* Is anyone aware of evidence of the extent of click fraud on Facebook? I see a lot of discussion about it but not a lot of data.
I personally think there's a reckoning coming soon to traditional ad tech. There is so much fraud, such ludicrous and impossible revenue and impression targets, so much pushback against shitty and annoying advertising, and so little incentive to be honest with the customer. It can't last like this.
I'm on both sides of the coin.
As an advertiser, click fraud is obvious when I see a huge disparity between google analytics and the ad network. You'd have to put a lot of effort into sending fake traffic that actually engages with the advertiser's landing page.
As a publisher, I know that if I send a bunch of BS traffic to my advertisers they will blacklist me and starve my website.
Regarding your estimates, I agree for really broad traffic.
If your targeting options are very vague/loose you can buy ad traffic for very cheap, but you end up with a large number of click fraud.
Most advertisers should want to be as narrow as possible with their targeting. The tricky thing is that when everyone is bidding on the golden goose egg demographics the price jumps quickly.
In some cases, you're better off buying the 'general audience' traffic for pennies just because of sheer volume multiplied by chance.
Done properly, the customer doesn't actually care if 99% of the exchange inventory and impressions are fraudulent, as long as they accomplish whatever objectives they set out to.
Put more explicitly, if the objective for an advertiser (or, more likely, their agency) is to build awareness of a brand, they can measure a measure awareness among their target group, run a few million dollars of advertising, and then measure awareness again to see the difference. Done with enough customers, and enough programs, you start to get a very clear sense of where the dollar is best spent.
When you're advertising on multiple channels, even if you did have a solid measurement of awareness (you don't), you still wouldn't know what was responsible.
This is the appeal behind things like Facebook ads. It seems you have a very clear sense of what return you're getting because there's such precise targeting and tracking. The issue is that neither of those have anything to do with the first portion of your comment: measuring awareness. You still need to measure awareness objectively and no one knows how to do it.
The people best equipped and most ready to act like they do, however, is — you guessed it — Facebook and similar ad networks!
Besides, the ad industry has been kept afloat with exactly the type of behavior that grandparent describes for decades if not centuries. What makes you think they're more sophisticated now?
And you run many, many programs on lots of channels in different demographics, and geographies - and see which program is most effective.
Also, I'm not saying it's all fraudulent and doesn't work at all - of course if that were the case everyone would know. What I'm saying is that the brand is massively overpaying, because a large percentage of their delivery goes to fraudulent bot traffic. And it's pretty much in everyone's financial interest except the brand's to turn a blind eye.
Also, it's important to know how campaigns are actually delivered: firstly, Sales sells something wildly optimistic like delivering 1m impressions a day to 45-50 y/o women in Florida. As time goes on and delivery can't be met we loosen the targeting parameters in order to "deliver in full" even though this delivery is no longer to the customers specifications. We might tell them we're doing that, or we might not and hope we get away with it. Then after the campaign we take whatever numbers we got and create a story out of it that makes us look good. If we're working with an agency we give it to them and they presumably spin it some more before presenting it to the customer.
Why do you do this work? You clearly know what you're doing; perhaps you're very good. And I won't presume to know your reasons or the underlying belief system that may or may not support it.
But from my personal perspective, it sounds like hell to spend my productive days essentially just selling stuff. Is it that you believe in what you're selling? Is it that in your moral framework being good at this job trumps what seems like a bad way to spend your time?
I'm truly asking out of curiosity, by the way. I'm a front-end developer and a significant portion of my work, when I work, is essentially dressing up ads with websites that don't add up to much of value (within the parameters of my own perspective on life). So judging you would be hypocrisy.
But I'm actively unhappy about it, and any excitement I show involves the craft, and I'm looking into ways to apply my craft in ways meaningful to me. And I'm curious to know how this works for you.
