Hacker News new | past | comments | ask | show | jobs | submit login
How Digital Advertising Markets Really Work (prospect.org)
137 points by bingden 76 days ago | hide | past | web | favorite | 28 comments

I’ve worked around the ad ecosystem for a long time as a media buying agency, then an ad network, then a publisher. I think this article is spot on. Few understand the iron grasp Google has on the ecosystem. Soon they are changing the model to allow for fair competition, but thus far Google AdX has had a massive advantage by having the first and last price in every bid request. They get to see what other ad networks are paying for each request. They get to optimize their profits using data from users, publishers, ad exchanges, and advertisers, mostly because of the absolute monopoly Google DFP/Ad Manager has in the ad server market. If I were a regulator that would be my first stop.

If AdX was more open, how do you think the market would change?

For starters that would revive the 3rd party ad server ecosystem, since currently the only way to access full AdX demand is to run DFP. It would put AdX on the same footing as others, meaning buying through Google wouldn't give buyers a price advantage. It would allow for privacy-centric ad networks on vetted, premium publishers to emerge. Maybe even header bidding could die, meaning front-end performance would increase 10X on many publisher sites (many header bidding stacks make 300-800 ajax calls in the first few seconds of page load!) which would eliminate the need for AMP. It would mean Google would have less influence over which publishers get to monetize their content. I could go on...

> Consumers can pay for ad-free video subscription services like Spotify Premium or Hulu—why can’t they in search or social networking? Why are these markets stuck?

The article has very pragmatic observatiosn but this is not one of them. This won't work, people won't pay with money for what they can easily pay with their targetting data. Private data is worthless in any other context than advertising, there is no opportunity cost.

I also think it didn't touch on advertisers themselves, who bear a big responsibility for this situation. Agencies like IAB have done less than nothing to fix this situation, they ve made things worse. We need more ethical ad platforms, but we also need more ethical advertisers.

The richest people are the ones who advertisers really want to reach (in the article, the hypothetical beer ad sells at $2 CPM vs the hypothetical private jet ad at a $2000 CPM). At the same time, the richest people are the ones who can afford to pay whatever subscription fee FB or Google would charge. I don't think that there is a lack of demand, it's just that the people who make up that demand create an adverse selection problem, the users who would opt out are the ones who make your platform valuable.

I don't think that antitrust regulation solves this particular adverse selection problem. It's certainly difficult to compete against the established giants, for many reasons, but this theoretical opting out is definitely not some consumer benefit that we can expect.

It's not really the same thing you're talking about at all, but I think there's also an interesting conflict of this type between where the largest sites are getting their user growth from now, and the users that are most valuable as advertising targets to them.

The larger sites (Facebook, Twitter, etc.) are pretty close to saturation in North America and a lot of other English-speaking countries. There just aren't many people left here that don't know what Facebook is and might still be convinced they should start using it regularly. Almost all of the growth comes from other parts of the world, but the users from those locations are worth far less in terms of advertising revenue.

If you look at Facebook's Q2 2019 presentation [1], their average revenue per user from US & Canada is now $33.27, Europe $10.70, Asia-Pacific $3.04, and Rest of World $2.13. One US/Canada user is worth more to them than 15 from "Rest of World".

However, 74% of their Monthly Active Users are from Asia-Pacific/Rest-of-World now, and that only corresponds to 28% of their revenue.

I don't really have a particular point I'm trying to make, I just think it's interesting. These companies worry a lot about showing continued user growth (or at least not shrinking), but the growth sources now don't have anywhere near the same value as the old ones.

[1]: https://s21.q4cdn.com/399680738/files/doc_financials/2019/Q2...

continued growth is good for their stock. It will of course reach a tipping point where the sheer amount of users will start yielding net negative revenue at which point there will be a reckoning

"Facebook made an average of $6.18 off each user in Q4". But your point is that price discrimination really matters. The same thing with whales playing "free" games. Gotta agree with your main point.

> The richest people are the ones who advertisers really want to reach (in the article, the hypothetical beer ad sells at $2 CPM vs the hypothetical private jet ad at a $2000 CPM)

Rubbish. The 1000x $2 eyeballs is definitely worth money to the beer sellers, independently to the 1x "$2000" eyeball for the private jet advertiser (ignoring profit margin etc). Note, from the article: "The private jet ad might sell at a $200 CPM as opposed to the $2 CPM beer ad targeted to an anonymous user" not sure why you changed value to 2k.

I at least would pay good money every month for a google suite free of all the bullshit they've been adding the last ten years. Give me relevant search results again please take my money.

This is absolutely correct.

Not only is an individual's private data nearly worthless, but they can protect it for free simply by using the Brave browser or other adblocking/tracking solutions.

People often claim that they would pay for web content if it were somehow easier or if micropayments worked better or something. I'm not sure that I believe them.

> I'm not sure that I believe them.

You don't have to take their word that people would pay. They already are paying. Millions[1] of people go out of their way[2] to pay creators monthly or per-work.

However, a lot of that "web content" needs to realize the actual value of their content might be ~$0. Advertising distorted the market; a lot of people were able to extract revenue greater than the actual "market value" of their content.

[1] https://graphtreon.com/patreon-stats

[2] While Patreon isn't very hard, having to take the extra step of vising (and maybe making an account) a 3rd party service is not a proper "micropayment" system. The goal of micropayments is to make trivial to immediately pay for something without any extra friction (perhaps a button/whatever in the browser to send a tip/donation, no need to worry about Paypal/Patreon/etc)

> a lot of that "web content" needs to realize the actual value of their content might be ~$0

If people are spending time on your site, you are providing value to them (excluding scams, of course). It's never 0.

