，不幸的是据我所知仍然没有好的 AI 视频编辑器通过推特作品集搜索、Discord YouTube 自动化社区和 Upwork 找到编辑。至于编辑理念：不要过度，保持干净。每次编辑有三个工作。可视化概念（当你说“这个工具自动执行邮件跟进”时，屏幕显示工具正在做这件事）。保持注意力（剪掉空白）。并让你的优惠有吸引力。 ## 让每个视频都成为隐藏的 VSL 当 YouTuber 讨论策略时，他们通常是在优化浏览量。当涉及商业和收入时，浏览量是一个虚荣指标。你应该做的是针对你的特定 ICP 制作高意图视频，并将你的解决方案定位为比他们见过的任何东西都好。浏览量比娱乐优化的内容少，TAM 更小，但转化率高 100 倍。同时教育和推销不要做一个很棒的教育视频并在最后贴上 CTA——“如果你想了解更多，请查看描述中的我的课程”不起作用。视频的整个目的应该与你的优惠解决的问题联系在一起。你的视频必须解决与你的优惠相同的问题（或子问题）假设优惠是 SMMA 课程，视频是“how to do cold outreach for SMMA (step-by-step guide)”（如何为 SMMA 做冷开发（分步指南））。冷开发是运营 SMMA 的一个子问题——任何搜索它的人都试图启动或扩展一个。你逐步教授开发，真正的价值，当你教学时，你自然地将课程定位为完整的解决方案…… 你在教育他们的同时表明你的优惠是结果的捷径。你在提醒他们他们有的问题并提出你有的解决方案。为什么这转化率高 100 倍传统的 YouTube 策略是制作广泛的娱乐内容，获得数百万浏览量，并希望 0.001% 转化高意图策略是为你的确切 ICP 制作特定的教育内容，获得数千次浏览量并以 0.1%-1% 转化。不要私藏最大的错误是因为害怕“给得太多”而保留价值。你给得越多，他们越信任你，他们买得越多。你的视频应该如此有价值，以至于某人可以实施它并在不购买的情况下获得结果。他们应该将你的优惠视为获得 X 结果的捷径。他们购买是因为你在视频中证明了你知道你在说什么。总结的公式：用高意图主题针对你的 ICP，教授你的优惠解决的子问题，始终将优惠定位为完整的解决方案，最后有一个清晰的 CTA，并且不要私藏。 # 渠道 #3: 短视频（Instagram Reels + TikTok）介绍中的快速提醒：这不是我的主要渠道，但我用它进行直接销售和线索挖掘，这是系统。Reels 和 TikTok 运行几乎相同的剧本，所以我将它们作为一个渠道覆盖，并在实际重要的地方指出差异。这些信息大部分来自我的好朋友 @casualhyped，他在短视频方面是个神，我推荐关注他。 ## 两个平台的不同之处在共享剧本之前，这里是两者之间的差异。美国 IP 在 Instagram 上是可选的但推荐——美国账号获得更好的触达，功能优先和更好的变现。在 TikTok 上，如果你试图专注于美国受众，这几乎是必须的。预热 Instagram 需要 1-2 天，你可以在第 2 天积极互动（关注、评论、保存）。TikTok 更严格，你应该花整整 2 天纯粹消费，然后是一个决定账号生死的关注测试。粉丝数在 Instagram 上很重要，但 TikTok 独立测试每个视频——无论你的关注者如何，每次上传都有公平的机会病毒式传播。长度 60-90 秒是 Reels 的最佳点。TikTok 是杠铃形的，要么 5-9 秒，要么 60 秒以上，避免中间 15-45 秒的死角。顶级信号保存是 Instagram 最强大的信号。在 TikTok 上是分享（一次分享值 10+ 个点赞）和重看率（20%+ 会给你巨大的推动）。字幕字幕在 TikTok 上可搜索，它们在那里是 SEO 表面，所以前置你的关键词。其他都是共享的。这是系统。 ## 账号设置旁注：以下内容仅适用于你试图最大化美国受众和多账号的情况，如果不是你可以跳过此部分你可以通过从 anyip.io 获取住宅美国代理（住宅，不是数据中心 - 选择主要城市），在 iOS 上使用 V2Box 应用运行它以及每个账号使用新电子邮件来针对美国。连接代理，在 Safari 中搜索“what is my IP”并在创建任何东西之前确认它显示美国。然后注册一个与细分领域相关的用户名，基于你产品的一行清晰简介以及简介中的漏斗链接。 Instagram 细节对于几个账号，你的家庭 IP 有效——只有当你扩展时，独立的 IP 才变成好习惯。 