Facebook reported fake numbers to advertisers

It took time, and many thousands of dollars, before I realized that the vast majority of “likes” my pages received as a result of paid campaigns on FB were from accounts which were clearly not real people.

A simple look enough of their profiles revealed that, like would he expected from any fly by night CPA network, FB was using bots, or at least straw man accounts run by low-cost staff, to like and view content which FB was paid to advertise.

Worse, I found that the clickthrough metrics reported by them to off-FB destinations I advertised NEVER was anywhere close to what was reported on the destination, including when tracked by Google Analytics.

In short: like-fraud, click-fraud, and more.

I cannot be the only person to notice these things. I assume it persists because most people, self included, simply complain and move on once we notice the “game” but don’t sue.

You're not the only person. Here's Veritasium reporting on it in 2014: https://www.youtube.com/watch?v=oVfHeWTKjag&ab_channel=Verit...

> but don’t sue

Ok, so, honest question, why not? If you're correct then it seems like a class action would be a slam dunk and wouldn't cost you anything personally because class action lawyers are happy to skim millions off the top of the settlement.

100%

Fraud in advertising isn't just a Facebook issue. The entire digital advertising ecosystem is chock full of fraud. Of course, ad networks aren't incentivised to do much about it, because they get their cut even when ad impressions are fraudulent.

I spoke with Augustine Fou late last year for my podcast about digital fraud, it was pretty eye-opening to say the least.

https://www.mql.fm/002-60-million-60-billion-ad-fraud-questi...

Unlikely that these accounts were run by FB. More likely by other companies with their own objectives, such as creating plausible fake users.

In theory.

In practice FB's lawyers will produce a compelling bullshit excuse to explain the poor results, the fake clicks, the fake accounts, and the lack of sales.

Likely it will be the user's fault and FB can't be held responsible if the user keeps spending money etc etc.

This is really an antitrust issue, because it's a particularly nasty form of market monopolisation combined with cultivated mental and emotional manipulation to keep buyers locked in and spending.

Click-bots abuse everyone; a feasible strategy is running ads on random news re-aggregator sites or other worthless sites and buying the lowest cost ads on FB pointing to them. The click-bots game the engagement/quality metrics to make the ads seem legitimate enough to keep serving, and the click-bots have to click on a wide selection of ads (including yours) to avoid drawing suspicion. Normal traffic mixed with more click-bots gets the revenue from the worst scammy advertisers (who may also be operating the sites and injecting their own scammy ads over legitimate ads they promised to serve)

It's basically arbitrage to sell spammy clickbait ads for scammers at higher prices than FB would pay if they would even allow the ads on their network.

Advertising networks try to detect and filter click-bots but of course some percentage will slip through.

Switch to ad networks where you pay for conversions, not clicks/impressions.

> such as creating plausible fake users

Define a 'plausible' user? Facebook just wants your phone number and uses that as the only metric to determine it's a human behind the account. And as we know, there is no shortage of phone numbers to use from places like Twilio where you can mass-generate an army of Facebook users.

If it's anything like the situation in Europe you can't effectively sue Facebook without at least a couple millions to potentially throw out the window. They pay a lot of good lawyers to make the court proceedings as long-winded and expensive as possible for the opposing party, relying on the fact that very few have the resources, time and determination to put up with this.

Oh yeah organizing a class action against facebook, a "slam dunk" :D

When I ran a campaign I experienced similar problems. Received likes from very dubious looking Facebook profiles who likes all sorts of random stuff.

Page Like ads are the least ROI effective ads on Facebook. You tell their machine learning algorithm to optimize for people who will click "Like" on every page they see. That'll increase your vanity metric and meet your stated objective, but it doesn't drive business results for you because you're getting the wrong type of customer.

If you tell Facebook's machine learning algorithm to optimize towards purchases on your website or visits to your stores or to users onboarding to your app, then you'll really see the power of their beast. Those are the ads that people are spending billions of dollars on, because the outcomes drive real business value and have too much friction to be faked at scale.

Do they even have a right to sue or does FB force them into arbitration?

I honestly don't know as I have never bought ads on FB. It would not surprise me if all but the biggest enterprise customers have horrible terms of service...

The fraud Facebook habitually engages in is one thing. The psychological effects of their apps on their users is another. Machine learning gears the content toward what spikes emotion in the moment. Social media is the new cigarettes.

The thing that confuses me about the Facebook like, follow, or click fraud is: it seems like Facebook is stringent with their real name + real person policy and bans people that violate it. So how are all of these bots surviving? And if somehow they are real people's accounts: how are they getting outside the geo fencing / geo targeting?