Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

“35-50% of clicks on Reddit Ads are fraudulent”

reddit.com

Post reply on HN
Print view View original

321–330 of 398 posts

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#321

Post by cbsmith »

stealthbot wrote: post #298
cbsmith wrote: post #214

I gotta say... this is an incredibly naive analysis. To demonstrate Reddit Ads isn't doing click fraud protection, you should try to generate a fraudulent click yourself. Looking at what is getting through and finding that a lot of it as easily identifiable as fraudulent is NOT a valid signal. It's incredibly likely that there are far more bots that are being blocked and that the bots that are getting through have in…

A company basically did what you described on Facebook, Google, Bing and Yahoo. https://www.methodmi.com/reports/in-plain-sight

A lot of companies have done this and do this kind of thing as a service.

That said, this report's objective isn't really to measure click fraud detection on those networks. It's to highlight that one doesn't have to be very sophisticated to commit ad fraud (which is absolutely true). It also has some really odd ideas (e.g. they contacted the Chrome team to let them know about the existence of the stealth plugin in, and were surprised that the Chrome team said that the browser was working as intended so there was no action for them to take on their part... they similarly drew a false equivalence between defending account creation vs. defending click fraud). They also seemed to think that sites should be blocking bots, which is of course not what you want to do, because that provides a feedback loop to the fraudsters so that they can figure out how to overcome the measures. You want to proceed as if it is working, even let it show up in the ad campaign data at least initially, and as much as possible make corrections in ways that make it difficult to determine which traffic has been identified as fraud.

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#323

Post by XCSme »

stevesearer wrote: post #166
Pick-A-Hill2019 wrote: post #160

Earlier quoted context omitted.

When you say 'on my site' would it be correct to say that you are talking about the one in your 'about' section? If so I was motivated to take a look as a way of expressing support for your design choices. However, you are using a Google-Analytics script???? Nothing wrong with that Buuuuttt.... uhm, yeah.

Yeah, I would like to not use GA too in the ideal world too :/ That said, we do use the Anonymize IP setting and have all of the demographic / interest settings turned off for Google Analytics. We basically whittled 3rd party stuff down to GA and a Google font, so while not 100% perfect, I am pretty happy with the strides we have made.

I almost managed to remove all 3rd parties too (apart from the tawk.to chat widget, until I implement a custom contact form).

If you are looking to replace Google Analytics with a self-hosted solution (to fully eliminate 3rd parties) you can check out the analytics platform I'm building: https://www.usertrack.net

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#324

Post by thatguy0900 »

GekkePrutser wrote: post #293
thatguy0900 wrote: post #142

Earlier quoted context omitted.

As others have said, any service doing fraud prevention might ban you. It doesn't seem to be happening yet, but that could change any time it got popular and they would be pretty justified doing it.

Since when is clicking on something but not looking at it fraud?

I know you're being obtuse, but the fraud is that someone is paying for you to look at those ads. You're under no obligation to actually do it, of course, and the site is under no obligation to let you use it, hence the ban.

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#325

Post by stevesearer »

XCSme wrote: post #323
stevesearer wrote: post #166

Earlier quoted context omitted.

Yeah, I would like to not use GA too in the ideal world too :/ That said, we do use the Anonymize IP setting and have all of the demographic / interest settings turned off for Google Analytics. We basically whittled 3rd party stuff down to GA and a Google font, so while not 100% perfect, I am pretty happy with the strides we have made.

I almost managed to remove all 3rd parties too (apart from the tawk.to chat widget, until I implement a custom contact form). If you are looking to replace Google Analytics with a self-hosted solution (to fully eliminate 3rd parties) you can check out the analytics platform I'm building: https://www.usertrack.net

I actually did try UserTrack a couple of months ago using the self-hosting option and it was easy to get going, though was missing a few metrics we use in addition to event tracking.

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#326

Post by stevespang »

My son is a software engineer by day, a pro-am Texas Hold 'em Poker player on some nights at "legal" card houses. Some criticize his "outlaw" endeavors, but he has over a year of serious winning statistics.

Compared to online ad scams - - his pursuit is a solid cake walk. No. of clicks as a metric should be outlawed.

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#327

Post by MaxBarraclough »

rightbyte wrote: post #264
hinkley wrote: post #98

Earlier quoted context omitted.

I'm constantly amazed at the frequency with which articles posted to Hacker News have a bar of clickbait ads immediately following the body of the article. Those rabbit-hole articles where they split eight facts across 12 pages and in some cases never mention the fact they were advertising in the original link. All so they can put six more ads per click in front of you.

Ye. These "You can't guess how 'random actor' looks now" and then a picture of some drug addict and not the actor. The actor in the ad is nowhere to be seen if you click the ad. I actually got fooled to click because I couldn't believe that the actor got like that and it looked like an article on the "honorable" newssite ...

What's the endgame for those ads? How is your click meant to convert to revenue? Plain old scams?

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#328

Post by a1369209993 »

zokier wrote: post #73

I find it so weird idea that people really click on ads. I don't know if I've ever done so (intentionally). As such I would have expected more like 90% of clicks to be fraudulent.

Yeah, 35-50% is just appallingly low; does anyone have ideas on how to improve this number?

Advertising delenda est.

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#329

Post by throwaway20382 »

earthboundkid wrote: post #8

Every newspaper in America has obvious scam ads on its homepage. The whole online ad market is diseased and needs to be uprooted. Common sense fixes: - Never allow dynamic ad content! Ads should be individually approved by publishers and not subject to change after approval. - Ban user tracking and targeting. It's a red queen's race and no one benefits. Everyone is better off once it's illegal. - Pay for ad time slot…

"- Pay for ad time slots, not user impressions. Instead of selling by CPM, publishers should say "independent analyst X says we have traffic Y on an average Sunday. You can buy placement for next Sunday for $Z." "

This would end fraud. This would end mis-clicks. This would benefit the purchaser of ads AND would make the ad market more efficient. So, asking for "world peace" might be a easier route to go.

Re: “35-50% of clicks on Reddit Ads are fraudulent”

#330

Post by justinbaker84 »

Advertising manager here - The real problem is that not enough companies pay attention to the results they are getting from their ad campaigns. Most big companies just dump money into a lot of different campaigns without putting in place strict measurement procedures. So this puts the ethical ad seller at a major disadvantage. Currently you have companies spending tens of millions of dollars on digital ads each quarter and only looking ad metrics like cost per click or cost per impression. So they are judging the effectiveness of the ad campaign on metrics the ad seller provides. They are not putting in place effective checks and balances.

Why you ask? Because if the huge companies buying ads wanted to actually measure everything it would be extremely time consuming and expensive. It is a lot easier to pretend the problem doesn’t exist and just go with the flow.

So small companies with small budgets are actually a lot better at measuring results in most cases. Also online only businesses are a lot better at it. Measuring the offline impact of online ads is very hard.

But the bulk of ad dollars are spent by huge companies with huge bureaucracies. So game theory says you need to capture those dollars to win.

Source – I’ve been doing digital marketing for the last 12 years at a variety of companies.

Post reply on HN