Reading this thread I really start to feel people don't understand what fraud is. You shouldn't be calling it fraud when in the next paragraph you say that FB does work for impulse buys and not "ozonetel"

Just because your ads didn't convert, or 90% of the traffic you buy didn't convert, doesn't make it fraud. Fraud is malice, with the intention to defraud someone and not deliver what you promise, with intent. This is completely different to you setting a target demographic on FB and Google and it not converting.

I imagine most people calling fraud in this thread aren't actually talking about fraud, just non converting traffic. There could be a million reasons for this. You haven't set up conversion tracking, your targeting could be off, or your product or landing page is just crap. Calling it fraud is not helpful.

I think you misunderstood what he wrote.

My business also sell only to other business, and Google kept calling us nagging us to do some "experiments".

One of them involved enabling "display network" ads on mobile.

Like the guy side... it "attracts" fraud.

So, what kind of fraud?

I am not talking about no conversion or no click, I am talking about all our clicks, or almost all to the point it was impossible to tell the difference, were fraudulent, for example most our ad spending was going into a musical keyboard app for kids, that has nothing to do with my business, despite me blocking the category multiple times, the app in question just kept being reuploaded in different categories to avoid bans.

When I went to check out the app Google Play page, it was rated 1 star, and lots of complaints similar to this:

"App is just a fraud, it keeps opening ads in the middle of the song while my toddler is playing to force my toddler to click the ads, then it automatically closes the page that opened and keep running, using all my battery and making my toddler see weird stuff."

So yeah, it is fraud, it is not "failed conversion", it is an app conning toddlers into clicking my ads that are selling stuff that shouldn't be anywhere near a toddler, both the parents and my business were being scammed.

You've misunderstood what I wrote. When I'm saying that "90% of what we got what was fraud", I mean that 90% of the traffic that did convert from FB ads were users that were trying to scam us (it was a financial services app), I mean literally steal money from us with things like stolen credit cards.

The general user quality you get from Facebook Ads is atrocious.

> You shouldn't be calling it fraud when in the next paragraph you say that FB does work for impulse buys and not "ozonetel"

I don’t believe they were saying the adds themselves were fraud, merely that they had a higher than average propensity to attract fraudulent customers.

Here are the California jury instructions for fraud and deceit: https://www.justia.com/trials-litigation/docs/caci/1900/

There are definitely multiple categories this would fall under.

Facebook ads are the worst from a users perspective. I bought a floor mat through them once and for the next week all the ads were for floor mats. I just bought one I’m not interested in buying another. I bought a PC chair recently with my credit card, now I see a bunch of ads for PC chairs, again, I don’t need another one.

Same thing with kitchen appliances, but not limited to Facebook. I'm not going to buy another refrigerator for at least 10 years, but still I'm treated as Mr Fridge.

Honestly this might be a good thing, it seems to show that ad networks don't have or at least don't provide your actual purchase history to the ad space buyers because given that info they probably wouldn't be spamming someone who just bought a $thing with ads for $thing_but_from_us. As annoying as it is it's probably better than the alternative of ubiquitous access to your purchase history.

Well we know these places buy your credit card history, so it’s more proof that their ad system is shitty and their customers should make a fuss.

Yup, this is the problem. There is no centralised log of what people bought, which is what you'd need to actually fix this problem.

A good argument, but it happens on Amazon when the search, research, and purchase were all on Amazon.

Yup, I get "buy it again" prompts on nearly all of my recent purchases, even when they are all durable goods that don't expire or require restocking, like say a tablet case.

Hmmm, that's really weird. It's probably an edge case that would be complicated to change in that case (never underestimate the power of inertia, especially in ML systems).

I feel like it might be almost trivially solved with a “buyer only needs one [for X months/years?]” property on the items.

It would likely have to be added by a human, but still: there are a relatively small number of items that people only need one of.

ML is pretty shitty at causal inference

Amazon enjoy that kind of thing too. Look at a product page, Amazon ads become for that product or similar ones. Buy it, and... they still want me to buy another one. And they wonder why people use ad blockers...

I've seen the theory floated that people return so much stuff that shortly after buying a thing you are indeed most likely to buy another, slightly different one of the same kind, and that explains the weird categories where it happens. (and for lots of things Amazon sells recommending similar items does make some sense, although I would expect a bigger spacing in time sometimes.

Also that brand satisfaction is related to how it performs relative to other offerings. If you just dropped $44k on a new vehicle, they want to make sure that you're seeing fancy cars, thinking about how nice and luxurious they are, etc.

And the emails! I keep getting suggestions I just brought a copy of a book why on earth might I want to buy another copy the next week.

One way to avoid/reduce this is to always place orders from a private browsing/incognito window, and to never log in using your normal browsing sessions.

Well no one has a plausible explanation on it yet but with some of the highest A/B testing ratio and best Data Scientist Amazon too seems to think this works.

And yes I agree it sucks.

Hah. I remember shopping for a wireless router and seeing things like “shoppers who bought this item also bought...”. Yes, we all buy three slightly different models of routers. Good job Amazon.

Actually there is. There is a certain chance that you're unhappy with your fridge and you'll return it. So the assumption is that you are more likely to buy a different fridge than any other item. That's why they show you other fridges.

I have an established personal opinion on this, and the opinion is that I'm glad that things are this way. It means that ads are pretty dumb and relatively harmless so far and I can easily ignore them. Floor mats? Tough luck bud, I just bought one.

Now, when the ads begin predicting what I would want the next day—that's when I start worrying.

Don't use facebook but surprised how my spouse finds instagram ads super relevant and interesting enough to check them out almost 100% of the time. They are mostly fashion/travel related that she is interested in checking out,if not buying.

I work for a retail app and can usually tell when we have or haven't been spending on FB ads when I check our usage numbers for the week.

They do work but like anything it's about choosing the right tool (platform) for the job. We wouldn't run ads on LinkedIn and B2B SAAS companies shouldn't advertise on FB/Instagram.