TikTok 细节代理是强制的，设备设置也是。 - 完全重置你的 iPhone（设置 -> 通用 -> 传输或重置 -> 抹掉所有内容和设置） - 不用 SIM 卡设置，地区美国，英语（美国），纽约或洛杉矶时区 - 创建一个美国 Apple ID - 如果被问到地址从谷歌地图获取，如果被问到从 textverified 获取临时美国号码。 - 与普遍看法相反，美国 SIM 不是必需的，但其他一切看起来必须像一个全新的美国设备。 - 注册时出生日期 25+。在创建账号之前验证 IP 和位置都显示美国——如果不显示，先修复代理。不要创建它并希望如此。 ## 预热两个平台上的新账号都会被标记为潜在垃圾邮件。预热建立一定程度的信任。 Instagram（1-2 天）第一天是被动的，滚动 Reels 20-30 分钟，观看细分领域内容每个 5-10 秒，点赞 15-20 个帖子，不关注也不发布。到第一天结束时，你的信息流应该充满你的细分领域第二天是另外 20-30 分钟，关注 10-15 个细分领域账号，2-3 个真实评论并保存 5-10 个 Reels——然后你就准备好发布了。 TikTok（2 天 + 关注测试）在 TikTok 上做整整两天的纯粹消费，每次 20-30 分钟…… 滚动 FYP，完整观看细分领域视频（完整观看是一个强烈的信号），点赞 15-20 个。两天都不关注也不评论，绝对不发布。然后运行关注测试：尝试关注 1-2 个细分领域账号。如果关注注册了（你的关注数大于 0），那么账号是受信任的，你可以开始发布。如果没有注册，账号被标记了——删除它，获取一个新的代理并重新开始。不要试图挽救被标记的账号。 ## 病毒式公式：4 个支柱两个平台上每个病毒式短视频都遵循相同的结构：钩子、留存、奖励、工程化互动。 ## 支柱 1：钩子（前 3 秒）有效的钩子公式：好奇心缺口：“关于 [主题] 没人告诉你的那一件事”，“知道这个感觉像违法” 震惊声明：“你对 [主题] 的一切认知都是错的”，“停止做 [常见行为]——原因如下” 相关痛点：“你知道当 [挣扎] 时的那种感觉吗？”，“这就是为什么你的 [东西] 不起作用” 基于结果：“我如何在 30 天内获得 1 万粉丝”，“看我如何在 60 秒内把 [X] 变成 [Y]” 模式中断：“等着看……”，“POV：你刚发现……” 钩子也是视觉和听觉的例如，具有对比的第一帧或显示真实表情的脸，粗体文字叠加（最多 5-7 个词，一秒内可读）以及立即开始的热门声音。用不同的钩子测试同一个视频的 3 个版本。仅钩子就可能造成 1,000 次浏览和 1,000,000 次浏览之间的差异。 ## 支柱 2：留存两种算法都首先将观看时间置于一切之上。70%+ 的完成率获得巨大的触达，在 TikTok 上，20%+ 的重读率在此基础上被视为令人上瘾的内容。在 1-2 秒内切入正题。例如，一个好的开场是“这个 AI 工具让我赚了 [金额]。这是方法。” 每 2-3 秒剪辑一次（TikTok 上每 2 秒），观众的大脑将每次剪辑读取为新内容。视觉单调会扼杀留存。使用开放式循环来留住他们（“最好的部分即将到来……”）并使其可循环——结尾流回开头（“一切始于何时……”），以便观众在不知不觉中重看。 ## 支柱 3：奖励在 5-15 秒内兑现钩子的承诺。观众不耐烦，如果他们觉得被诱饵了，他们不会互动。给他们一个可以立即使用的快速胜利，娱乐（即使不是“有用的”笑声也很有价值），顿悟时刻（“等等，这就是它起作用的原因？”），前后转换或验证（“如果你有过这种感觉，你并不孤单”）。尽可能堆叠它们——一个教授 + 娱乐 + 验证的视频优于只做一件事的视频。 ## 支柱 4：工程化互动两种算法都严重权衡互动速度。故意设计它。评论诱饵：以问题结束（“你是哪一种：A 还是 B？”）或热门观点（“不受欢迎的观点：……”）。评论告诉算法这值得分发。分享诱饵：“把这个发给需要它的人”，“标记你的朋友/老板/伴侣”。在 TikTok 上，分享是单一最强大的信号。保存诱饵：“保存这个以免丢失”。参考内容如清单、模板，任何逐步的东西都会自然被保存。在 Instagram 上，保存是最强的信号。重看诱饵（TikTok）：隐藏只在重看时显示的细节，或以需要重看的反转结束。