🀅@ecomchigga
found a guy who reverse engineered the entire twitter algorithm and I'm gonna expose everything he told me
he's not a coder. not some tech genius. just a dude who got mass obsessed with understanding why some tweets blow up and others die.
spent 14 months tracking over 50,000 tweets across different niches. recorded the time posted, hook style, format, length, topic, everything.
built a spreadsheet so big it crashes his laptop.
and what he found goes against almost everything the "gurus" teach.
**finding 1: the first 47 minutes determine everything**
not the first hour. not the first 30 minutes. specifically 47 minutes.
he tested this hundreds of times. if a tweet doesn't get meaningful engagement in the first 47 minutes, it's basically dead. the algorithm gives up on it.
but here's the weird part: engagement after 47 minutes barely matters. a tweet that pops off in the first 47 mins will keep getting pushed even if engagement slows down.
the algorithm makes its decision early and sticks with it.
what this means for you: stop posting and walking away. the first 47 minutes you should be engaging, replying to comments, being active. signal to the algorithm that this tweet is worth pushing.
**finding 2: likes are worth almost nothing**
everyone celebrates likes. they're the least valuable metric.
his data showed:
- 1 retweet = roughly 15 likes in algorithm weight
- 1 comment = roughly 8 likes
- 1 bookmark = roughly 12 likes
- 1 quote tweet = roughly 25 likes
the algorithm wants distribution and conversation. not passive approval.
a tweet with 50 likes and 2 comments will get buried.
a tweet with 20 likes and 15 comments will get pushed.
what this means for you: optimize for comments. ask questions. say controversial things. create content people want to respond to, not just agree with.
**finding 3: threads are dying**
in 2022 threads were king. not anymore.
his data showed thread performance dropped 40% in 2024 compared to 2023. and it's still dropping.
why? people's attention spans got shorter. they see "thread" and keep scrolling. they don't want to commit to 15 tweets.
what's working instead: long single tweets. you can fit a lot of value in one tweet now. and people are more likely to read a long single tweet than click through a thread.
the exceptions: storytelling threads still work if the hook is insane. but educational "here's 10 tips" threads are basically dead.
**finding 4: images hurt more than they help (usually)**
this one surprised me.
he found that tweets with images get 15% more impressions on average BUT 23% less engagement rate.
why? images make people stop scrolling but they also make people feel like they've "consumed" the content without engaging. they look at the image and move on.
text-only tweets force people to read. reading creates investment. investment creates engagement.
the exception: screenshots of results, DMs, or testimonials. those images ADD credibility and increase engagement. random stock photos or graphics do the opposite.
**finding 5: the hook has a character limit within the character limit**
the first 35-40 characters matter more than the rest of the tweet combined.
that's roughly 7-8 words.
his data showed that tweets where the "hook" (the interesting part) appeared within the first 40 characters performed 3x better than tweets where the hook appeared later.
bad: "so I've been thinking about something lately and I realized that making money online is actually pretty simple"
good: "making money online is simple. here's why everyone overcomplicates it:"
same message. second one front-loads the hook.
people decide to keep reading or scroll within the first second. you have 7 words to earn their attention.
**finding 6: posting frequency has diminishing returns after 4**
he tested accounts posting 1x, 3x, 5x, 7x, and 10x per day.
the jump from 1 to 3 posts was huge. almost 3x the growth.
the jump from 3 to 5 was solid. about 40% more growth.
the jump from 5 to 7 was minimal. maybe 10%.
the jump from 7 to 10 actually showed negative returns. less growth than 7.
why? posting too much cannibalizes your own reach. your tweets compete against each other. the algorithm won't show someone 10 tweets from the same person in one session.
sweet spot based on his 4-5 tweets per day. more than that and you're just creating noise.
**finding 7: the bio matters more than any single tweet**
he tracked how often people check profiles before following. it's basically 100%.
then he tracked how many profile visitors convert to followers based on different bio styles.
vague bio ("entrepreneur | dreamer | building cool stuff"): 8% follow rate
specific bio with result ("I help freelancers get clients. grew mine to $20K/month"): 31% follow rate
that's 4x difference from changing 15 words.
your bio is your conversion rate. most people treat it like an afterthought.
**finding 8: weekends are underrated**
everyone says post on weekdays. his data says different.
saturday and sunday had 23% less competition (fewer tweets from serious accounts) but only 12% less active users.
the ratio is better on weekends. your tweets have a higher chance of standing out.
especially sunday evening. that's when people are scrolling before the week starts but most accounts aren't posting.
**finding 9: negativity outperforms positivity (but there's a catch)**
tweets about problems, frustrations, and failures got 2.3x more engagement than tweets about wins and celebrations.
people connect with struggle more than success.
but here's the catch: negativity attracts followers, positivity converts them to buyers.
if all your content is negative, you'll grow fast but sell nothing. people will see you as someone who complains, not someone who has solutions.
the balance: 60% problem-focused content (for growth), 40% solution-focused content (for sales).
**finding 10: replying to yourself works**
tweets where the author replied to themselves within 10 minutes performed 34% better than tweets without self-replies.
why? self-replies count as comments. they signal engagement. they also give you more space to add value without making the original tweet too long.
the move: post tweet, immediately reply with additional context or a question, then go engage with other accounts.
**what he does with this **
he runs 3 accounts now. all between 15K-40K followers. makes around $70K/month combined selling digital products.
doesn't post more than 4x per day. writes all his hooks in under 40 characters. optimizes for comments over likes. posts his best content on sunday evenings.
no fancy tools. no growth hacks. just understanding how the game actually works.
he offered to sell me his full dataset for $2K. I said no because I'm sharing everything here anyway.
the algorithm isn't magic. it's a machine with patterns.
learn the patterns. exploit the patterns. win the game.