I am not against advertising per-se. I think there is some truth to the idea that advertising keeps the Internet free. It's also a part of the sales cycle that drives the economy, and keeps a lot of people in work. I think the fear of tracking is a little overblown, too - everything I've seen about tracking in my company is extremely focused on keeping user data anonymous (at least as much due to fear of a lawsuit and bad PR as out of ethical standards). I've never heard anyone suggest doing anything with that data other than trying to use it to deliver better results for our clients. We have massive amounts of tracking data, but we don't know who anyone is or where they live. We seem to try hard to keep it that way. Now, it may be possible for someone with sufficient access to take the data and figure out names and addresses, but we never even let that data into the system, and we have stringent, externally audited controls, so it would have to be a malicious act. You'd be out of a job in a heartbeat if you got caught doing that.
So, I don't have a moral problem with the way my company in particular is using data. I think the fact that your mobile phone operator will sell your location is much more concerning. There are a lot companies in the ad tech space, and I'd bet a lot of them care a lot less than we do about personal user data.
On the issue of fraud in particular I can truthfully say that I've only just become aware of it. I am an engineer on a small component of the platform. Communication between departments is not great, so I have been disconnected from the reality of how campaigns are run. It wasn't until I looked into our data while diagnosing another issue, and subsequently reached out to account management, sales etc that I realized what was going on. That was in the last few weeks. Since then the issue has blown up and we're working on a plan to deal with it. Whether that plan has teeth and is enacted will tell a lot about the willingness of the company to do the right thing for our customers. I'm cautiously optimistic, but that optimism is tempered by the general negative attitude toward transparency that I've seen from leadership in the company.
I also am planning to leave the company before the end of the year. This is as much for personal reasons as work ones. The company is dysfunctional and oftentimes its own worst enemy. I wouldn't work for a company like this again, and would ideally stay out of ad tech entirely.
So even if lots of clicks were fraudulent, they're looking at ($ earned from clicks)/($ spent on clicks). Based on Facebook's continuing increase in earnings, it's safe to bet that lots of major advertisers are seeing this ratio returning >1, and are therefore increasing their spend.
All advertising platforms deliver fraudulent/worthless/imaginary/bogus views. Print has a problem with publishers making up circulation numbers. Television has a problem with ads running while no one is watching. Radio has a massive problem as their listener numbers have collapsed. The display ad exchanges have tremendous problems. Mobile apps have an incredible amount of accidental clicks which may often either be 10 year olds trying to play a game or just outright fake. Video ad networks have a problem with ads auto-playing on repeat that are seen and heard by no one.
Google and Facebook are about the easiest to deal with, by far.
Hopefully someday all ad networks decide that they only get fraudulent clicks from me, and stop their practices.
I think their is a certain amount of Black Market SEO Google/Facebook/Whoever will write off as an acceptable loss because the technology just isn't their to stop it (or at least the financial incentive is not there).
Basic heuristics however seem to be able to keep the big offenders in check, and that's all that really matters it seems.
This probably counts more as an anecdote than as data, but at my last startup we bought about $100,000 of Facebook advertising and never saw click fraud (we tracked downstream behavior and revenue very carefully). Far and away Facebook provided the best traffic compared to any other paid channel.
source: https://www.youtube.com/watch?v=oVfHeWTKjag
click fraud less so. i've run marketing campaigns and you do see some clicks missing in google analytics, but it's not as bad as the like fraud.
There are multiple articles about the quality of the demographic you get when you use their official platform. Conspiracy theorists point out that likes are a very weak signal of interest, and thus it's hard to judge if you got good likes or bad likes.
Facebook has a conflict of interest in choosing who to show your like ad to: if both Alice and Eve are willing to like you with equal probability, but Eve is less likely to be an engaged fan, Facebook should show your ad to Eve.
You got your like, so you're happy, but your content is less likely to organically go viral, because Eve isn't as engaged as Alice would have been. Notably, not all of your fans will see your posts: Facebook no longer guarantees delivery to people who have liked you. Luckily, Facebook will help you run a campaign to target the fans of your page that you just paid to get.