How Instagram deals with Automation

[Tommy T]

Saw this blog post about the Instagram changes and thought it was interesting. The facebook research paper is also a super interesting insight into the new follow blocks people have been experiencing.

https://blog.goldnitro.com/2019/06/08/how-instagram-deals-with-automation/
https://research.fb.com/publication...rizing-account-automation-abuse-and-defenses/

 

[tubatesmal]

Thats actually a nice share reading it rn8!

 

[Chris.Roark]

Great mate! Thank you!

I might add this to your two links:
https://techcrunch.com/2018/11/19/instagram-fake-followers/
https://instagram-press.com/blog/2018/11/19/reducing-inauthentic-activity-on-instagram/

 

[GenesisOne]

I don't believe their stats.
I have child accounts running bellow their suggested threshold that are experiencing follow blocks.

 

[ProGIG]

Saw this blog post about the Instagram changes and thought it was interesting. The facebook research paper is also a super interesting insight into the new follow blocks people have been experiencing.

https://blog.goldnitro.com/2019/06/08/how-instagram-deals-with-automation/
https://research.fb.com/publication...rizing-account-automation-abuse-and-defenses/

After reading all this and

Great mate! Thank you!

I might add this to your two links:
https://techcrunch.com/2018/11/19/instagram-fake-followers/
https://instagram-press.com/blog/2018/11/19/reducing-inauthentic-activity-on-instagram/

this I am in delimma whether I should start with Botting or not, any help will be appreciated...!!!

 

[tubatesmal]

I don't believe their stats.
I have child accounts running bellow their suggested threshold that are experiencing follow blocks.

Obviously there are more facts besides the following & like treshould. This article have 1 year. They should probably already included other data to predict bots traffic & quick block the actions instead.

 

[Tommy T]

I don't believe their stats.
I have child accounts running bellow their suggested threshold that are experiencing follow blocks.

Which threshold did you try using? It also appears that which proxy you are using will also influence your threshold.

 

[IG Professor]

Great share. this is a must read from anyone who is doing social media marketing. even if he don't automate
Ty

 

[GenesisOne]

Which threshold did you try using? It also appears that which proxy you are using will also influence your threshold.

That's my guess too, and the reason I'll start using mobile proxies.

 

[Tommy T]

That's my guess too, and the reason I'll start using mobile proxies.

Let me know how that works. I've heard that 200 follows per day is now the recommended speed. We've seen some success with accounts on this speed setting.

 

[Julian_Si]

Yes, 200 is the number most people recommend these days.

 

[slickkilla]

I don't believe their stats.
I have child accounts running bellow their suggested threshold that are experiencing follow blocks.

Which stats don't you believe. If its this research papers stats
https://research.fb.com/publication...rizing-account-automation-abuse-and-defenses/
Then its crazy not believing stats from Instagram employees themselves

In that paper they do not mention a suggested threshold they clearly state that the threshold being implement would be a subnets 25th percentiles amount of follows per day on a subnet that only has botted accounts(Datacenter Proxies) or the 99th percentile if the subnet has a mix of real users and botted accounts (Mobile proxies).
But this research paper was from around Fall of last year so i could imagine they evolved on this and now the threshold is a mix of the above and other flags raised by an account.

 

[Heisenberg's Hat]

Which shows they know exactly what's going on and that they know more about botting than most who use a bot.

They could easily reduce the fake accounts of celebrities to rubble - but they also need those big accounts.

The algorithm only uses activity as a measurement of quality, which is another reason why an artist creating great content and using Instagram just as is has no chance of gaining enough followers to reach a wider audience.

In addition, 60% of internet traffic is already bot traffic, and the amount of this on Instagram should be much higher. Cutting this traffic and deleting the fake accounts (which Instagram can perfectly identify) would lead to a drop of activity and "active users" and hurt the network.

What I think Instagram will be doing is to try to get the worst trash posting spam accounts out of the way and those who use bots too much.
They will raise the financial bar to use a bot, so not everybody can buy a bot like Instazood for 10$. I expect Instazood to be shut down any moment.

