I forwarded my post about fake social signals and the potential danger of buying 2000 likes on fiverr to the guy that made the suggestion to me and he asked a few more questions.
One of his questions was about how someone on Fiverr can add 2000 “likes” to a page so rapidly. The answer is they need to be using some sort of “bot” or automated software. These usually use proxies and multiple spam accounts that can go off and generate the 2000 likes with a single click once some basic data has been setup.
Codename:Like shown in a screen shot below is probably what they would use.
I have not used this package but used many similar automated tools. They are brilliant in how they work and what they do but I am convinced that they are not the way to go and could actually have a detrimental affect on the popularity of the page.
How They Work And Why We Will Never Use Them
You need to provide three types of data to setup this bot.
- List of proxies (private or harvested free ones)
- list of Facebook accounts to use
- URL of the target(s) you want “liked”
The proxies are used so that each Facebook account providing the social signal has a different internet address making it appear to come from a different location. This works fine as long as the proxies are highly anonymous. We take extra care when creating social profiles by allocating each of our profiles with a private proxy from a location that matches the residential address of the profile. It’s this attention to detail that will help keep our profiles under the spam detection radar.
The Facebook accounts being used for the 2000 likes will also be created using some sort of automated bot. We manually create all our social accounts and invest time in giving them all a unique human touch that an automated script can’t achieve.
The auto-generated Facebook accounts will have been created in a cookie cutter fashion and will be easily detected and flagged as a spam account despite attempts to add individual pictures and other unique account information.
I also noticed that the bot allows you to add delays between actions which is an attempt at emulating human activity.
Does the bot fool Facebook?
The bot is trying to fool Facebook. The fact that it uses proxies and time delays to emulate human activity is evidence of this. The question is does it really fool Facebook or just temporally circumvent the anti-spam measures Facebook have in place?
It was Google that first started to talk about “natural” activity when it launched its war on link spam. They have invested a lot of effort in creating systems to analyze online activity and detect what it considers to be unnatural activity. It would be naive in the extreme to think that Facebook have not done the same. In fact there is plenty of evidence to show that this is exactly what they are doing.
Bots can’t emulate all real human activity
It’s not just about identifying accounts that look spammy but they are also able to review the activity on the account to determine if it appears natural. To really emulate human behavior the bot would need to log into the Facebook account for extended periods and engage in a range of activities unrelated to the generation of social signals. These activities may include playing games, personal messages, chatting, browsing the profiles, reading posts and viewing photo albums. To my knowledge these are not activities that are easily emulated by bots.
Consider how easy it would be for Facebook to create a system to review the online activity of accounts and flag those accounts operated by non-humans. It should be easy to examine these accounts to determine if the bot activity is legitimate such as service like hootsuite or a black hat bot engaged in spam.
Natural velocity is monitored
We have not yet considered the target account. There is nothing natural in getting 2000 new likes in a day or so then back to zero activity. Generally growth occurs with gradual and consistent velocity. Two thousand people do not suddenly wake up one day and decide to open their Facebook account solely for the purpose of “liking” something and then log off again. It should also be relatively easy for a system to distinguish between something going viral and getting a burst of real likes and this type of social spam.
This activity on the target account alone should trigger an alarm. If I were working at Facebook on removing spam I would flag not just the recipient account but also the accounts providing the social signals.
Easy for Facebook to identify spam accounts and fake social signals
When you think about it identifying spammy social accounts and social spam on their network would be easy for Facebook and the increased number of existing accounts now being issued with anti-bot challenges and requiring SMS verification would suggest this is exactly what they are doing.
The systems for detecting social spam will continue to improve and the black hat people will continue to try to out smart them. This is not a game we are interested in playing.