字幕：这些在 TikTok 上可搜索，所以前置关键词——“AI 赚钱技巧：我如何在 30 天内赚了 1 万美元”是一个很好的例子。使用人们会搜索的确切短语并以互动钩子结束（“评论‘AI’我会发给你工具”）。 ## 视频结构模板 0-3 秒：钩子。模式中断或好奇心循环，粗体文字叠加，热门声音已经在播放。 3-5 秒：背景——我们要解决问题的一两句话。 5-10 秒：价值传递，即实际奖励，快速剪辑，文字强化每一点。 10-15 秒：回报。完成承诺，添加反转，循环回开头。在字幕中添加丰富的关键词钩子 + 互动问题 + 主题标签。对于音频，热门声音（在 TikTok 上，理想情况下少于 48 小时）或带有强大开场白的原始音频。 ## 发布时间表 Instagram 第一阶段（0-1k 粉丝）是每天 2 次，间隔 4 小时（比如上午 9 点和下午 1 点）——这确保一致性而不压倒新受众，每个帖子都有自己的算法推送。第二阶段（1k+）是每天多达 15 个试用 Reels——试用 Reels 在向所有人发布前在非粉丝上测试，所以你可以发布大量并只将获胜者发布到你的主个人资料。 TikTok 第一阶段（测试）是每天 1 次，每天同一时间，直到一个视频突破 1,000 次浏览——TikTok 给每个视频 10-30 分钟证明自己，在获得关注力之前发布更多不推荐。第二阶段（扩展）是每天 2 次，间隔 4 小时。前 30 分钟规则（TikTok）初始分发在前 10-30 分钟内决定——这是通过完成率、重看、互动速度和分享计算的。如果你的前几个视频失败了，不要气馁。TikTok 独立测试每个视频——你的突破可能是视频 #5 或视频 #25。 ## 短视频清单 - 账号已预热（IG 上 1-2 天，TikTok 上 2 天 + 关注测试） - 第一条帖子前信息流显示你的细分领域 - 前 3 秒有钩子 - 每 2-3 秒剪辑一次（TikTok 上每 2 秒） - 每个视频都有文字叠加和字幕 - 热门音频（TikTok 上少于 48 小时） - 长度：Reels 上 60-90s，TikTok 上 5-9s 或 60s+ - 5-15 秒内交付奖励 - 互动诱饵：评论、分享、保存或重看从那里，剩下的就是持续发布并将流量发送到你的漏斗。 # 结论写这篇文章花了一段时间哈哈，如果你喜欢，我会感激你的反馈回复这篇文章中的所有内容都是我在 X 和 YouTube 合作项目中亲手运作的剧本。在过去的一年里，我一直在帮助那些擅长自己工作的人——开发者、营销人员、创始人——将他们转化为拥有经过验证的可重复框架的公众人物。它让这里的许多大账号走得很远，有些甚至超过了 10 万粉丝，并在此过程中为我建立了一个庞大的网络。如果你认为你有能力，希望我帮助你建立你的品牌，我很乐意阅读你的申请，看看你能为平台带来什么。在这里申请：eptwts.com （附言：我不收费）希望你喜欢这篇文章。 ## 相关链接 - [@eptwts](https://x.com/eptwts) - [103K](https://x.com/eptwts/status/2082083321383997553/analytics) - [vidiq.com](https://vidiq.com/) - [thumbnailpreview.com](https://thumbnailpreview.com/) - [@casualhyped](https://x.com/@casualhyped) - [anyip.io](https://anyip.io/) - [eptwts.com](https://eptwts.com/) - [Upgrade to Premium](https://x.com/i/premium_sign_up) - [8:38 PM · Jul 28, 2026](https://x.com/eptwts/status/2082083321383997553) - [103.9K Views](https://x.com/eptwts/status/2082083321383997553/analytics) - [View quotes](https://x.com/eptwts/status/2082083321383997553/quotes) --- *导出时间: 2026/7/29 09:35:36*

🔗 相关文档