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 $10k. 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
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.
