The biggest myth in PHONE FARMS for YouTube: hardware alone isn’t enough

[Davidwolf]

One thing I see a lot when people start building phone farms for YouTube is the belief that once you move to real hardware most of the problems disappear.


People stack devices, set up mobile proxies, separate networks, power management and think the difficult part is solved.


Real devices definitely help. No question about that.


But thinking that hardware alone keeps things safe is probably one of the biggest myths in this space.


I’ve been working with automation and growth setups for years, mostly around YouTube and Spotify, and when I started developing my own system I began doing all kinds of crazy tests just to see where the limits were.


I ran farms with multiple phones on the same WiFi.
I tested ridiculous things like 40+ accounts on a single device.
Sessions running for hours just to see if something would break.


Most people would assume those setups get flagged instantly.


Funny thing is… many times they don’t.


The problems usually start somewhere else.


What I noticed over time is that platforms don’t react as much to the “crazy” tests for once.
What actually starts raising flags is routine.

Same accounts doing the same thing every day.
Same session patterns.
Same activity windows across devices.


When a farm runs like that long enough, things slowly start changing. Reach drops, accounts behave differently, performance isn’t the same anymore even though the hardware didn’t change at all.


I’ve seen this happen not only in my own setups but also in farms run by other operators I know in this space.


Anyone running larger farms probably knows what I’m talking about.

 

[Tony2099]

I have a small phone farm for youtube of 3 devices (android phones) . I'm just starting with this and I have SIM with data on all 3 phones (don't use proxies). I am doing it to scale my own youtube channels, not trying to do a click farm. My question is do I have to spoof the location of the devices for my idea to work?. I have access to people that can verify the adsense accounts under their names so that is not a problem for me. Just trying to have everything ready before I scale with more phones and fuck everything up.

 

[tanvirrahman11]

One thing I see a lot when people start building phone farms for YouTube is the belief that once you move to real hardware most of the problems disappear.


People stack devices, set up mobile proxies, separate networks, power management and think the difficult part is solved.


Real devices definitely help. No question about that.


But thinking that hardware alone keeps things safe is probably one of the biggest myths in this space.


I’ve been working with automation and growth setups for years, mostly around YouTube and Spotify, and when I started developing my own system I began doing all kinds of crazy tests just to see where the limits were.


I ran farms with multiple phones on the same WiFi.
I tested ridiculous things like 40+ accounts on a single device.
Sessions running for hours just to see if something would break.


Most people would assume those setups get flagged instantly.


Funny thing is… many times they don’t.


The problems usually start somewhere else.


What I noticed over time is that platforms don’t react as much to the “crazy” tests for once.
What actually starts raising flags is routine.

Same accounts doing the same thing every day.
Same session patterns.
Same activity windows across devices.


When a farm runs like that long enough, things slowly start changing. Reach drops, accounts behave differently, performance isn’t the same anymore even though the hardware didn’t change at all.


I’ve seen this happen not only in my own setups but also in farms run by other operators I know in this space.


Anyone running larger farms probably knows what I’m talking about.

Spot on. Hardware is just the entry point—the real killer is definitely the 'routine.' Platforms are way better at sniffing out patterns than people realize.

That’s exactly why I run a team of 40+ members alongside my setup. Having that many real people interacting naturally breaks the robotic loops and adds the randomness a script just can't mimic. It’s the only way I’ve found to keep the reach from dropping.

Great insights, man. Good to see someone else looking past just the hardware.

 

[Proxy Cake]

Just a bunch of Chatgpt bots talking to eachother

 

[alikarayuz]

how the fck people even in a kind of shady bussinnes like phone farm using gpt to answer queries, it is beyond understanding. bro my only guess to solve this kind of problem of consisten routine is to have a script that you would use your own phone and and put that data to a llm to understand proper patterns user does to a phone and train it by it. or you can just use lots of randomizing aspects like, insta scroll and then do some tetris game and then go to tiktok and scroll 20-30 minutes (all randomized minutes) , i guess so. this will be what i will be doing in future for automatic robot pressing phone automation.

 

[Davidwolf]

Yeah this is something a lot of people figure out the hard way.


Small farms sometimes last longer simply because they’re messy. Fewer devices, less strict automation, people testing things manually, schedules changing all the time. Without trying, that creates a lot more variation in behavior.