They will still allow robots for quality accounts as they need those quality accounts, and quality accounts cannot be established without robots, as no artist in his right mind will spend 8 hours a day clicking on Instagram.

Instagram without good content is dead, and already artists are very discontent with how bad they are treated as true content providers by Instagram.

The solution would be if an AI based algorithm could distinguish good from bad photography - it can't be done perfectly, but you can likely select out about 95% of the posts, the selfies, the food on plate images, and all the other junk that drags down Instagram more than the bots do.

 

[GenesisOne]

Which stats don't you believe. If its this research papers stats
https://research.fb.com/publication...rizing-account-automation-abuse-and-defenses/
Then its crazy not believing stats from Instagram employees themselves

In that paper they do not mention a suggested threshold they clearly state that the threshold being implement would be a subnets 25th percentiles amount of follows per day on a subnet that only has botted accounts(Datacenter Proxies) or the 99th percentile if the subnet has a mix of real users and botted accounts (Mobile proxies).
But this research paper was from around Fall of last year so i could imagine they evolved on this and now the threshold is a mix of the above and other flags raised by an account.

I don't need to believe anything. I talk from my OWN stats. Which I find as a better source of knowledge than just BELIEVING.

 

[slickkilla]

I don't need to believe anything. I talk from my OWN stats. Which I find as a better source of knowledge than just BELIEVING.

Lol quite ridiculous not believing a research paper from Instagram themselves on something regarding exactly what you are having issues with.
Yes your OWN stats might show that your accounts get blocked after 2000 follows or what ever amount it get blocked at but again as i mentioned, that research paper never gave a "suggested threshold" they stated how they come up with the threshold to place on an account.
So you saying that you don't "believe" their thresholds doesnt make sense when they done mention any specific number in the research paper.

 

[boseki]

I don't need to believe anything. I talk from my OWN stats. Which I find as a better source of knowledge than just BELIEVING.

Congrats on sharing the mindset and intelligence of a 12-year old

 

[Tommy T]

Which shows they know exactly what's going on and that they know more about botting than most who use a bot.

In addition, 60% of internet traffic is already bot traffic, and the amount of this on Instagram should be much higher. Cutting this traffic and deleting the fake accounts (which Instagram can perfectly identify) would lead to a drop of activity and "active users" and hurt the network.

What I think Instagram will be doing is to try to get the worst trash posting spam accounts out of the way and those who use bots too much.
They will raise the financial bar to use a bot, so not everybody can buy a bot like Instazood for 10$. I expect Instazood to be shut down any moment.

They will still allow robots for quality accounts as they need those quality accounts, and quality accounts cannot be established without robots, as no artist in his right mind will spend 8 hours a day clicking on Instagram.

Instagram without good content is dead, and already artists are very discontent with how bad they are treated as true content providers by Instagram.

I tend to agree. It's going to be interesting to see how this develops.

 

[Br Stoner]

Saw this blog post about the Instagram changes and thought it was interesting. The facebook research paper is also a super interesting insight into the new follow blocks people have been experiencing.

https://blog.goldnitro.com/2019/06/08/how-instagram-deals-with-automation/
https://research.fb.com/publications/following-their-footsteps-characterizing-account-automation-abuse-and-defenses/

Think this thread is really outdated referred, 3 months happens a lot of things on the algorithms of any social media and these notices have more than 7 months. Some information is available, but you cant use their references now because they already adapted to new versions

 

[Tommy T]

Think this thread is really outdated referred, 3 months happens a lot of things on the algorithms of any social media and these notices have more than 7 months. Some information is available, but you cant use their references now because they already adapted to new versions

That's definitely true. Some of the behavior we're seeing is definitely different than what's described.

The interesting takeaway for me is more their philosophy for tackling automation. It seems that they 1) run multiple experiments 2) use different kinds of time windows 3) enforce blocks and 4) use delayed removal. IMO it's safe to assume that IG will still rely on those tools in their toolbox.

 

[MertOmerta]

Thanks for sharing