> Advertising distorted the market; a lot of people were abl

It did not. The percentage of GDP that companies spend in advertising is historically the same for more than 100 years. What happened with the internet is that Google and later FB progressively sucked up ALL the profits, leaving literally scraps to publishers. And over time they are eating more and more value from publishers because all their competitors have vanished (there is currently NO competitor to adsense). Publishers are seing their ad CPMs being lower , despite the fact that online audiences are growing fast, and the shift to online ads has accelerated explosively in the past few years.

Patreon and subscriptions have their place, but they are not scalable. Users will revolt if they have to pay 30 yearly subscriptions for reading 30 articles , and are forced to pass from an spanish inquisition to unsubscribe. Subscriptions can also be bad for the quality of content: If creators try to appeal to their patrons/subscribers instead of trying to reach as wide audiences as possible, they tend to become more partisan and biased ; you get what you incentivize.

micropayments would indeed be far better especially if they were anonymous, fire-and-forget payments. crypto payments would be ideal for that but its not gonna happen because they are untaxable

This is nowhere near the level of actual content consumption. Payment models have been tried endlessly for decades and none have been successful yet in gaining any serious usage.

I've been happily using ad blockers for years, so every time I get to use someone else's computer I am reminded of what a horrible place the web has become.

On the other hand, I'd like to support creators as well, so I've created a web site that allows readers to make payments easily. It's like scroll, quid, patreon, etc. but I've tried to reduce friction even further by only going with a HN-style sign up form (email/pass) and charging readers only once they've reached a minimum account balance:


My initial impression is that most people will just continue blocking ads everywhere, because it's the easiest and safest thing to do, and I'm not even sure I disagree with that approach.

Any feedback is appreciated.

If you look at the acquisitions of Google in the 00's, it's pretty obvious : in 03 they buy AdSense, 05 Analytics and Android, then in 08 Doubleclick, having their hands every part of the supply chain.

Facebook then came along and followed the playbook with the social buttons.

Brilliant move, regardless of how we feel about that.

It's the advertising market, not the companies, that are the root problem.

Go ahead and get rid of Google and Facebook and the rest overnight and all you'll do is find a new vector for advertising spring up virtually overnight.

Every ad tech company is responding to market incentives. Breaking companies up or killing them with regulation etc... won't make a dent in the long run.

Outlawing advertising as a business however will end the surveillance economy swiftly.

I don't understand the conclusion: "But if we understand that user identity and user data are drivers of ad pricing, and that the terms that Google and Facebook are able to extract from independent publishers drive the health of the wider news ecosystem, it is impossible to continue to make this argument today."

Earlier, the article explains pretty well how tracking increases the CPMs that publishers can achieve. In the author's example, cross-site tracking increases CPMs by 200%.

This seems like a benefit to the "health of the wider news ecosystem". If the author believes Google/Facebook increase the pie for publisher, why then say that they "extract from independent publishers"?

This is a really important topic that I wish the author had addressed. The publisher often doesn't make more money from the $200 CPM ad than it does the $2 CPM ad. Most of the difference goes to the ad network.

Also - pretty much all ad networks are rounding errors compared to Google and FB, which make up 85% of all new ad spend. The last 15% is mostly Amazon, LinkedIn, Twitter, Pinterest.

Yahoo/AOL have to still be in top 5 right? I’d think Microsoft (even outside of LinkedIn) and Snapchat should have decent percentages of the leftover too. At least the former.

Yahoo/AOL have terrible targeting and no bot protection, so generate very little revenue.

What do you mean? They generate billions in revenue. Are you saying they aren’t ahead of every company you mentioned except Amazon? Microsoft as a whole would be ahead too, but you only mentioned LinkedIn. Verizon Media is certainly ahead of those.

Maybe the regulators can say to Google or Facebook that they can't collect those user data, or they limit the collection somehow. I can't imagine how else you can stop this duopoly. The problem with the whole ecosystem is that every system they have feed the other one, ie. Chrome gathers the search data to be used for Ads etc.

This might be solved with AI. Basically you will have a personal AI that will shadow browse with you in the beginning to understand your tastes. After some time, the AI will do the browsing for you, and if needed can take adversarial routes to confuse the ad networks.

Meantime, the AI can update a private site with content that interest you.

One of the best articles I've read on the internet so far. Eye opening.

The core issue here is that companies uses AI against humans. Humans can fight back only by using AI against AI.

An client side AI is the only way to make tracking useless.

“That 22-year-old communications major has had to make way for data scientists, mathematicians, and computer programmers who, behind the scenes, use statistics, calculus, and linear algebra to optimize advertising campaigns, by micro-targeting users and constantly tweaking algorithms.”

Click the source link, and it’s an article about how companies are wasting money on data scientists when they only need data analysts, contradictory to the point made by primary article. This was an intriguing article, but I could not continue reading after noticing this.

You should read the rest - it's a good article. You are right about that point , and more broadly i m not convinced that advertisers can justify their spending and their bids well. Advertising seems to be a constant for businesses - they will spend those money to marketing regardless. And ads are effective to a degree - but i have yet to see a rigorous study of the correlation between amount spent or bids to ROI. Judging from the iffy nature of advertising statistics and google's "invalid traffic" alchemies, i believe google is also selling large amounts of bullshit along with their basic ad product. It's bizzare that an industry that is so valuable seems to be operating in a rather thick haze.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact