This analogy doesn't hold at all. Casinos are engineered for the house to win and the customer to lose because the casino is paying the customer directly. They are disincentivized to let the customer win. If the customer always won, the casino goes out of business very quickly.
Ad auctions are completely different, since they are just a marketplace. A customer's (advertiser's) win comes from a third party, and the point of the ad platform is to connect that third party to the customer. The ad platform would love to be able to make every single customer win all the time. Although, if they did that, it would create a lot more demand and even out the ROI with increased ad prices. But that's not the "evil ad platform" engineering the customers to always lose in the way casinos do, that's simply basic supply/demand economics.
COI disclaimer: My current job is a small part of an ads platform on a website you have heard of. I dislike ads but realize they are a necessity for running some parts of the internet.
Yea, I also work in ads and the article doesn't make sense at all. A way bigger factor than anything mentioned in this article is simply what you as an advertiser can afford as a CAC (cost to acquire a customer). There are some advertisers who are selling low cost/low repeat products that advertising is nearly impossible to make profitable. Meanwhile, other advertisers are selling high cost/high repeat products to that same person. The latter can afford to price the former out of ad inventory.
Many companies are spending significant amounts of money (many millions) and not even telling the advertising platforms the actual revenue they generated from the ads, just the number of "leads" or "customers" because they are paranoid about the very thing in the article (which I don't actually believe to be true anyway).
You have picked one dimension, incentives, and say the analogy doesn't work because on that dimension it's different. Analogies aren't supposed to be exactly the same.
The point of the analogy stands - most players using paid advertising can't win. They don't realize the entire structure of the market makes it close to impossible for them to win. Google (and a small number of others) have an extremely strong position in what is effectively the distribution channel for a lot of businesses, and a pricing mechanism which enables them to take a giant chunk of economics of the transaction.
It's an entirely AI-generated article. There's almost certainly no human making a good-faith argument here. You're wasting your time arguing with an LLM, which is a pretty good illustration of why I think we should be shunning such content. Only so many hours in a day.
I'm now shunning any domain I don't recognize and de facto upranking any domain I recognize from being around for more than five years. Which is unfortunate because not every old domain deserves a positive reputation, they've only gained that by default relative to most of what's "published" today.
Oh I'm very aware of that. I'm not trying to arguing against the article. I'm trying to inform the dozens of people who are upvoting this post and implicitly think it makes any semblance of a good point.
Not strictly necessary, but functionally necessary. Many websites provide services that are expensive to run and can't be funded by other products from the company, so the website itself has to be profitable. I'm not convinced that these websites (e.g. most social media) are societally necessary, but I don't see any scenario where they continue to exist without advertising (barring extreme measures such as government funding).
People who believe ads are necessary have an underlying assumption that commerce is also a requirement. It's a shaky basis, and leads to a bunch of other messy things (ads to demographic to profiles to tracking)
I use to think we lived in an enlightened age, where we have now learned that we need to use scientific reasoning and analysis to make decisions. It is what lifted us up from superstition and folk tales.
But society, still, collectively, falls for snake oil, over and over again.
Anyone that has an interest in the topic should listen (or read) about what
Steve Tadelis did in the 2012 timeframe. He was an economics professor at the University of California, Berkeley, and then worked at eBay, where he led the famous experiment on Google ad effectiveness.
People are born in ignorance. Reason takes energy. Nature optimizes for energy efficiency.
This is likely a hole us meat puppets can never crawl out of. Any moment we stop spending energy to climb out, we're sliding down the ant lion nest to ignorance.
i can't square away all of this with the simple math that if you make a product that makes sense for online advertising, you make boatloads of money by using online ads, and make less as more people compete in it.
this is true of GLP1 ads and for many, many subscription services like video streaming.
i'm not sure it's about falling for something so much as it is that people who like, sell products through amazon, well, they are fucked no matter what.
and that somebody is looking seven generations (or even a seventh of a generation) ahead seems so strange in 2026 when so many people think they can enjoy being poor forever (like billions of years) if they don't manage to pound out 15 ai slop blog posts to prove to everyone else that they know where ai is going.
