Re:Google and Facebook ad traffic is 90% useless
by revicon on Sunday February 26, 2017 @01:23PM
(#13738729)
This line stuck out for me...
Also when I turned off "Google Partners" and "external
sites" from our campaigns, the ratio of idiot-clicks went
down dramatically. I believe that most of the useless ?
traffic (i.e.: bots or confused users) come from sites
external to Google.
Most of the adwords best practices docs I've read, and going through the udemy course Isaac Rudansky put out on adwords, it sounds like the extended google partners network is best left turned off for all adword campaigns.
Re:Google and Facebook ad traffic is 90% useless
by Animats on Sunday February 26, 2017 @01:18PM
(#13738686)
Google's third-party ads are known to be almost useless. Referrals directly from a Google search have some value. But this article goes beyond that. It indicates that most clicks on third party ads are coming from bots.
It would be useful for a few sites to publish IP addresses and times from which they get bot-like traffic. Maybe a pattern will emerge.
Re:Google and Facebook ad traffic is 90% useless
by ssharp on Sunday February 26, 2017 @01:34PM
(#13738798)
A simple way to try and block out bots is to just not allow your ads to show to Linux users. It's a tiny fraction of most website's traffic but the largest share of bot traffic. I'd further speculate that real desktop Linux users are substantially more likely to have an ad block running anyway.
Re:Google and Facebook ad traffic is 90% useless
by cannonpr on Sunday February 26, 2017 @01:48PM
(#13738909)
Eh, bots these days use a good distribution of user agents strings and you can define the type when buying traffic typically. There are ways of finger printing browser engines but that's not done for ad bidding.
Plus even if it comes into play bots will just shift to virtualised windows stacks.
Re:Google and Facebook ad traffic is 90% useless
by fictioncircle on Sunday February 26, 2017 @01:58PM
(#13739007)
A non-negligible portion of the bot networks are teenagers and college students learning how to manipulate the system.
Inexperienced operators like that make amateur mistakes like leaving in default user-agent strings for a variety of tools that are frequently linux-based as its cheap to go on something like LowEndTalk.com, pick up a couple $2/month VMs, pick up some proxies, and test out a small group of bots designed to fake traffic to generate Ad revenue.
The reason I said teenagers & students is you have to be really, really sophisticated for that sort of operation to scale beyond $100/month in profit so its generally something I've seen people "learn" on, realize the limitations, and move on. (i.e. Making no amateur mistakes that flag your traffic as non-legit and being able to use an IP address only a couple of times a day in random patterns)
Re:Google and Facebook ad traffic is 90% useless
by downandout on Sunday February 26, 2017 @01:29PM
(#13738766)
I have had similar experiences, but even if the traffic is 90℅ useless, that means 10℅ isn't. It's an auction, so bid an amount where clicks from each particular network are profitable to you. There are some niches where brand advertisers and well-funded startups willing to lose money will suck up all the inventory unprofitably. You simply won't be able to compete with that type of activity unless you also have money that you are willing to light on fire.
However, there are pockets of profitability within the PPC arena. They shift quite a bit at various times based on the competitive landscape...an unprofitable niche today may be a profitable one two months from now after someone that was driving up bids burns through all their cash. You just have to monitor and deploy when the CPC is low enough for you in your niche.
Re:Google and Facebook ad traffic is 90% useless
by cm2012 on Sunday February 26, 2017 @01:25PM
(#13738735)
Either your ads are effective at driving more sales than the ads cost or they aren't. It doesn't matter what % of the traffic is "fake" as long as the bottom line works.
Re:Google and Facebook ad traffic is 90% useless
by edblarney on Sunday February 26, 2017 @01:04PM
(#13738601)
It's useful to some things, esp. when there is a high value potential click conversion, i.e. amazon and ebay items.
Some ads are bought by brand advertiser with little understanding of the impact because a) it's 'brand' and hard to measure in the first place, and b) they have a huge budget, need to spread it across channels, so they just 'buy X$' of 'social ads' on the 'trusted networks' because that's the marketing mix required of their bosses or client. To be fair - print and TV ads have the same problem. Marketing is not a hard science in many places, and they have big budgets :).
But it clearly works for some folks.
There should be no surprise at in-links from reddit: theres usually a degree of clear interest in the first place, and second, there is a 'voting' structure that naturally moves better quality stuff to the top of the pile.
Re:Google and Facebook ad traffic is 90% useless
by xiphias on Sunday February 26, 2017 @01:04PM
(#13738603)
If the advertiser used the same bids on different properties, the campaign wasn't yet set up well. Also the device (mobile/desktop/tablet), query, language and many other features of the user session change the conversion rate, which means the bid should be different.
The fact that the ads were useless meant that either there is too much auction pressure and the advertiser wasn't able to compete, or he just didn't set up the campaign well (which is of course a hard task).
It would be good to know if the advertiser was using Enhanced CPC or not, which can increase the effectiveness of the ads, but even then it takes money to set up the bids correctly.
