Hacker News Reader

Comment by kiernanmcgowan | Hacker News Reader

This ad block debate makes me think about patio11’s post “The optimal amount of fraud is non zero”. Essentially in trying to make a system as fraud-free as possible you make it too difficult to work with / the cost of fighting fraud is higher than the cost of the fraud.

Ads vs Ad Blockers seem to be in that space, but on the cost of customer good will. Can YouTube squeeze more money from ads by doing this? Sure. Is it worth the customer anger? Maybe?

https://www.bitsaboutmoney.com/archive/optimal-amount-of-fra...

View on HN

> patio11’s post “The optimal amount of fraud is non zero”.

Thanks, I just skimmed this post. It seems to me that "optimal" is substituting for "cost effective". The optimal amount of fraud is always zero, it's just not cost effective given the existence of people willing to go to lengths to commit fraud.

> Can YouTube squeeze more money from ads by doing this?

The third option is that this would incentivize a portion of the viewership to purchase Youtube Premium. Netflix seems to be benefiting with the same strategy.

It's not just cost effectiveness. It's about how awful you will make the user experience for real users if you tried to bring fraud truly to 0. It's essentially the security-convenience tradeoff. In terms of YouTube, it's the monetization-experience tradeoff.

Yes, because the corporations are optimizing for cost effectiveness (return on sales or return on investment).

Corporations don't care what the user experience is for its own sake. They care about keeping and retaining paying users. After a point they don't bother making the user experience better, because it isn't cost effective to do so.

But cost effectiveness makes it sound like you can fix it by throwing money at it. The point is that you cannot fix it, even with infinite money, because you will reach zero users before you reach 0 fraud.

The way to fix it with infinite money is to have individual human curators (concierge services) for every customer. This eliminates fraud and makes a very, very easy-to-use "user interface" when the human concierges are well trained.

Regardless, even if I'm wrong with "cost effective", the word still isn't "optimal".

However well trained they may be, humans are notoriously easy to manipulate and engineer around. Humans in a customer service role especially, since they are by definition expected to be helpful by default.

We've reached maximum reply depth.

> This is not a hypothetical. In many healthcare systems, that is the approximate timeframe for medical practice, doctors are amongst the most expensive and selectively educated service professionals on the planet, and can still be readily hoodwinked by a bad faith approach.

Yes, because they are trained, licensed, and incentivized to place other things above fraud detection or customer service itself. I was literally discussing with the other person about Youtube, for which 5 - 10 years of training is sufficient to train a concierge and still leave 4.5 - 9.5 years dedicated to fraud detection and prevention.

People complain about health care all of the time, yet most of them still come back to it. The threshold for flouncing due to bad user interface is a lot higher than for Youtube.

If they're paid enough you can select for humans with superior talents and, at least as important, training. Heck, with infinite money you could have them train for 5 or 10 years before ever having a real-case person to interact with.

I think part of the problem here is a mistaken interpretation on my part. I was thinking of "optimal" in terms of an ideal society in which no fraud takes place. I think the article writer was thinking of "optimal" in terms of optimizing the solution to a problem given the constraints and freedoms of current society.

The willingness and ability of people to commit fraud is not a constant. It's highly variable depending on social context. The more you allow fraud also changes society to accept more fraud, which is a vicious circle.

Cost effective in a way. A more precise thing to say might be that the Nash equilibrium of the system contains a non zero amount of fraud.

Or that fraud is an evolutionarily stable strategy within the environment.

> Or that fraud is an evolutionarily stable strategy within the environment.

I like this take on it.

I doubt YouTube is going for zero. I suspect they only took action recently because ad blockers have become very common.

I thought it was still something stupid low like 10% of people using adblockers?

More than 10% of telemetry-enabled Firefox users worldwide use adblockers. In Poland it's 15%. In the United States, Germany, and Russia it is closer to 20%, and in France it is closer to 30% [1].

[1]: https://data.firefox.com/dashboard/usage-behavior (see near the bottom of the page)

But Firefox is under 5% of browser market share (generous estimates still seem to put it under 5%—the low-side estimate is closer to 2.5%) and its users are almost-certainly far more likely than most to use ad blockers.

Not in Germany

You're arguing against a straw man.

I don’t follow.

“Not a lot of people use ad blockers”

“quite a few telemetry enabled Firefox users do”

“but Firefox users are a tiny share of total users and probably way more likely to ad block than most web users”

Where’s the straw man?

Do you see how you used the word “probably” to signify that the thing you’re claiming is just your opinion and not based on fact?

It’s a fact that it’s probable.

Do you want to pretend-argue with me, or to positively assert that Edge, Chrome, and Safari users ad block at a rate similar to Firefox users? This is silly.

Also, that’s not what a straw man is.

And the population that will block ads will include a lot that block telemetry as well on principle, so the percentage is higher than that.

But the population that blocks (opt-out) telemetry is likely low, at least that is what I would assume. So I don't think it makes a big impact

The ~20% using adblockers probably aren't a random cross section of society, but more tech-savvy audiences that advertisers will pay big $ to reach.

Most users have never even heard of Firefox. I just installed it on someone's phone yesterday, along with uBO; they had never heard of it before, but were very excited about being able to watch YouTube without ads. My social circle isn't very large, so me helping occasional non-technical users install FF+uBO on their phones isn't going to affect the numbers very much.

Even in the tech company I work in (not an adtech company), most people use Chrome, not Firefox.

Germany, mostly older users. We had a manual HEAD request to our google ads script for a while where we’d log the results (successful: yes/no). 30-35% blocked ads.

Statista has a pretty nice writeup available here. [1] Granted some of the data is going to be contradictory, because they are an aggregator and not a source. So for instance in one section they claim the overall rate of ad blocking in the US is 26.4%, yet in another section their data suggests 30% of mobile users are blocking ads, while 51% of desktop users are. Pretty hard to see how that adds up to 26.4% overall.

Regardless of the noise in the data, it at least seems clear that whatever the number is - it's way above 10%, and growing. This is probably why YouTube is willing to go to war with their users over it, one of the very few things Google has ever done that could directly imperil one of their monopolies. Potentially driving off a substantial number of your users at the time when there are more viable alternatives than ever is a quite a bold move. Let's see how it plays out.

[1] - https://www.statista.com/topics/3201/ad-blocking/#topicOverv...

I suspect they also want to push users onto YouTube Premium, since they are offering a free trial when you click through from the banner. (At least, this is the route that I took as a user.)

Reply on news.ycombinator.com