I don't really think this is true, as if advertising didn't work people wouldn't use it.
That being said, there is something to be said about how extractive it has become.
It used to be you ran a TV ad, and if you made a great ad, you got amazing ROI, you were rewarded for being clever.
Now, it's all captured algorithmically, there are no "great ads", you can only have good ads or bad ads. They know full well if your ad is great then they can charge you whatever they want, inevitably making your ROI only "good enough". Sure go to another platform where the algorithm will price you out eventually.
Look at Him & Hers / Ro / GoodRx / etc. They basically take every dollar they make and throw it into advertising, it's the only way to stay relevant.
Psychological elements aside, this sort of misses that goods and services require multiple inputs from different counterparties (or different markets) that have different pricing power (some of which might be close to monopolies). It’s implausible that whenever there is one monopoly seller anywhere that all added economic value (or more than all, as in the article??) has to go to that one.
If I have to necessarily buy ads from google and water from the local water company, where does the economic value go in equilibrium? Clearly not all to the water company and not all just to google.
The gambling analogy doesn’t quite hold here. Gambling is a zero sum game. Advertising is not a zero sum game. There is room for both sides to walk away with gains from trade.
It's often constant sum, because in ads can move sales between organisations instead of creating value.
With the automated bidder maximising ad prices for shareholder returns it can become zero sum equivalent. Every gain is matched by a loss elsewhere.
This changes if ads are creating new demand, or if consumers are getting value they otherwise wouldn't have found. But in many markets that won't be the case.
The article explains why that isn't the case: it's because advertisers (Google's clients who run ads) are competing with each other to push the cost of advertising up to the point where the cost of an ad and the value it provided reach a rough equilibrium.
And in economic theory, perfect competition drives the margin to zero. The gambling part is not explaining that theory, but rather the effect on human behavior when the short term results hover around that zero margin line.
That wasn't the analogy. It's gambling in the sense "the house always wins".
It's not a simple "both sides" issue - the scenario has multiple sides. You find a good keyword, someone finds you, you both benefit. Great!
Then a third party finds and uses your strategy, and bids higher in the ad auction. This reduces profits. As more parties join in, the profits decrease.
The thesis is that those profits don't simply go to zero, but can actually result in a net loss. In the meanwhile, those higher bids end up as profits for the auction house.
Uh, yeah the house always wins because it is zero sum. In a competitive market there is gains from trades. Rational actors can drive consumer and producer surplus to zero but once rational actors won't continue to buy at a loss.
Gambling, advertising, trading, and AI have one thing in common: if you don’t do well it’s your fault. You didn’t prompt right, or target the right audience, or you don’t have the right harness, and the only way to remedy is always to spend more money.
The only way to keep out of those traps is to have real metrics of your own and benchmarks of your own. Performance changes over time so you can never trust the current state will continue.
The whole ecosystem is extractive with the goal of improving quarterly earnings by removing control from the people spending the money. Online ads started out being trackable and much higher ROI than traditional media. Then they ran traditional media out of business while continously raising their own rates (opaque quality score, nebulous targeting options) and then go even further, essentially stealing publishers' content for their own walled garden models.
Amazon got fined a bajillion dollars for deciding to add undisclosed reserve prices for their ad auctions. They just decided that too many auctions ended too low, so they added fake reserve prices and just chose what the price would be...
Google and Meta have also both been caught in advertising shenanigans -- I forget the details but I recall getting some (tiny) class action settlement check.
I drive by an Ikea on my way to work. They just installed a video wall in view of the highway. It is huge, four stories tall and square. It gets noticed but is not intrusive. It is not animated. In daylight it just looks like a big painting. At night it isn't any brighter than the parking lot around it. Yesterday it was advertising an end table for 65$. This morning it was something with a plant in the background. I actually saw and remember that ad. I almost never see ads on websites and certainly don't remember those I do. But that Ikea wall... THAT I remember.