Re:Google and Facebook ad traffic is 90% useless
by synicalx on Sunday February 26, 2017 @06:40PM
(#13740716)
When you think about it, "real" advertising is the same - think how many people drive past a billboard vs how many of them actually buy whatever is on the billboard.
Google/Facebook really shoot themselves in the foot by claiming that they only target your ads at people who are already interested in what you're selling. That makes their ad services sound like some sort of marketing panacea and sets the expectation that you'll get a lot more interest than you actually will. What they don't want to acknowledge is that people are just as desensitised to, and disinterested in online ads as they are in "real" ads.
Re:Google and Facebook ad traffic is 90% useless
by benologist on Sunday February 26, 2017 @01:50PM
(#13738936)
Gleaning value from their advertising will become the dialup of this generation as entrenched manipulation and adblocking grow concurrently and they can't thwart either group.
AOL still milks dialup, but the debate has long-shifted to how they could possibly be getting away with it. That's Google/FB style advertising in another decade when 1 in 1000s of clicks is an actual person you want to pay to reach but they continue to use contrived metrics to justify billing for 100s.
Re:Google and Facebook ad traffic is 90% useless
by guyb7 on Sunday February 26, 2017 @01:56PM
(#13738986)
Had the same experience with Google and Facebook ads. ~80% of the clicks are simple bots that don't run javascript or store cookies.
Additional ~10% are more sophisticated bots that try to mimic real people behavior, but when you compile their actions to a video it's really easy to see that they are not human.
Most of these bots are run by agencies with the purpose of clicking on their clients' ads.
Re:Google and Facebook ad traffic is 90% useless
by uladzislau on Sunday February 26, 2017 @01:42PM
(#13738867)
It really depends on the product and the industry as well as many other factors like proper targeting, ad copy, matching expectations of the user, etc.
I've seen many ad campaigns with fantastical results or at least very positive ROI and many spectacular failures too.
Re:Google and Facebook ad traffic is 90% useless
by pacomerh on Sunday February 26, 2017 @01:49PM
(#13738926)
I tried it once for a month and I got results, probably more than 10% (To sell a music album). I was super specific with the demographic, I guess that helps.
There are decent sites too of course, but you cab target then manually I believe. Arstech for an example.
It would be useful for a few sites to publish IP addresses and times from which they get bot-like traffic. Maybe a pattern will emerge.
Inexperienced operators like that make amateur mistakes like leaving in default user-agent strings for a variety of tools that are frequently linux-based as its cheap to go on something like LowEndTalk.com, pick up a couple $2/month VMs, pick up some proxies, and test out a small group of bots designed to fake traffic to generate Ad revenue.
The reason I said teenagers & students is you have to be really, really sophisticated for that sort of operation to scale beyond $100/month in profit so its generally something I've seen people "learn" on, realize the limitations, and move on. (i.e. Making no amateur mistakes that flag your traffic as non-legit and being able to use an IP address only a couple of times a day in random patterns)
However, there are pockets of profitability within the PPC arena. They shift quite a bit at various times based on the competitive landscape...an unprofitable niche today may be a profitable one two months from now after someone that was driving up bids burns through all their cash. You just have to monitor and deploy when the CPC is low enough for you in your niche.
Some ads are bought by brand advertiser with little understanding of the impact because a) it's 'brand' and hard to measure in the first place, and b) they have a huge budget, need to spread it across channels, so they just 'buy X$' of 'social ads' on the 'trusted networks' because that's the marketing mix required of their bosses or client. To be fair - print and TV ads have the same problem. Marketing is not a hard science in many places, and they have big budgets :).
But it clearly works for some folks.
There should be no surprise at in-links from reddit: theres usually a degree of clear interest in the first place, and second, there is a 'voting' structure that naturally moves better quality stuff to the top of the pile.
It would be good to know if the advertiser was using Enhanced CPC or not, which can increase the effectiveness of the ads, but even then it takes money to set up the bids correctly.
Disclaimer: I'm working on a relevant product
Google/Facebook really shoot themselves in the foot by claiming that they only target your ads at people who are already interested in what you're selling. That makes their ad services sound like some sort of marketing panacea and sets the expectation that you'll get a lot more interest than you actually will. What they don't want to acknowledge is that people are just as desensitised to, and disinterested in online ads as they are in "real" ads.
AOL still milks dialup, but the debate has long-shifted to how they could possibly be getting away with it. That's Google/FB style advertising in another decade when 1 in 1000s of clicks is an actual person you want to pay to reach but they continue to use contrived metrics to justify billing for 100s.
http://www.cnbc.com/2017/02/10/facebook-to-audit-advertising...
Additional ~10% are more sophisticated bots that try to mimic real people behavior, but when you compile their actions to a video it's really easy to see that they are not human.
Most of these bots are run by agencies with the purpose of clicking on their clients' ads.
I've seen many ad campaigns with fantastical results or at least very positive ROI and many spectacular failures too.
And 10% useful doesn't sound too bad....