Threads algorythm breakdown pt.2

[lissaaa]

Hey,

the phone farm is almost optimized. We´ve noticed that if we copy content 1 by 1 even with spoofing we go viral every time. BUT we dont know why that happens. Does the algorythm rank you higher if you do that? Or is that cause the post is just good in general? (good engagement bait, ...). If we recreated the content with our own pictures we dont get as much traffic. We are testing right now to use different captions and same pictures and samce captions but different captions from viral posts. It seems like same pictures is performing even better. BUT it looks like meta is not really scanning the content for what it really is. (Only nude check and if its mass duplicated over many many accounts). So whats the best plan? Just copy viral captions and optimized the pictures? It feels like pictures are just your multiplier after the caption allows you to get tested if people are making engagement (scroll back, clicks, comments, ...). Does anyone have some knowledge about how meta handles the content exactly? And how to perfectly abuse viral captions? Also does meta slow down certain copy paste captions after some time maybe? If too many people reuse the same ones.

Thanks for the replies!

 

[ToyBox Marketing]

In case you continue to post duplicates, the other accounts could get flagged. It is important you change the hash and meta data for each upload

 

[gg11]

where are you buying your threads accounts & whats your warm up process?

 

[lissaaa]

where are you buying your threads accounts & whats your warm up process?

im not buying we have automated creation

 

[Dragon Tia7]

Viral captions trigger fast engagement so the post gets pushed. Images just amplify once its being tested. Reusing the same captions works until engagement drops then they slowly lose power.

 

[lissaaa]

Meta doesn’t rank copied content higher by default. What you’re seeing is usually because proven viral captions familiar visuals trigger faster early engagement, which pushes the post into testing pools. Captions act as the hook, and images/videos act as a multiplier for retention and interaction. Recreated content often performs worse because it removes familiarity and proven engagement patterns.


Meta mainly reacts to user behavior signals (watch time, scroll-backs, comments, shares), not deep semantic understanding of captions. Reused captions can work, but overuse eventually weakens them as engagement drops or duplication signals increase.


In short: performance comes from engagement mechanics, not copy-paste magic.

Thank you! Very informative.

Yes we´ve seen this before. But somehow we cant really control yet how big the testing pool will be. It seems like if we copy 1 by 1 we get an insane big one since views can sometimes absolutely explode. We´ve tried to hook up to some trends but only went viral a few times. The goal is to get everything automated so we can create automated viral posts so every accounts brings about 250k+- views a day. We were at those numbers before but now it dropped to only 20% of accounts hitting those numbers on a daily basis. Do you have any idea how the "spoiler" tool works with the algorythm? Is that doing anything with how the post gets ranked? OR is it just a feature to maybe make more engagement bait but not really affecting the algorythm or how much he will test your post

 

[neylontouria]

It's great to see the discussion around content performance and engagement mechanics. I totally agree that viral captions and familiar visuals can trigger faster early engagement, which in turn pushes the post into testing pools. Can you guys share more about your strategies for creating automated content and warming up newly created accounts, as gg11 mentioned they have an automated creation process - what tools or methods are you using for this?

 

[lissaaa]

It's great to see the discussion around content performance and engagement mechanics. I totally agree that viral captions and familiar visuals can trigger faster early engagement, which in turn pushes the post into testing pools. Can you guys share more about your strategies for creating automated content and warming up newly created accounts, as gg11 mentioned they have an automated creation process - what tools or methods are you using for this?

Selfmade UI from our dev. Connected to phone farms which make the creation process. Cant say too much for the warmup since everyone has their own methods to this. But a huge tip is to have everything organised meaning make sure your setup + devices are ultra clean and dont have any anomalies in fingerprint/proxy and all of that. Thats like 90% of the work. Also dont listen to warmup tips from youtube 99% of them have 0 clue what they are doing. Just test test test test. Its the hard way but the better one and after that you have your own method which no one can steal and the best thing is you have knowledge and experience on how everything works. Thats what weve done and we just learn more and more every day

 

[gideyirangifb0r]

it’s less about copy-paste being “rewarded” and more about proven engagement patterns getting tested faster. Once a caption or format already triggers strong signals, the same visuals seem to amplify it, but Meta usually throttles things once duplication becomes too obvious over time

 

[Yoa Elizabeth]

Hey,

the phone farm is almost optimized. We´ve noticed that if we copy content 1 by 1 even with spoofing we go viral every time. BUT we dont know why that happens. Does the algorythm rank you higher if you do that? Or is that cause the post is just good in general? (good engagement bait, ...). If we recreated the content with our own pictures we dont get as much traffic. We are testing right now to use different captions and same pictures and samce captions but different captions from viral posts. It seems like same pictures is performing even better. BUT it looks like meta is not really scanning the content for what it really is. (Only nude check and if its mass duplicated over many many accounts). So whats the best plan? Just copy viral captions and optimized the pictures? It feels like pictures are just your multiplier after the caption allows you to get tested if people are making engagement (scroll back, clicks, comments, ...). Does anyone have some knowledge about how meta handles the content exactly? And how to perfectly abuse viral captions? Also does meta slow down certain copy paste captions after some time maybe? If too many people reuse the same ones.

Thanks for the replies!

Early engagement patterns are key testing slightly different captions while keeping visuals compelling often extends reach without relying on exact copies that the system may devalue over time.