Ad auctions are completely different, since they are just a marketplace. A customer's (advertiser's) win comes from a third party, and the point of the ad platform is to connect that third party to the customer. The ad platform would love to be able to make every single customer win all the time. Although, if they did that, it would create a lot more demand and even out the ROI with increased ad prices. But that's not the "evil ad platform" engineering the customers to always lose in the way casinos do, that's simply basic supply/demand economics.
COI disclaimer: My current job is a small part of an ads platform on a website you have heard of. I dislike ads but realize they are a necessity for running some parts of the internet.
Many companies are spending significant amounts of money (many millions) and not even telling the advertising platforms the actual revenue they generated from the ads, just the number of "leads" or "customers" because they are paranoid about the very thing in the article (which I don't actually believe to be true anyway).
The point of the analogy stands - most players using paid advertising can't win. They don't realize the entire structure of the market makes it close to impossible for them to win. Google (and a small number of others) have an extremely strong position in what is effectively the distribution channel for a lot of businesses, and a pricing mechanism which enables them to take a giant chunk of economics of the transaction.
Can you expand on this? Why are they necessary for some parts?
But society, still, collectively, falls for snake oil, over and over again.
This topic was talked about in 2019. https://news.ycombinator.com/item?id=21465873
Freakanomics did a great podcast about if Internet Advertising actually works. https://news.ycombinator.com/item?id=47906485
Anyone that has an interest in the topic should listen (or read) about what Steve Tadelis did in the 2012 timeframe. He was an economics professor at the University of California, Berkeley, and then worked at eBay, where he led the famous experiment on Google ad effectiveness.
This is likely a hole us meat puppets can never crawl out of. Any moment we stop spending energy to climb out, we're sliding down the ant lion nest to ignorance.
this is true of GLP1 ads and for many, many subscription services like video streaming.
i'm not sure it's about falling for something so much as it is that people who like, sell products through amazon, well, they are fucked no matter what.
Anyone with an IQ in that range or lower absolutely needs a reliable source to just tell them what to do and how to live.
Social media and weaponized right wing "news" have set all those reliable sources on fire.
Like lately I have been going to talks by Native Americans whose spirituality is rooted in place, relationships, commitment and all that
https://americanindian.si.edu/environment/pdf/01_02_Thanksgi...
and that somebody is looking seven generations (or even a seventh of a generation) ahead seems so strange in 2026 when so many people think they can enjoy being poor forever (like billions of years) if they don't manage to pound out 15 ai slop blog posts to prove to everyone else that they know where ai is going.
That being said, there is something to be said about how extractive it has become.
It used to be you ran a TV ad, and if you made a great ad, you got amazing ROI, you were rewarded for being clever.
Now, it's all captured algorithmically, there are no "great ads", you can only have good ads or bad ads. They know full well if your ad is great then they can charge you whatever they want, inevitably making your ROI only "good enough". Sure go to another platform where the algorithm will price you out eventually.
Look at Him & Hers / Ro / GoodRx / etc. They basically take every dollar they make and throw it into advertising, it's the only way to stay relevant.
If I have to necessarily buy ads from google and water from the local water company, where does the economic value go in equilibrium? Clearly not all to the water company and not all just to google.
With the automated bidder maximising ad prices for shareholder returns it can become zero sum equivalent. Every gain is matched by a loss elsewhere.
This changes if ads are creating new demand, or if consumers are getting value they otherwise wouldn't have found. But in many markets that won't be the case.
This means is often
It's not a simple "both sides" issue - the scenario has multiple sides. You find a good keyword, someone finds you, you both benefit. Great!
Then a third party finds and uses your strategy, and bids higher in the ad auction. This reduces profits. As more parties join in, the profits decrease.
The thesis is that those profits don't simply go to zero, but can actually result in a net loss. In the meanwhile, those higher bids end up as profits for the auction house.
I'm picking on the gambling analogy. It is wrong.
I agree with the general thesis of the article.
The only way to keep out of those traps is to have real metrics of your own and benchmarks of your own. Performance changes over time so you can never trust the current state will continue.
Every marketing channel has its pros and cons. Advertising is no exception.
Online adds are a fool's errand.