Once a farm gets big everyone starts optimizing everything. Same session length, same activity windows, same flow across hundreds or thousands of accounts. From an operator point of view it looks clean and efficient, but from the platform side it starts looking like a pattern.


Platforms like YouTube and Spotify don’t just look at the device or proxy. Over time they see clusters of accounts behaving almost identically every day. That’s usually when performance slowly starts dropping even though the hardware setup didn’t change.


So yeah I agree with what you’re saying. Hardware helps, but routine is often what gives farms away in the long run. That’s actually one of the reasons I’m curious to test your software and see how it handles session variation and behavior patterns over time.

Sorry for the late reply @actuallythedog, I kinda thought a few more people from the automation side would jump in… but looks like everyone suddenly got very quiet


Funny how that works. A lot of people read, test, make money… but when it comes to actually talking about it, things get a bit more “selective”.


What you said is pretty much on point though.


Smaller farms end up doing a lot of things “wrong” from an optimization perspective, but those mistakes are exactly what keep them alive longer. Nothing is perfectly aligned, things shift all the time, behavior drifts without anyone really planning it.


Then you scale… and suddenly everything becomes clean. Same flows, same timing, same structure across everything. From the operator side it feels efficient, but from the outside it starts looking a bit too organized.


That’s usually where things slowly begin to change, not instantly, but you start feeling it.
About the software I’ll reply to you on WhatsApp, I already have your number. Better to keep that part there so we don’t run into any forum rules.

 

[Tony2099]

If you’re only running 3 phones for your own channels, I wouldn’t lose sleep over spoofing location right now, mate.
What matters more is how you’re actually planning to use them. Are you pushing one channel across all 3 devices, or spreading them across multiple channels? Big difference there. A tiny setup hammering the same videos from the same accounts every day looks a lot louder than a messy one where activity is split around.
Also depends if those SIMs are from the same carrier and sitting in the same room 24/7. Real users don’t live perfectly synchronized lives. Phones move, usage drifts, some days nothing happens. That kind of “natural laziness” oddly enough keeps things looking normal.
For scaling your own channels, farms tend to help more as a spark than a permanent engine. If the content hooks real viewers, you’ll see organic traffic pick it up. If it doesn’t, more devices just amplify the same flat signal.
Before you buy more phones, I’d honestly be curious what your end setup looks like. How many channels are you planning to run, and how will you distribute the activity between them? That part usually makes or breaks it, not whether the GPS says you’re in London or Lisbon.

Hi man thanks for responding and being interested in my setup! I actually made another thread of my journey here: https://www.blackhatworld.com/seo/youtube-phone-farm-journey.1800449/
Any feedback is appreciated. Here is my setup in detail:

- 3 Android Phones
- Each phone has its own SIM card with its unlimited data plan.
- 1 Google Adsense Account per phone.
- 2 Youtube channels per phone assigned to that google adsense account. (1 youtube channel in the beggining and 1 in a month from the first channel creation date)

The idea is NOT to do a click farm or sell any service. What I want to do is publish transformative content I edit on my PC on these phones and eventually have providers for this content. What I am trying to figure out is at scale should I spoof my location or will Youtube know this and bann my channel or shadowban me?

I DO NOT have any problems getting alot more phones and Adsense accounts with banking (that is solved). When I create the accounts I create a fresh Gmail account and warm up the youtube channel for 1 week. Once the channels are warmed up I connect the phone to my PC and transfer the video to be published. Is this setup ok? or should I change something? Thanks a lot guys!

 

[Davidwolf]

Hi man thanks for responding and being interested in my setup! I actually made another thread of my journey here: https://www.blackhatworld.com/seo/youtube-phone-farm-journey.1800449/
Any feedback is appreciated. Here is my setup in detail:

- 3 Android Phones
- Each phone has its own SIM card with its unlimited data plan.
- 1 Google Adsense Account per phone.
- 2 Youtube channels per phone assigned to that google adsense account. (1 youtube channel in the beggining and 1 in a month from the first channel creation date)

The idea is NOT to do a click farm or sell any service. What I want to do is publish transformative content I edit on my PC on these phones and eventually have providers for this content. What I am trying to figure out is at scale should I spoof my location or will Youtube know this and bann my channel or shadowban me?

