The game is as jiggy said. Force competing ad buyers to pay for what appears to be non-productive traffic by forcing ad clicks. Since the real views are dominated by false views, the conversion rate on whatever is being sold appears to be horrible. The competitor goes elsewhere, and the scammer maintains access to artificially cheap ad inventory, since there are fewer bidders.
It's a well known problem. Considering AdWords is Google's primary source of revenue, they work hard at it. For a tiny bit more directly from Google, see: http://support.google.com/adwords/bin/answer.py?hl=en&an...
By far the most likely explanation in my opinion. Even if it's not the correct one, some of the traffic is almost certainly these kinds of bots.
I don't really suspect this to be true, Facebook CTR are incredibly low so if they are faking it they are doing it badly.
I suspect the most likely reason is some sort of spam scraper thing that visits every url or something. I would have assumed Facebook would be well protected against this sort of thing in regards to charging advertisers.
[1] Bot registration on FB is really easy. If you try register 100 fake users in direct way you need to deal with javascript and web forms with strange ids, cookies and so on in your script. But if you enter fb.com without enabled javascript, they ask you change browser or use mobile version, which is not require* js at all.
[2] You can register number of accounts from one ip without CAPTCHA
[3] Only one "difficulty" is to find batch of emails, it is not that big issue since we have services like mailinator.com.
So 15 -25 loc script on python do all your job. It is not required use any tools like selenium or special skills , only basic understanding how http get/post works.
Why fb doing all this super easy bot registration stuff? I think we all know why.
Edit: spelling. Sorry non native English speaker here
The article implies they could be used by facebook or competitors to burn up ad revenue, but this approach would just kill the golden pay-per-click goose.
[1] You can win a bunch of contests, competitions like those who have lot of likes can win iphone or other tech.
[2] You can sell likes.. just google and you will see, there are alot of places where you can buy it.
[3] Manipulation of public opinion.
Edit:
From FB side:
[1] they still have growth of users
[2] more bots == more clicks, more likes more FB value "on paper", more revenue
[3] and when they stabilize their market value they can easily kill those bots, just require them activation by phone.
vk.com Russian social network, had same issues, but they enabled phone activation for averybody and its not that easy to buy a lot of phone numbers in Russia, but some hackers just pay to vk.com admins...
Edit2: spelling
So Facebook is knowingly profiting from this, and they have the resources to fix this...but they don't since this is making so much money for them.
That doesn't seem like a very sustainable strategy though.
Does anyone know what kind (if any) of anti-bot measures Facebook already has in place for paid advertising?
Assuming the problem is as bad as many are saying, I personally think it's fair to draw the same conclusion you have -- running and letting run a botnet is one and the same. It's costing other businesses money; it's too big for Facebook to be able to play dumb on the issue.
Advertisers monitor to see performance on the other end, and when the campaigns don't yield actual success they move on.
The only real exception to this are mega-brands that advertise just to build namespace. Everyone else expects direct action from some hefty percentage of their ads.
Imagine a 100 person company with 5 people dedicated to advertising. Leadership says to increase on-line presence. Some recent college grad on the advertising team buys FB ads. Next month leadership is pleased that they're advertising on Facebook. Also leadership is pleased that sales increased x%. The FB ads could be completely worthless. Some non-trivial percentage of clicks could be from bots and no one would notice.
Most ads target a specific outcome: You buy a product, create an account, joining a mailing list, etc. Even the most technically incompetent advertisers establishes these fundamental metrics such that they can say "I spent $1,000,000 on Facebook ads and only got ten signups. I spent $1,000,000 on iAds ads and got ten million signups." That sort of thing.
What they (Facebook) are doing now seems to be some version of DADT (Don't Ask, Don't Tell).
Does me pointing out the barely shrouded subtext mean that I believe what they are saying? Of course not. It's just as likely, it seems, that their test is simply flawed (as others said, let's see the technical honey trap), that it's some other technical fault, etc. I highly doubt, just on intuition, that Facebook is running such broken bots: Come on, they'd do a better job than make something so easily detected.
That to me sounds like FB is aware of the issue and isn't interested in discussing it. That's the subtext imo.
My own suspicion would be because someone is gaming their ads and they haven't been able to prevent it as of yet, and not because they themselves are gaming the numbers. No doubt they believe that discussing it is tacit admission that they haven't been able to prevent it, which is bad for business.
And it doesn't matter who it is, because they're leaving for a different reason: Facebook never returns their calls unless it's with an offer of extortion.
Is it crawlers crawling Facebook? Are they alleging that facebook is doing it to boost revenues?
Their hypothesis seems secure, given that you can't really use the facebook site at all without javascript enabled.
More alarming is Facebook holding their name for ransom for the sum of $24,000 a year!