Hacker News Reader

Comment by Mistri | Hacker News Reader

I have absolutely no faith that Google will be able to fix this. I run a small site that gets ~5K views per day, and a few weeks ago we had a heavy traffic day (~15K views in one day) due to a retweet on Twitter. Google marked it as fraud, and my account was "temporarily" banned. The ban never got lifted, so I did some research and finally got it resolved (if anyone's curious, I made a video detailing how: https://www.youtube.com/watch?v=_pCoBoK1hEM)

If Google marked organic traffic as fraudulent, their algorithm has a long way to go before they can start distinguishing between sabotage and actually fraudulent traffic.

View on HN

Back when I was in college google stole the ad money my website had generated. They let me run ads for months, but when I went to check out they accused me of click fraud. No recourse, no customer service, and absolutely no respect from Google about this.

Even now, more than a decade later, I refuse to use many Google services (such as their cloud) because I know they built part of their business by stealing from people like me.

Did you consider seeking recourse in court at the time? I wonder if a wave of lawsuits or even small court claims would tip the incentives for Google to create a proper system for contesting.

I wonder if you could file a complaint today in small claims. Do you have any records from the time?

I've often wondered what would happen if someone were to try that. I suspect it would result in Google hell banning you from all of their services in perpetuity, just like they do to people who attempt to dispute a credit card charge.

Just last night I listened to something on NPR about CAPTCHA v3, which doesn't involve checkboxes nor pointing at parking meters anymore. It samples page activity and interactions by the user and supplies a human-likelihood score to the site (h/t Twitter bot scorers).

So we live in a future where our activity is bounded by a variety of quality scores. You're mostly a human, your site's increased traffic is less legitimate...every Taylorist-Goodhartian aspect of existence is priced into the marketplaces of society.

Perhaps a decent analogy could be Black Mirror's "Nosedive" episode, but with endless dimensions instead of just one (which admittedly could be a derived score from the dimensional ratings), and ratings that aren't derived from the moods of peers and other interactors, but passively gathered from everyday behavior.

I think this is the NPR show episode: https://www.npr.org/sections/money/2019/04/24/716854013/epis...

I like the way Qwant search engine deals with this. It gives a small challenge to solve which is a lot less cumbersome than finding traffic lights on photos or far less invasive than the fingerprinting technologies others use.

You haven't actually said anything about the challenge, so we have no idea if it is actually any different.

I wish there was a challenge that helped verify OpenStreetMap edits!

> Just last night I listened to something on NPR about CAPTCHA v3, which doesn't involve checkboxes nor pointing at parking meters anymore. It samples page activity and interactions by the user and supplies a human-likelihood score to the site (h/t Twitter bot scorers).

That turns the web into a walled garden you can't access without submitting to google's tracking. If they'd just drop the pretense of not being a walled garden this wouldn't even be a problem that needed solving.

So you train a bot to show 'human like patterns'.

The problem with relying on ML on one side is it can be gamed by ML on the other.

If it was easy to train a bot to show humanlike patterns, we'd have driverless cars by now.

Hmm... I think training a program to fool another program is a much, much, much, much easier problem than driving a car.

As an example, you only have to learn how to emulate a single 'human like' pattern to fool the ML, where you target function is a simple fool or not fool ML - once you achieved fooled you are done.

Whereas to drive a car you need to deal with all the infinite variation in inputs you might experience in driving a car, with complex target functions ie safety versus speed.

Googles classification algorithms are shit. They mark spam as ham, ham as spam, block good accounts and promote fraudulent ones. It's ridiculous the amount of trust we place in this company.

That's on point, I just got my Google Ads account banned for no reason and support is non-existent of course. Google just doesn't know how to deal with fraud at all.

There's a criminal fiction book to be written here. Disgruntled Google employee ends up playing both sides, as the man supplying the classification algorithms and also selling 'backdoors' to it. Only a rogue loose cannon dev ops can stop him.

It's not really trust. It's really that there is no alternative, and the laws are such that they have no accountability.

Most google products have alternatives. The ones with a network effect (where you need Google servers even if you have an open source client) are Youtube and the Play store (I can't get most content from either in any other way, most companies don't give apks for download), but the rest is replaceable or has open source variants (ungoogled chromium and AOSP for example).

I do agree on the accountability part. When I can't take part in normal life without a certain product, it governs people's lives and they should be held to similar standards as other governing bodies.

Agreed. I hear their Ad Words is tough to beat for those who make money that way.

I don't get why they would ban your account, and not the fraudulent clickers. They are the problem. Surely Google is able to detect where these clicks are coming from, and can tell that it's the same as other fraudulent attacks?

15K is not so much honestly... I would have expected this to happen on much higher numbers.

Thanks for the video btw.

It doesn't really happen to big companies because they have the power to sue.

Every individual I know who's run google ads has been banned right before a payout at some point. They generally all in the range of $3000-6000. It's an amount that's a decent sum, but only to people who don't have deep enough pockets to afford a lawyer to go against a megacorp.

I'm not convinced it's not intentional. The banned people probably aren't generating that much in profit for google, and seizing a few thousand dollars from a customer who's likely to just cancel services at some point probably has a better payoff after a certain threshold. No google employee will ever explain how it works and there's zero support, so I doubt my mind will ever be changed on this subject. One thing I do know is Google isn't a stupid company--just evil. The problem has been too consistently occurring and too consistently ignored that it can't just be incompetence. Say it's spam and show a couple suspicious figures if anyone really tries to investigate and you get off clean.

Knowing Google they probably are using ML to maximize revenue, and the black box is causing the system to fuck little customers over. Oh well, it isn't fraud, they're just doing what the algorithm tells them to.

Reply on news.ycombinator.com