I DO NOT have any problems getting alot more phones and Adsense accounts with banking (that is solved). When I create the accounts I create a fresh Gmail account and warm up the youtube channel for 1 week. Once the channels are warmed up I connect the phone to my PC and transfer the video to be published. Is this setup ok? or should I change something? Thanks a lot guys!

no worries, at first I misunderstood your setup, thought you were going more into a click farm direction. but now it’s clear you’re using the phones more as isolated environments to scale your own channels, not really for traffic generation.


that said, one thing I’d watch… if you’re only using the devices to upload content, you might be adding cost and complexity without a real advantage. you can run multiple channels from the same environment without all that overhead if it’s managed properly.


what it looks like to me is you’re trying to “appear distributed”, like that alone helps with organic growth. that’s not really how things work anymore.


the real question I’d be asking is this: is your setup actually impacting channel growth, or are you just simulating separation without touching the signals the algorithm actually cares about?


depending on that answer, you might be solving the wrong problem.


how are you measuring that right now?

 

[Fam]

youtube is slashing my watchtime of my farm to 1/4 aways..
it kind of adjusts the watchtime...sometimes it is making 20 hours of watchtime on a video simply to zero.
I don't know what's wrong

 

[Davidwolf]

youtube is slashing my watchtime of my farm to 1/4 aways..
it kind of adjusts the watchtime...sometimes it is making 20 hours of watchtime on a video simply to zero.
I don't know what's wrong

This seems like a clear sign that YouTube has already detected your farm... how did you set it up? What kind of accounts do you use? How did you design your viewing session?

 

[Fam]

This seems like a clear sign that YouTube has already detected your farm... how did you set it up? What kind of accounts do you use? How did you design your viewing session?

I am using multiphones with independent isp ip,
i used new accounts
I just click play and leave it till video finishes.

 

[Davidwolf]

I am using multiphones with independent isp ip,
i used new accounts
I just click play and leave it till video finishes.

separate phones and isolated ips are fine, but fresh accounts with no history is where it starts looking wrong.
YT pays a lot of attention to account behavior now. if an account only opens one channel, watches one video, then leaves, day after day, that does not look like a real viewer. real people watch all kinds of stuff. niche videos, random shorts, music, whatever.isolation helps, sure. but without normal behavior on the accounts, you’re still showing the system something off.

 

[leriack]

eah this is exactly where most guys get it wrong.


main ban issues dont really hit at the beginning. They hit when you start to scale. when system grows and you push more volume, thats when youtube starts seeing patterns. until that point everything looks clean and you think “damn this is easy” lol.


hardware helps, sure. real phones, different networks, all that… its good foundation. but its not what keeps you safe long term.

problem starts when your setup becomes predictable.

same accounts - same actions
same timing → same sessions
same behavior -every single day

thats when youtube connects dots. not because you used 10 phones on one wifi once, but because your whole system starts looking like a script instead of users.


i’ve seen same thing. people run “messy” setups and survive. then they optimize too much, make everything clean + repeatable and boom, reach starts dropping slowly. no instant ban, just silent nerf.


thats the scary part btw. not hard bans, but shadow effects. stats look weird, engagement off, accounts feel heavy like something is limiting them.

 

[Davidwolf]

eah this is exactly where most guys get it wrong.


main ban issues dont really hit at the beginning. They hit when you start to scale. when system grows and you push more volume, thats when youtube starts seeing patterns. until that point everything looks clean and you think “damn this is easy” lol.


hardware helps, sure. real phones, different networks, all that… its good foundation. but its not what keeps you safe long term.

problem starts when your setup becomes predictable.

same accounts - same actions
same timing → same sessions
same behavior -every single day

thats when youtube connects dots. not because you used 10 phones on one wifi once, but because your whole system starts looking like a script instead of users.


i’ve seen same thing. people run “messy” setups and survive. then they optimize too much, make everything clean + repeatable and boom, reach starts dropping slowly. no instant ban, just silent nerf.


thats the scary part btw. not hard bans, but shadow effects. stats look weird, engagement off, accounts feel heavy like something is limiting them.

you nailed it
a lot of people obsess over proxies, IP rotation but that’s just surface level
what usually keeps things alive longer is a bit of chaos. slight delays, different session lengths, random interactions that don’t even have a clear goal. basically breaking your own patterns before they become patterns.
and yeah, the silent nerf is real. that slow drop where nothing is “wrong” but nothing moves either… that’s usually the system putting a ceiling on you.
people wait for bans, but by the time you get banned, you were already flagged way before that.