Live data from Hacker News

Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

facebook.com

1–10 of 381 posts

Re: Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

#3
"We built a page logger. Any time a page was loaded, we'd keep track of it. You know what we found? The 80% of clicks we were paying for were from bots."

How do you tell if it's a bot? Just by user-agent string or repeated slamming by IP address, or is there some other technique?

Re: Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

#4
post #3

"We built a page logger. Any time a page was loaded, we'd keep track of it. You know what we found? The 80% of clicks we were paying for were from bots." How do you tell if it's a bot? Just by user-agent string or repeated slamming by IP address, or is there some other technique?

The main technique is to notice a pattern that is only explainable by bot activity.

Generally speaking, in addition to the patterns you mentioned (invalid user-agent strings, rapid succession of same-IP access) there are also other patterns, such as failure to load referenced CSS or image files, or 100% failure on Captcha challenges.

Re: Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

#5
post #3

"We built a page logger. Any time a page was loaded, we'd keep track of it. You know what we found? The 80% of clicks we were paying for were from bots." How do you tell if it's a bot? Just by user-agent string or repeated slamming by IP address, or is there some other technique?

Most likely a combination. In addition to analyzing the user's navigational path, exit points, IP ranges, etc...

Re: Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

#6
post #3

"We built a page logger. Any time a page was loaded, we'd keep track of it. You know what we found? The 80% of clicks we were paying for were from bots." How do you tell if it's a bot? Just by user-agent string or repeated slamming by IP address, or is there some other technique?

They said that for 80% of their clicks , Javascript was off. Now there will be plenty of users who have disabled Javascript but nowhere near 80%.

Although they stated they built their own analytics software, so there's also a chance that it wasn't working as they thought it was.

Re: Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

#7
Wow, even if this is a little true the implications are big.

Is this similar, or related to other ad space inflation? I'm reminded of Google (and others)(1) sending out free $75 adwords coupons to easily allow the costs of bids to inflate since it's not my own real money (thereby increasing the percentage of real cash being made on the ads)?

(1)Edited and clarified

Re: Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

#8
post #4
post #3

"We built a page logger. Any time a page was loaded, we'd keep track of it. You know what we found? The 80% of clicks we were paying for were from bots." How do you tell if it's a bot? Just by user-agent string or repeated slamming by IP address, or is there some other technique?

The main technique is to notice a pattern that is only explainable by bot activity. Generally speaking, in addition to the patterns you mentioned (invalid user-agent strings, rapid succession of same-IP access) there are also other patterns, such as failure to load referenced CSS or image files, or 100% failure on Captcha challenges.

"failure to load referenced CSS or image files" -- is there a way for bots (in particular, mechanize in perl/python) to also download those files?

"100% failure on Captcha challenges" -- are captchas generally deployed on click-through ads?

Re: Company withdrawing from Facebook as analytics show 80% of ad clicks from bots

#9
I wish their post was a bit more technical. Show us the code you used to detect javascript, how did the analytics work? Did they test their code by using their own browser to click an ad and see if the code detected javascript, or their hit in the analytic software?

More technical info and data would be nice!

Post reply on HN